CSS Button No Image Css3Menu.com
New! Search comments:
(NOTE: Relevance, Author, and Article are not applicable for comment searches)
Not possible. But <span class="statdef"><a href="http://www.baseballprospectus.com/glossary/index.php?search=RP" onmouseover="doTooltip(event, jpfl_getStat('RP'))" onmouseout="hideTip()">RP</span></a>'s who come in in tied games always have plenty of innings.
Plus, it's a very high leverage situation. I prefer Scoresheet's reliever usage.
To maximize quality innings, make him a closer with an earliest inning of 8.
This isn't the right way to go about this, I think.
I can get awesome correlations for (say) organic food availability vs. autism diagnoses. The question is what the best combination of slugging and on-base matches the runs scored.
There are a couple of ways to do this. One is to go to a multiplicative method - which is how baseball works. Run scoring is (very roughly) <span class="statdef"><a href="http://www.baseballprospectus.com/glossary/index.php?search=OBP" onmouseover="doTooltip(event, jpfl_getStat('OBP'))" onmouseout="hideTip()">OBP</span></a> x Slugging.
Another is to look at a linear weights for each season.
This page: http://www.fangraphs.com/library/principles/linear-weights/ has linear weights relative to outs. Additional bases from a single/double/triple/homer range from 0.27 runs to 0.38 runs.
Simple calculation: I get a double and a single, you get a homer and an out. That's 1.000/1.500 for me, .500/2.000 for you. wOBA or linear weights are going to give me a higher expectation, despite the fact you got a homer, and homers are good (upgrading a triple to a homer is better than upgrading a double to a triple by a fair amount.)
If your finding is correct, it defies effectively all prior research on the subject. (A walk is better than an extra base.) This does not mean you are wrong, but I believe everyone else is right and you are mistaken. Correlation does indeed imply causation, but only loosely.
Finally, an error that this article had on initial publication elsewhere still is here: "On-base percentage has a higher correlation coefficient to scoring than slugging percentage for the period 1913-2015. But slugging percentage explains scoring better in the period 1939-2015 and every subsequent span ending in the present." I think there's an error there.
If I'm wrong on this, I hope someone corrects me. It seems to me that regression analysis is much stronger than simple correlation coefficients. Also if I'm wrong, you're the first author I've seen to point this out, so congrats.
But I'm tentatively putting this in the interestingly-mistaken bin.
I hadn't considered the accessibility issue, and appreciate the reasoning. I still suspect it's better to go off the site numbers, (because people here will be more likely to be subscribers, and the book-buyers can still play the other way), but I get it.
It's a great contest idea and I'm looking forward to the results.
I love the contest idea. Can the crowd beat <span class="statdef"><a href="http://www.baseballprospectus.com/glossary/index.php?search=PECOTA" onmouseover="doTooltip(event, jpfl_getStat('PECOTA'))" onmouseout="hideTip()">PECOTA</span></a>? As an occasional PECOTA complainer, I ought to be able to do it.
Execution complaint: Seriously, you want me to look up all these in the book? That's a massive PITA, and there's no good reason for it. (I own the book.)
I'm gonna go off the site numbers, as you suggest. And then I'll score myself at the end - based off the site numbers. There are differences, and they are not huge most of the time, but they are not nominal.
I enjoyed the piece. The initial anecdote strikes me as at least a partial fallacy, though.
While I think confidence does help in some fields, a substantial confidence/competence ratio - where confidence outstrips ability by a substantial amount - is a problem.
Take some of your readers. They may be quite confident in some things - perhaps math or chess. But, to stereotype from one example, they may be unconfident in their dancing.
The direction is competence --> confidence, as it should be. Assuming that the correlation means (say) I should be more confident in my keen fashion sense in order to have more strikes me as misguided.
Thinking you're a little or somewhat better than you are may have advantages. But thinking you're way better than you are ends too often in failure modes, right?
(Oh, and the amount you can think you are better than you are? 40% Because I've been calling it the 40% rule for 30 years, and I am *so much better* at quantifying excess confidence than other people that my number should stand despite its resistance to objective analysis.)
You say that Sax was imprisoned for crimes he didn't commit, but there were a lot of unsolved murders in New York that year.
A lot of them.
First example: Take vaccine, zero percent chance of getting Sudden Death Syndrome.
Don't take vaccine, 0.1% chance of getting SDS.
Second example: Don't volunteer, zero percent chance of getting SDS.
Volunteer: 0.1% chance.
There's *no* practical difference here. It's action vs. inaction, yes, but whatever 0.1% of your life value is is the right answer. It's possible to logically end up slightly different because of decreasing marginal utility of money, but there's no way that difference should be this severe.
I'm with you, Jeff, and I'm not a senior citizen - I only look 80 years old.
Or I didn't get the joke. One of those.
Great gadget, but I have some questions:
Addison Russell appears to have been kept in 253 leagues, or 10.3%. Wait, what? That would give us 89.7% of the leagues which protect one guy or have dumb children in them. Also, it gives us about 2500 leagues total.
There are enough other similar weirdnesses that I think I misunderstand what at least one of the columns meant. (I thought it meant: Number of leagues kept in BL/AL/NL, then percentage of time kept.)
Could you clarify what the numbers mean for the slower readers? Thanks.
Excellent. A global list like this took a ton of work, I'm sure. I salute the effort.
Thanks for the list and participation in the comments, guys!
For posterity's sake, I'd like to record my prediction now that Meadows is non-trivially better than Clint Frazier. (Frazier was listed above Meadows on almost all lists last year, too, and I thought that was a mistake then.)
If you screech about a trade, be right.
If the guy who just acquired the rookie of the year (let's say, Angel Berroa) for a pick has stolen, you should be willing to offer substantially more and make for instant profit for that guy.
I think the best solution in a high-end league is a Trade Whinge rule: You can Trade Whinge, but either of the people involved in the trade can Rebut by betting future picks on the outcome of the trade. (There are some more details, but this ought to cure the overwhingers more quickly than anything else.)
Oh, and if you want to bet, shop the books. You'd think the lines would run even or evenish - and that equalizing force does exist - but the Mets and Mariners lines were much better at the spots I bet them.
Other tip: As the article notes, it's all about the odds. Have an idea of what odds you will take on your preferred bets walking in; this prevents anchoring issues. ("Hey, look, the Astros are 20-1 at Book A, but 40-1 at Book B! I should bet on them at Book B!")
There are a bunch of different books:
1. Caesar's Entertainment books. They are at lots of hotels.
2. Venetian/Palazzo (run by Cantor). Terrific books, non-awful vigs. They were where I got my Indians in playoffs bet.
3. MGM Resorts.
4. LVH (formerly Las Vegas Hilton, now Las Vegas Hotel.) Off-strip, but accessible by monorail. Tons of betting options.
5. Riviera (run by William Hill.) Substantial vigs, but tons of prop bets available.
I think I'm missing one. Look for the betting sheets; lots of bets available that are not available on the big board.
(Finally, I have the Caesar's sheet here; teams must play 160 games for over/under action there, not 162. But, the over-unders are +120 each, with a push at +300. That's a very light vig, *and* the +300 bet strikes me as fabulously overoptimistic; very tempting option.)
What book is that at? I've never seen a requirement that teams play all 162. I saw some really sweet lines at one book that had the pushes at 3-1 but +105 for the over/unders (pushes losing). There were some tempting lines there.
Anyway, I bet:
Indians to make the playoffs at +350
Indians to win the WS at 45-1
Mets to win the division at 28-1
Mariners to win the division at 12-1
Blue Jays to win the division at 13-1
Trumbo to hit most homers at 25-1
Rizzo to hit most homers at 100-1
Bet I would have made if it were easier to collect smaller wins or if I had enough disposable income to make a very large bet: NYM + 13 1/2 wins over NYY at +105.
The Astros opened up at the Atlantis with an over/under of 57 1/2, IIRC.
Also, bet on baseball. There is still profit to be made, despite the fact that the books and the big money have gotten smarter.
Wow, I was enjoying this....
A point and a complaint:
Point: I am a huge fan of betting on the future; if your tools don't let you predict what's coming, they're fundamentally uninteresting. Bill James famously rooted hard for a former Red Sox prospect his methodology had as the batting average leader. I think the prospect got traded for Larry Andersen and was never heard from again. I liked the future-prediction-makes-rubber-hit-road observation very much.
Complaint: Seriously, "No atheists in foxholes." Seriously. Really. So Pat Tillman converted? That would be news to his brother. Ted Williams converted and de-converted? Each flight?
It's a provably false aphorism; there are plenty to, erm, testify that they faced imminent deadly peril and did not convert. I'm one. (Non-military setting.)
Rougned Odor's stock's probably up too, though he won't be ready in 2014. Rosenthal's reporting that the Rangers may not be done moving middle infielders.
I've edited Braun's statement to make it shorter and more credible.
Now that the MLB investigation is over, I would like to apologize to everyone for getting caught twice.
Getting caught was the most embarrassing thing to happen to me. The first time, I successfully blamed others, and I regret that did not work longer.
The use of something disallowed occurred when my grandmother was dying and I had been diagnosed with testicular, lung, and brain cancer. I took two pills to cure the cancer, and they also caused positive tests, which are my fault, kinda. And then my grandmother died.
I would like to apologize to people for getting caught again. You would think I would be able to avoid getting caught again with my vast resources and the fact that I only took two pills to cure my cancer and save my grandmother, who undied because I took the pills only to die again later out of grief because of MLB's intrusive anti-Ryan-Braun policies. But it's my fault in the end for getting caught.
I would like to apologize to the specimen collector. No matter how much of an insignificant God-hating peon he is compared to me, I should not have gotten caught lying about him.
I remain filled with regret over this incident. I swear on all that is holy that I shall never get caught again.
I liked the article; Sobsey gave 110%. Carleton's a gamer. No one wants it more than R.J. Anderson.
Lindbergh plays the game (of writing) the way it was meant to be played. Ferrin's a great clubhouse guy. Sussman's a scrapper.
Rozenson is, pure and a simple, a baseball player (writer-about.) Cole keeps it light, helping the whole ballclub.
I thought maybe Mike Fast would have one here, but he gone.
What about Bryce Harper's 763rd home run? Or are you thinking 2014 for that?
Joe, thanks for this.
As a pitchfork-and-torch salesman to the PECOTA mob the past few years, I've been very disappointed in the contrast between confidence level expressed and actual performance. (When Wayne Causey is your best comp for Bryce Harper, you're doing it wrong; predicted results for teams should average fewer than 86 wins.)
I hope that things go well on this front, and I wish y'all the best. I agree that there's a significant process problem. There are some good things BP is doing (Scoresheet Draft Aid!), and there are good articles. I'm glad to see that efforts are being made to fix some product rather than just have BP go to pure cash-cow mode.
Shorter ramble: Humility - while often a vice - is good if it leads to fixing.
I always thought that everyone had confirmation bias, and this proves it.
OK, I'm confused: Why is using multiplication and division the wrong way to use park factors?
CAIRO projected Hamilton to hit 27 homers in Arlington in 2013. I'd look at the three year park factors for homers (1.26 for Arlington and .79 for Anaheim), take the 15 we'd project for Arlington and reduce it by about 5.5 HR's. And that's a lot, for sure.
What I can't figure out is why that's going to radically overstate the loss, assuming no park changes in the last three years (and assuming the 2011 outlier result in Arlington is properly factored in at full weight.) Are you asserting that it's wrong because even three-year park factors are too noisy? I'm genuinely baffled, and would be pleased to be unbaffled.
In some respects, yes, the Royals cashed out their risk; Shields is probably worth c. $22 million per year; they'll pay him $21 million total, for a $23 million profit. It would be surprising if those numbers were way, way off.
The question is what Myers will do. If Myers is excellent, it's a huge loss; Myers is cost-controlled for a long-time. Even if he's just good, there's a lot of (performance-cost) profit to be made.
If the Royals don't contend in the next two years, it would be very hard to win this trade. Like a lot of people, I view this as an effort by Dayton Moore not to get fired; I'm not too convinced it will work. But I also agree that full garment-rending sobbing may not quite be called for.
I bet on the futures every year in Vegas. This year, I bet on the White Sox to win the AL Central at 15-1 (OK, and the Mets at 50-1); ah, well.
I don't bet online because I think it's illegal. But the idea that the line is unbeatable strikes me as mistaken.
The comparison of odds on a multi-pick to a single game seems unfair to the linemakers. If you make one significant mistake as a linemaker on one of those teams, you might get killed and eaten. While that overall vigorish looks daunting, the vig on any single team isn't.
And I'm convinced the futures lines are beatable. (On the above lines, I like TB, Detroit, KC, Seattle, and Colorado.) The Venetian's book will have lines on the divisional races and a lightly lower vig. Shopping lines is a good idea. Plus, it's fun.
I mean, unless you missed betting in 2008 because a trial went long and you wanted to bet on the Rays to win the AL at 45-1. If you did that you might be bitter and mocked for years by your friends, who are all terrible people who should be eaten by landsharks. Theoretically.
"Both would have paid him 90% of their entire budget to get a deal done."
Wow. That seems really unwise. Buxton got a $6 million signing bonus, $200K under slot for pick 1.2. If you were to give Buxton 90% of your draft money, you'd have to commit to that up front, then significantly overdraft players who would sign in the rest of the draft. (You forfeit draft money for picks who are unsigned.)
Houston's method - drafting Correa perhaps a touch early to free up signficant slot space later - seems a much more likely outcome than committing to run most of your draft money into pick 1. I don't think I'm going to hire those scouting directors as negotiators; I think they're not just wrong, but badly, foolishly wrong.
I'm looking forward to your rankings - they should be Parksian and fun.
I disagree that the rankings are just kitchen appliances, a matter of preference. Once you have sample size, you can see who has done better over time; projections are and should be a serious business where results matter.
I'd warn you away from some of the frequent failure modes:
1. Certainty due to non-stat factors for hitters. One prospect list left Brett Lawrie completely off some time ago. They said because he was a redass (my word), he didn't deserve to be in the top 100+ of prospects. Ignore data at your peril; factoring in outside stuff is probably good, but overusing it is hazardous.
2. Fumbles. Your predecessor was a terrific writer and well-connected. But when you don't put Xander Bogaerts on your midseason Top 50, you've fumbled. KG took the position, as near as I could tell, that the critics were misguided Sox fans who needed more in their life. In the case of Bogaerts, the better defense would have been, "Crap, I fumbled."
3. Young pitcher bias. Young pitchers get hurt. Be cautious.
4. Failure to adequately account for hitter age. Hitter age is huge. (See: Jazayerli, R.) Xander is a better prospect than Travis d'Arnaud.
5. Velocity worries for lefties. Velocity worries for righties are completely legitimate; I was (I think) the first to publish evidence that you could improve projections by using velocity. Velocity worries for lefties are often overblown, though.
6. Small Sample Size Sucker Syndrome. You'll see a minor league player not-enough. Four PA's is not enough. (Pitching-wise, five innings might be enough to give you an idea of velo and movement; caution is warranted, but less caution.)
7. Surprise picks are cool! Except if you drift way far away from the pack. Then you're probably just wrong. You should have a *very* good reason to *seriously* deviate from the consensus. (e.g. Special knowledge you have; sharing cocaine with the player; being John Sickels and therefore usually right.)
Best of luck. Enjoy the crafting.
I think I don't understand this post.
To the extent I do understand it, it seems to say that measurements of run and win values and valuation of intangibles is at issue, but that if you agreed on that, you'd clearly agree on who was the most valuable. If that's the assertion, I disbelieve it.
Value's a particularly hard problem even if you have good quantitative valuations. I don't understand exactly what "Value is being good at helping your team do the things that lead to scoring more runs than your opponent," means, but let's try some out:
1. Most wins added. This seems a good way to do it, but there's a fatal problem: A guy 90 runs over replacement adds more wins to a .350 team than to a .650 team. So, if "Value is being good at helping your team do the things that lead to scoring more runs than your opponent," means "Most added wins," it's not quite right.
2. Most runs added/subtracted from enemy. Not good enough, and by a longshot. Leverage matters. By my view, if you have a team with 93 wins and a zero run differential, some players on that team should get credit for the 12 wins over Pythag. But you don't have to agree with me.
3. Most runs added/subtracted from enemy, adding a Pythag adjustment for overperforming or underperforming teams and adding some small bonus for playoff teams. That is not "pretty clear" to me.
So we're going to disagree on value even where we agree on the accuracy of the numbers, the precision of the numbers, the value of intangibles for any given player... we're not reaching a place of agreement. How is "what value is" "pretty clear"? I'm baffled.
As nrs5000 notes, the deltas are way, way off. The playoff calculations generally are also suspect, to be kind.
See: Detroit and CWS. Note that the Tigers gained 1.4% on their division chances (they're actually up 1.4% with CWS down 1.4%) after a rainout between the two teams. And then there's the deltas, which are dramatically miscalc'ed, lending a certain lack of confidence in the process. This sort of thing is not a new complaint, and the continuation of it is regrettable. (Props to Kory for his good work and lack of citation to the facially errant playoff odds.)
Sam is correct on the main question - the Red Sox could have walked away and left the Dodgers with those the Dodgers had claimed. See, for instance:
The Red Sox thus could have left the Dodgers with Beckett, but it appears roughly certain (as Sam surmises) that they were in advanced trade talks and the Sox had agreed to pull back Beckett if they didn't make a deal involving Adrian.
Aside: Does anyone have a reliable figure on the amount of salary offset the Red Sox sent over? I heard it was $10 million, in which case this deal seems like a great opportunity for the Red Sox to reboot - that's Hamilton+Greinke money off the books.
The subhead on the front page starts "Toronto's Travis Snider," which perhaps could be read, "The forgotten Travis Snider."
Hit the yellow arrow up top. You can go back and forth and note the curious road the playoff odds take.
Cleveland went from 53.5% to win the division to 51.2%; Detroit went from 24.8% to 29.5%. To quote Ralph Wiggum, "That's unpossible!"
I enjoy the hit list; I enjoy Mr. Kory's style.
But referencing the playoff odds, even occasionally, seems misguided. Continuing to publish them in their present form seems regrettable. The Indians beat the Tigers yesterday, and Detroit's chance of winning the division went substatantially up and Cleveland's chance at the division win went down (not in real life, just here.) This is a continuing issue, and others have also noted it.
You're right that the deltas are broken (Good call!), but the odds are also broken. The Indians did have a measurable decline in playoff odds after beating the Tigers, and that's wrong.
Better if they were still the same, maybe. Rays win, go down 3.9% for playoffs. Cleveland beats Detroit, and their chance of getting in the playoffs declines non-trivially.
This is not plausible. (This could happen if a serious injury occurred and that was factored in, but that isn't what happened.) Prior years did not have these implausible results.
The alternative explanation, that the change to WC's changes the odds in very weird ways, doesn't appear to be correct. (The second alternative explanation, that the new way is right and the old way was wrong, is facially implausible because it is mistaken, I think.)
The playoff odds do not appear to be value added and should be scrubbed pending a fix or a convincing explanation.
Nope. Still a problem. Maybe it will be fixed on the next daily update?
As to pronunciation, it's TER-dos-SLA-vich, per the man himself. Don't know about the nickname. He is annoyed when people say "Turd-O-slav-ich," probably due to a lifetime of that.
(Jed Gyorko is JERK-o.)
It is vitally important that people understand that there is a right and wrong answer to this question. Randy Brown is right; the increase in diagnoses has not meant an increase in conditions. And the vaccine issue has been stomped to bits. (While I'm here, the facilitated communication scoundrels are deluded or criminal or some from each column.)
Good luck, Maury. I hope things get a little better over time.
Ohhh. As a SSSS (Small Sample Size Sucker), I love this. Thank you, from the protector of Ryan Raburn.
I've enjoyed your tenure; hope the ride of rest of your life goes well.
I think Sickels published a Top 120 on his site.
Fixing things is good. As I've noted ad nauseum, I think that sticking to your guns on some of PECOTA's perceived flaws has been mistaken; finding and fixing flaws is important. There should be a mental bonus, rather than a mental penalty, for fixing things.
On this topic, I think "Colin proved they should," in regards to using older data weighted higher than other projections should read, "Colin asserted they should." If there's proof in the linked article, I didn't see it.
I liked the book better than Richard did. It wasn't The Best Book Ever, but I enjoyed the writing. (Aside: I think Profar did get a writeup.) I felt like I knew more after reading it.
I'm sympathetic to the typos; there's no way to get a book like this out without them. Misspelling Gary Huckabay's name in the acknowledgements.... I found it amusing. The actual copy editors probably shouldn't.
I'm unsympathetic to the PECOTA comps, which are broken, and known to be broken. Oddly, they're less broken in the book than they are here; Bryce Harper comps to Travis Snider rather than Wayne Causey. (The change in comps causes almost no change in projection.)
Hope this is helpful.
Did the tweet also talk about Beltran's career-ending injury, *****'s steroid use, Rose's reinstatement, or Strasburg's visit to James Andrews?
This isn't just Scoresheet-friendly, it's Scoresheet-excellent. I'm impressed.
How much do we have to pay to bar our leaguemates' access?
I think the questions are a little oversized, but I wouldn't bring the font all the way down, just a little down.
Like others, I do enjoy the chatter format.
I can't imagine that Bautista's numbers aren't a good reflection of Bautista. The projection should be accurate for him, if the system works.
The reason there are screams of horror over PECOTA outlier projections is that over the last couple of years, some of those projections have been obviously wrong. The Kila/Bowker problem of last year was severe and there were articles citing Kila as good based on PECOTA. This was despite the repeated and persistent observations by readers that this was wrong. I specifically outlined why the system was doing this, and why it was wrong, and I did so before Kila or Bowker or any of the other 26+-aged minor leaguers who were being misprojected got a 2011 plate appearance.
I would therefore be cautious in saying it's reader error to dismiss projections that look wrong by eyeball.
PECOTA now seems to have very serious problems with the very young. The comp lists are blatantly fouled. I believe the book PECOTA's are even worse for them. I'm ready to bet the over on Mike Trout, Freddie Freeman, and Eric Hosmer, among others. (There are other problems, like the Chipper problem.)
Maybe I'm wrong about this year; that's always possible. But, again, before you discount your readers' concerns about PECOTA, you should remember just how right your readers have been in the past.
Quick hits, because I gotta go to work:
1. The obvious problems last year were the Kila/Bowker problem, the minor league comps problem, and the long-term aging curves looking ridiculous.
2. Some efforts have been made. Still, Bryce Harper's number one comp is Wayne Causey. I'd categorize that as a failure. It indicates an inattentiveness to the obvious errors of effectively dropping minor league seasons out of the comps and comparing minor leaguers to major leaguers only (and then, if your answer is Wayne Causey, doing it wrong.)
3. These are updated PECOTA's; I expect that the books are going to have different ones. I was part of the Beta Testing, and I think my comments thereto are part of my public comment log.
4. The claims that this PECOTA is (roughly) better than everything in history have been made the last two years. For any comparisons (and one to ZiPS seems warranted), transparancy will be vital.
5. For pity's sake, implying that these are iterative changes to the already-Deadly Accurate PECOTA's is worrying. Authors cited the projections for Kila and Bowker repeatedly and BP never said, "These were clearly wrong."
6. There are some weird projections. Chipper PECOTA: 281/375/450. Chipper ZiPS: 260/348/438. We'll see if this is right. But I doubt it: Weightings for the very young or very old should trend more strongly to more recent seasons. (See: Silver, Nate, Rearranging PECOTA, in the 2006 annual.)
7. Speaking of 2006 conclusions:
A. Level adjustment for minor league players by age. Since comps are not selected from minor leaguers, I doubt this is done. Also, you should not comp more than one level away by age, according to Silver.
B. Starter/Reliever adjustment. Chris Sale will lose velocity and K rate as a starter.
Crap. Went long. Gotta run.
Yeah, we obviously have incomplete facts, but if indeed it was Hamilton who called for help, that's a very positive sign. A one-off relapse (though obviously not Hamilton's first) with desire not to relapse further may be resolvable.
Hamilton's long and improbable road back from the gutter has had its share of bumps. And I think there is a place for social excoriation if such deters either Hamilton or others from the addiction lifestyle - addicts tend strongly toward behaviors that are harmful to society.
Hamilton's less likely to engage in some of those behaviors because he's got money and a very vested support structure.
And, of course, that's just my opinion. I could be wrong.
I wonder who has the highest total? Highest average? (This one probably won't help me.) But it's kind of a fun feature.
History is not a repeatable experiment, so I understand there are limits to the methodology. Still, to take one aspect of this, are we asking all pitchers to slow down their delivery? Hey, Justin Verlander, you're throwing too hard? What evidence is there that guys who throw 98 get hurt more than guys who throw 92? If not everyone, how do we decide which guys to slow down? And slowing down pitches impacts performance negatively.
The Rangers seek very hard throwers and run them hard. Are they wrong to do so?
How much more time between appearances do we need? Isn't it an easy post hoc explanation to say "strength and conditioning" when someone gets hurt?
Can't you have shoulder or elbow inflammation without improper mechanics, or diagnosable fatigue or weakness? I'm no athlete, but I hurt my shoulder doing approximately nothing. An explanation which explains everything explains nothing; we can ascribe all of these injuries to something after they happen.
If you can predict future injuries with good accuracy, then I withdraw all of this and I am duly impressed. But if you're saying that all these pitcher injuries are preventable if only the teams weren't so dumb, I remain skeptical.
"A large component of preventability"?
This strikes me as argument by assertion. You assert that a variety of maladies have a large component of preventability, but it certainly appears that most organizations are trying to prevent them, and pitchers still get hurt.
I think the reason more pitchers get hurt is that throwing the ball that much or that frequently with that velocity is bad for you. I disbelieve that the majority of the pitching injuries are preventable. Well, I suppose almost all of them are preventable - Adam Wainwright wouldn't have hurt his arm if he were selling insurance rather than pitching.
If these were preventable, I'd like to see a chart of actual pitchers and what could have been done to prevent these injuries. Without that, your argument is raw assertion. Sure, Mark Prior and Kerry Wood had preventable injuries. But tons of guys break when they are being handled quite carefully.
I'm with everyone. I noted after one of your articles that it should have an "Awesome" tag so people don't miss it - later articles showed that "Mike Fast" was a good proxy for "Awesome."
Your work genuinely advanced the field, and if only we lived in a proper sort of society, you'd be obliged to continue here; from each according to his abilities and all.
Sadly, Kim Jong Il is dead and you're moving on. Well, at least one of those is sad. For us. But not for you. Well deserved, Mike Fast. The 'Stros have a very nice park in a highly populated area and a barren organization; I wish you the best in helping the Astros achieve competence.
Maybe we can have you erased under SOPA. Your articles sometimes talk about Pirates.
It's unfortunate that you've gotten this so wrong. Yeah, you've made a lot of arguments, but still....
Writers who cite fancy-dan stats like OBP merely clog the internet.
Reviewing past performance by writers removes their ability to show the human element of the game. When Bill Plaschke said that only six MLB teams even had a chance to get into the playoffs, we could get all technical and county or we could accept the human condition, as held by Plaschke, players... everyone but Ken Funck.
And let's face it - Ken's not the most dominant sportswriter of his era. When you've got Jay Mariotti, Plaschke, and Bill Conlin around, Ken can't compete, so he relies on soul-killing assertions like "Home runs are basically always good," when it's well known that even a triple will do.
A lot of people look at Mike Barnicle and Ron Borges and criticize their work as "empty" or "plagiarized" or "not actually factually correct." But you didn't watch them write their columns, so you just don't know.
Finally, look at all of these other writers, who are getting paid much more than Ken Funck. It's a capitalist system; the people that do the best earn the most. Ken's not earning, so it's time to just start burning his material.
Either you misread my argument, or I miswrote it.
The net evidence for Bagwell being on steroids is much greater (and less susceptible to proof otherwise) than the totality of the evidence that Obama is not a natural-born citizen. The speculation based on the initial lack of a long-form birth certificate is some evidence. There is lots of evidence regarding the birther claim, and the evidence that Barack Obama was born in Hawaii is overwhelming. This does not mean there is not contrary evidence.
Bagwell (or Martinez or Jeter) cannot generate that kind of evidence of non-steroid use. These situations aren't comparable in that way. But even if they were, I didn't say what you ascribe to me. (I think it's me, anyway. If I've misfollowed the track of this conversation, I'm sorry for that.)
I never, ever said Bagwell should be out of the Hall based on that evidence, and observed that using such sorts of evidence to make decisions about what should happen to people creates some fairly serious issues. The conclusion that I did say that, I'd submit, seems based on a theory that anyone who says such speculation is evidence must necessarily be a steroid hawk; this theory is mistaken. (I would vote for Jeff Bagwell for the HoF; I do believe Bagwell likely took steroids, and I do not believe it likely that Edgar did, but my estimates are soft.)
I think you're mistaken on this. Argument from authority may be weak evidence, but it often is evidence. (To be sure, it also used fallaciously.)
On another forum, I argued about what California's Three Strikes law actually is. By any definition, I am an expert in this field. You should take my word as having more value than a New York lawyer's, or a California lawyer without expertise in criminal law, because I have actual expertise in this subject.
If we are talking about the relative values of Ron Santo and Ron Cey in this forum, the argument from authority would be somewhere close to worthless; most of us have some expertise. But if we're talking about actuarial studies, I'd take the word of an actuary.
I know very little about Edgar Martinez personally. Someone who knows almost anything about him knows more than I do. I assume that Rob Neyer does.
Let's define our terms. "Evidence" is some indication that something is more or less likely to be true.
Low-value evidence is still evidence. I had no opinion on the steroid use of either Edgar or Derek before this discussion. Because Rob Neyer suspects Edgar more than Derek, and Rob Neyer is more likely to know something I don't know, I now ascribe a higher likelihood to Edgar than Derek as a potential steroid-taker.
Now, that evidence is very weak, to be sure. But it is some evidence.
On other issues: I believe the weights used to say that steroid users didn't get bigger are worthless; those are their listed weights and they did not rise with muscle gain.
I assume the subtraction difference in shortstop HR/PA is due to rounding. It certainly appears that the home run rate relative to baseline was much greater for shortstops.
As a general rule. I believe that Rob's making differentiations or proposing people should care about a "scintilla" of difference is misguided; if we say that Jeter had a 6.3% chance of taking steroids and Edgar had a 6.4% chance, that doesn't support making differentiations in treatment except at the most severe margins. There's also a good argument that we should not use such wispy evidence against people when they have no control over such evidence.
In short, I think Rob's use of the term "evidence" appears adequate, if not particularly helpful. The fact that I went to business school is some evidence tending to show I am more likely to be an accountant than the average human. I am not an accountant; my assertion of that (alleged) fact is quite a bit stronger evidence against the theory that I am an accountant.
OK, I've found some comps. I don't know if I missed them before, or if they're new.
Bryce Harper's first comp is Roger Marquis, who is kind of famous for his single, unsuccessful plate appearance in the majors. B-ref shows his age-19 season as hitting .227 with a slugging percentage of .315, with unknown walks. His age-18 season was the single PA previously mentioned.
Nelson Mathews is the second comp; he hit .294 with a .419 slugging in his age-18 season, mostly in B-level ball. (This was when there was D, C, B, A, AA, and AAA. The low-level leagues survived on selling games to locals.)
The third comp is, amazingly, John Paciorek. Paciorek hit a productive .219 for Modesto in his age-18 season. If you're using his age 19 season, he hit .135.
Jesus Montero's first comp is Marc Hill. That's not possible.
Machado comps to Tom Carroll (presumably Tom E. Carroll) and there's the same problem there.
Mike Trout projects to hit .204/.261/.281. (ZiPS says .267/.338/.414.)
If you're right about all these things, I congratulate you. But I don't think you are.
Loved the article; the writeups had me genuinely laughing. (The Colon one is priceless.)
The reason that Curt Schilling was a comp to Al Alburquerque is that the comps were broken; the stated intent to compare minor leaguers to minor leaguers didn't happen, and this led to bizarre comps like Alburquerque to Schilling. (Or Rich Poythress to People A Lot Better Than Rich Poythress.)
Helpful suggestion to BP writers: Don't rely on the 2011 comps to minor league players. At all.
I resolve to crush my enemies like the insects they are. To my leaguemates: Hate you guys. So much.
I resolve to find a way to fit the 15-pound travelling trophy around my neck so I can wear it about town.
I resolve to plug Scoresheet to dumber owners, because the smarter ones are catching up to me.
I resolve to be ready for my fantasy drafts, including football. (I misprioritized work over the draft.)
[Actual conversation about this between boss and I:
JRM: This [month-long work project] has deeply interfered with my fantasy draft. I am not ready at all, and I would like to file a formal complaint.
Boss: It seems that there's a very simple remedy for this problem.
JRM: I do not think I like your proposed remedy.]
I resolve to stop trading prospects for too little.
I resolve not to reveal my secret strategies to the aforementioned insects any more than I already have.
I resolve to be only slightly less of a small-sample-size-sucker.
I resolve to root for Jose Altuve even if he's not on my team.
I resolve to win money playing poker in Vegas before the draft.
More on aging:
There are some very curious projections. Torii Hunter's got a brutal projection. It might be that Chipper and Helton (who also got a cheery estimate) are outliers.
Where Chipper is projected to do better than he has in the last three years, Hunter's expected to be significantly worse than he was over the last three years, inclusive of 2011, which was his worst year. Chipper's expected to have an OPS 19 points higher than his best year of the last three.
If there's a good reason for this, that's great. (Genuinely great: If this diffentiation is correct, PECOTA has an advantage over other prognostication algorithms of which I am aware.) I can't imagine this is entirely due to using data from four years ago, but you know the system better than I do.
My most serious concerns last year were:
1. Plainly mistaken projections for older minor leaguers.
2. Very flat aging curves.
3. Baffling comps, which were not done the way they were claimed.
4. Projections for people who were out entire seasons were wildly optimistic.
5. A wide variety of technical errors, reducing confidence in the projections considerably.
Let's take a look:
1. I called this the "Kila-Bowker problem." Nate knew about this problem, and adjusted for it in later versions of PECOTA, harshly degrading projections of level-repeaters. Kila didn't get a projection in this Beta, but Bowker did and while it lacks the spectacular nature of last year's projection, it's still high.
Matthew Brown is a good example of this - he's projected to hit 239/299/405 for the Twins; I suspect he's out of baseball after OPS'ing 650 at Rochester in limited time last year at the age of 28. It's my guess that something significant has been done about Kila-Bowker, but not quite enough.
2. The aging curves appear to me to be too flat. Chipper's projected at 285/382/455. Chipper has hit below .285 in each of the last three years. Is PECOTA going back four years? I believe earlier versions only used three years. Either it is going back four years, or it is permitting hyperoptimistic views of older players, or both.
3. There's no way to tell if the comp issue has been fixed that I could find. Comps for prospects and minor leaguers of any ilk were seriously befouled.
4. There are multiple projections for people who were gone last year (Lowell, Bradley, Glaus, off the top) and while there will be no way to demonstrate this since they're permanently gone, the projections look optimistic to me.
5. I didn't look at everything. I have hopes that this early start will lead to fewer pitchforks and torches.
I am, to be sure, pleased that you are crowdsourcing this early, because the process the last two years didn't work. I appreciate the opportunity to help. It is my firm hope that BP remembers the lessons of its past; BP put out articles using PECOTA's projections on Kila and Bowker without once saying, "These projections are obviously nonsense." That is not good.
BP has also claimed that the new PECOTA is better than Nate PECOTA. Perhaps this is so, but when the comps don't work right, the level-repeating adjustment is gone, and PECOTA is losing to Marcel rather than beating it, it causes doubts. This iteration of PECOTA appears to have some improvements, but it also raises some questions.
I hope this has been helpful.
OK, Part 2, Neyer's Foreward:
Comma after "published" in line two of the forward should be excised, says me.
In line five of the forward, I think the "There" after the ellipsis should not be capitalized, or it should be four-dot ellipsis.
The last two paragraphs should be punctuated in parallel; I suggest a question mark after "Arrogance."
Good luck, guys. Hope this sells well.
Thanks for access! Cool beans!
Quick editing notes from the first 13 pages or so (up to the first pre-existing article.)
Thank You: BP uses serial commas throughout; should be a comma after "Goble."
Foreward: I'd use a dash in paragraph two rather than the comma after "bases," but King might not like that.
Irregular use of umlauts in "Voros," should be standardized (Forward - no umlauts. Table of contents - umlauts.)
From Jaffe's opening on offense, I think "mid-Nineties" should be "mid-nineties."
That's all I have, which isn't exactly adequate payment for free access. Thanks again.
"But that sourced material is often used as a shield against outcomes..."
This assessment is full of win. It could not have been only me that echoed back to a former BP'er who did exactly this. It also applies to every fantasy owner who takes credit for his own wins and blames RotoSuperGeniusGreatness for his failures.
I know it's not everyone's cup of tea, but I think the meta-analysis - whether by design or not - pushes readers to be smarter. I find these presentations charming.
As a side note, I appreciate the acknowledgement that outcomes matter. If you stick to a #6org theory where you don't even care what the outcome is, then it's a little hard to change the paradigm that got you there. That's not to say good decisions can't end very badly through nothing but bad luck, but outcomes just have to matter.
RA: As a Republican, Dodger fan, and lawyer with some expertise on what constitutional rights actually are, I say to you this: Don't feed the trolls.
I think the (effective) hard-slotting is good for the small-market teams. Signability becomes a non-issue.
It occurs to me that there's a way to game the system if you wanted to pay a high draft pick more; you could draft people you don't want and use that money on another player. Still, since you'll get to sign everyone if you don't screw it up, I don't see this happening on any kind of regular basis.
I think the CBA is good for the fan experience of baseball.
Not that it was impossible to find the reasoning?
There is absolutely no reason whatsoever for this disgraceful result. I challenge anyone to give me a single reason why Nova is a better choice for Rookie of the Year than Pineda. Just one.
Further commentary seems to indicate (via your reliance on FIP) that you don't believe pitchers have any control over BABIP, which is a provably false statement. If you're going to FIP, you are ignoring a real skill variance. Granted, that variance is often obliterated by luck in a single season, but there's still a skill component.
The reason this sort of argument irritates me so is that it can only set back stathead arguments. We, in fact, have come an enormous way. Park effects matter. Defense matters. Wins matter very little. These principles are making progress.
But when you argue that just because Nova had better results than Pineda is not a reason to vote Nova ahead of Pineda.... that's alarming. FIP has known flaws; shouldn't we agree that those are flaws? Fangraphs' WAR for pitchers really crushes the value of guys like Mariano who routinely outperform their peripherals, because it bases WAR on an ERA simulator.
Let's not take a bite out of a writers' group that I have to believe would not have made some of the Burroughsian mistakes of the past, unless we have a real argument to make. I don't think the heat shed by this article brings much light to the proceedings. We don't want to be an insular little club scoffing at the stupid; that's one of the fastest ways to get stupid.
I think Nova was better than Pineda. I think Kershaw and Halladay were better than Lee, even if not by much.
But that's just my opinion. I could be wrong.
Hey, Aumann's Agreement Theorem strikes at BP! Cool.
To be clear, rational actors with the same knowledge should always agree. Differing knowledge may lead to non-unanimity. But the general point stands well.
Cliff Lee and Roy Halladay played for the same team. I don't know that they faced equivalent teams, but it strikes me that the reason Lee may have faced slightly stronger players is because he's left-handed, which is a quality of him, not of luck.
You leaned very hard into the voters because no one voted for the guy with the lowest ZpHX or xFIP or something, but if you do that, you'd best be taking everything into account yourself.
Did you account for this handedness issue? It doesn't appear that you did. And if the platoonmates vs. lefties were stronger, it's very hard to give Lee a bonus for that.
I think that the general argument that A and B are close so A should get about as many votes as B is flawed at the outset. If we're voting for Most Giraffes and Steve has 198 and Joe has 199, we're all going to vote for Joe. (Christina Kahrl made the a similar baffling argument re: Votto's MVP, which I understood to say that Votto probably deserved it but so many people voting for him was a crime against math.)
I'm far from a supporter of the voters; I'm still mad Alan Trammell didn't win the MVP when he obviously deserved it. But the tenor of the article is unjustified by the actions of the voters.
Plus, over enough time (way more than one year), ERA is a better indicator of skill than the estimators. (See: Rivera, M.) It's also an actual result.
Nova's B-R WAR is 3.6, while Pineda's was 2.8. That's a huge difference. That's "one reason." If you look at the actual run prevention results - the thing that helped their team win games - Nova was just better.
Is Pineda the better bet going forward? Absolutely. But Michael's argument is completely sound, and while you can disagree with it reasonably, it's unjustifiable to paint the other side as dumb and crazy.
I'd really appreciate the functionality R.A.Wagman suggests. That sort of focus would be a great option, else I won't likely use it often (if at all).
Thanks. This is very helpful.
Could you expand on your statement that OPS+ undervalues OBP? OPS+ values OBP at about 1.25 slugging this year. What's the approximate ratio that your stat uses?
As I read the start of the article, I was concerned with the comparison pool and with the ability to show causation. I find that often - too often - analysts come up with statistical conclusions without looking at possible artifacts.
By the end, I was not concerned. I join the chorus of "Awesome!" that this article richly deserves. Might I suggest that those articles which clearly and substantially move forward analysis on a key point get placed in a special section? Two weeks or four years from now, some interested person might not know how to find this, and this deserves to be found. This reminds me of some of the best long analytical articles from the earlier books.
There's clearly more to go on this issue, and there's been great work by others indicating an effect. But this robust effort seems to me to have significantly moved knowledge forward. And the video clips are fantastic.
(One aside: I think one of the possible statistical artifacts is that umpires call different pitches strikes on different counts; if you had a team with a lot of 3-0 counts, you'll end up with more fake strikes called. However, measuring for that is likely to befoul your sample sizes sufficiently as to make the noise overwhelm the likely very slight effect. But that's just my opinion. I could be wrong.)
I must be doing it wrong.
For an example of the old format at Cot's:
The new format either eliminates a lot of information I want (like an overview of the entire contract for the player) or has lost me. Additionally, many players have no information about their contract status other than current salary (and this includes many big names.)
Is there a way currently that I can find out from this Carl Crawford's contract terms without going through multiple pages? From my point of view, the original format was much more readable and had much more information.
Hope this helps.
I almost skipped this. Glad I didn't. The only thing it needed was an "awesome" tag to avoid the potential of missing it. Well done!
Whee! Super-fun article.
The new Upside numbers look much improved, and most of the general curves of the Upside numbers look better than the long mid-career flatline of the 10-year projections.
1. Those who didn't play in 2010 get their first year as 2010. Sometimes, those who did play get their first projection year as 2010.
2. The Upside scores are befouled.
2a. (Discovered midway through) Some of the projections for the 10-year don't match the one-year for 2011. Hmm. Looks like many of them, actually. If it were just a park adjustment thing, VORP shouldn't change, but it does.
3. For young hitters, the years from age 26-30 are all very, very similar, no matter the player type (Jesus Montero and Starlin Castro have the same situation.) This makes it appear that the comp system isn't being used in favor of a regular aging curver. And that aging curve is often uncurvy; Michael Brantley's year-by-year projections are very nearly the same for every year, forever.
4. Old hitters decline very slowly, much more slowly than in real life. Derek Jeter being useful at 44 is possible. OCab getting 500 plate appearances at the age of 42 isn't. Lance Berkman being pretty good at the age of 43 isn't likely to happen. The projections indicate that age-40+ people will proliferate to a degree never before seen in MLB.
5. Halladay (for one) has a pronounced decay curve. I've done some work on pitcher projections, and I'm convinced that's too strong. A pitcher of Halladay quality will typically not suffer the type of age decay that hits the ordinary.
6. Felix's projection for 2016 has him with 152K 59BB and a 4.09 ERA. I'm therefore assuming the 10-years have no park factors; this should be explicit.
That's what I got. Hope it helps.
--JRM, and I think this is what we call the alpha stage.
My comment above says or seems to say there are no 10-year projections now; that's wrong. My apologies.
Bayesian priors are in my wheelhouse. I am absolutely convinced that's the right way to do it. All good there.
But this is from the July 19 thread:
Rest-of-season PECOTA: We've made changes to the weighting of recent results that reduced the impact of 2009-2010 performance on players' rest-of-season projections, making them less susceptible to being swayed by small-sample stats. Those changes are now reflected in the "2011 Projections" table at the top of each player card, as well as in our Playoff Odds.
And several people noted that you'd previously said that Fangraphs was wrong for its too-2011 heavy stats. While the paragraph was parsed different ways (and we never found out which way was right - the logically coherent and grammatically coherent readings appear at odds), it appears to say that prior years are having their weighting reduced, to get closer to Fangraphs' weightings. (Yes, Fangraphs is doing something else that appears mistaken, but you vigorously stressed the goal of not overselling recent performance.)
Step 1. Fangraphs is wrong! They're weighting 2011 way, way too much! Dummies.
Step 2. The approximate weighting Fangraphs is using for 2011 (for players with similar PT rates in each year) looks right to some of the unwashed masses.
Step 3. We've changed our system to weight prior years less (and, by necessity, 2011 more.)
Either Step 1 or Step 3 has a flaw. What I want to know is if I've misread Step 3.
To my way of thinking, your criticism here is well-explained, and doesn't feel personal. The article did not succeed on those terms, at least to me.
This almost exactly reflects my views. The 10-year upside projections - the only 10-year projections - were immediately identified as broken, and we got promised that it would be looked at. That got nowhere. BP isn't announcing that it's taking down the broken PECOTA's, or broken 10-year upside.
There are other problems I discussed in the last thread that I won't harp on further.
Unlike evo34, I have a dog in this fight; I've published twice on BP, I've been acknowledged in the book credits, and I've had a number of pleasant conversations with BP authors both live and via e-mail. I want BP to succeed.
But an article saying, "Math is hard; incremental improvement isn't useful," and a somewhat opaque statistical attack is either anti-stathead or appearing to stem from personal animosity.
Steven Goldman deserves some heat for this, too. There's a time when an editor needs to slow down the horses.
No catches, no one adjacent to me catching one, majors or minors. I'm 45, and have probably attended a little over 100 games in person.
Great article, though. I think you've captured the spirit of getting a ball, and that's really fun.
What's the relative quality (if you know) of the complex leagues vs. the Venezualan Summer League? I'm curious about Oscar Hernandez' thoroughly insane performance in the VSL, and how much meaning that has.
I am, even if it's disallowed.
I agree that open disputes are often healthy (like this one!). And I concur with the methodological point you make; however, I expect that the more 2011-centric ZiPS projections will be more often closer from the original date/RoS PECOTA projections. Maybe I'll turn out to be wrong.
I am certainly pleased that you'll revisit this - I'd love to see an article at the end of the year to see how RoS PECOTA fared against RoS ZiPS from the inception of RoS PECOTA.
I'd very much like to see an article early in the offseason (or earlier) about the other flagged PECOTA issues. I view them as serious.
Thanks again for your response.
Charming! Thank you.
That seems an interesting (and counter-intuitive) reading, but is at least a potential explanation for the baffling language. Further reducing current performance seems misguided to me given Tango's writeups on the subject, but it's a possibility.
I am a huge longtime fan of BP. I'm very frustrated. Y'all are better than this. Or maybe not.
I appreciate the general admission of errors, but there are some specific unacknowledged errors that ought to be addressed.
1. What makewayhomer said.
1a. You started the RoS PECOTA by announcing that "Fangraphs is wrong." Now you've apparently gone toward Fangraphs. My quick calculations indicated Fangraphs' weighting was pretty good; rather more tellingly, Tom Tango's did too. I'm pretty sure an apology is in order; if you're going to wrongly disrespect competitors, you owe them an apology when you're wrong.
This is really bad form not to do that.
2. If you're going to talk about PECOTA problems and not address the Kila/Bowker problem - which was obvious and mentioned at the time by more than one person - that seems unfortunate. Some indication that you're trying to repair the error would be nice.
3. If you're going to talk about PECOTA problems and not talk about the Trout problem, that seems unfortunate. Some indication that you're trying to repair the error would be nice. The continued insistence at the time that the comp list problems for those with recent minor league histories weren't problems was very annoying, and remains so. One of PECOTA's great assets has left the building and there's no indication that it's coming back.
4. There's an ongoing bug in the A's in the Playoff Odds for at least a week. The Rockies are similarly problematic due to a Mark Ellis bug. Someone should look at these intermittently.
5. The way Nate did the Playoff Odds, if I understood it correctly, was to assume team quality around a baseline; the Monte Carlo sim used a distribution of quality assumptions vs. the remaining schedule. (That is, if we assume the A's are a .642 team the rest of the way, the sim might play them out as .655 or .617, rather than just as .642.) The playoff odds report appears to neither assume a distribution of potential goodness nor does it appear to take schedule into account. I could be wrong about this.
Short version: Grownups acknowledge error. I get that you want to sell the product, but there are some of us in the unwashed masses who aren't buying diet pills from Steve Garvey, penis extensions from Jimmy Johnson, and Fangraphs is Wrong from BP. Admit specific screwups.
I want to love you guys. And there are lots of good articles. And it makes me happy that Team Tracker's improved in some details (if not in daily reliability.) But you're making it hard.
Something has gone seriously awry. Per this, A's are a .642 team as of today, and Mark Ellis' thousands of PA's will be a big help.
I don't think Rose's number should be retired. Retirement should mean something.
Of course, as also pointed out, Steve "Too Stupid To Have Committed Fraud" Garvey managed to get his number retired. (Try Googling Enforma, if you don't remember that episode.)
I tried to reply to the long list above (well-researched!) because Cesar Cedeno was found to have accidentally shot a woman to death. Given Cedeno's other adventures, that's a particularly Cedeno-favorable view of the facts.
And I agree with the comment that, generally, fewer numbers should be retired. I had no idea whose numbers were retired in Houston until I was at Minute Maid last week.... shocking, really. Not Garvey-shocking, but pretty shocking.
Pick 7 is for a pick Arizona didn't sign last year. There's now a rule that if you don't sign a guy from the prior year, you get a pick one notch down the following year.
However, you can't keep playing this game; the compensation picks for unsigned draft picks can't then go unsigned and result in yet another bite at the apple the following year. Hope that helps.
Is that referencing the start of the season? Would he rank higher now?
(BTW, like everyone else, I love this feature.)
I regret the effect my actions had on my teammates, family, and fans. And, of course, those kittens.
I was at a party, and someone spiked the punched with androstenedione.
I know that I will be exonerated. This is a difficult period for me and my family, and I'm sure you'll all understand that further comment will come from my lawyer.
I was quoted out of context.
Yeah, I said it. I'm Brian F-king Cashman. What the hell else are they going to do to me?
Sky, I'm way late on this, but I'm confident that increased velocity decreases BABIP for right-handers that throw 90+. It probably decreases BABIP for everyone, but there's a ton of noise in the data.
Could you mark articles like this with an "Awesome" tag? It might prevent people from missing them.
I liked the article, but I think you've made two errors here in trusting BP's metrics:
1. The Playoff Odds Report does not currently change the expected win rate based on past performance. Estimating the Royals as exactly the same quality as was estimated at the start of the season seems misguided.
2. PECOTA was and is just wrong on Kila. The projection was broken from the start, and there's a specific reason: The way comps are done is especially problematic for older minor leaguers. (Kila and John Bowker were the clearest versions of this.) Apart from that, level-repeaters are known to need adjustments. The current iteration of PECOTA has principles for older minor leaguers which are known to be quite unsound.
I have a pretty good idea; it seems apparent. The Indians expected winning percentage has been at either .462 or .461 for as long as I've been tracking it. Which one of those it is doesn't appear to depend on the day before's results. I suspect it's due to rounding issues with the projections.
Thus, while I may be wrong, it seems extremely likely that the amount that actual stats are being used is, "None." There might be adjustments to the depth charts, but not to the projections and there's no nod to prior performance.
BP has lots of goodness in it still, but the performance on PECOTA/future projections. over the last couple of years has taken backwards steps. Some concession that the critics of PECOTA's Kila-Bowker problem were right might be a start. Or you could acknowledge evo34's point here, by withdrawing the claim that current performance is used in a sensible (or any) manner.
I'm not Colin, but take a look at:
Let's take away, top of the eighth, no one on. Going from down two to down three is a loss of about one-twentieth of a win. Going from up two to up one is a loss of about one-eighth of a win.
You're in much better leveraging shape if you're up a couple than down a couple. Teams are right to treat them differently.
They're on pace to throw Walden until his arm falls off, but I think you're dead wrong on Walden's potential excellence or lack thereof. Dudes who throw this hard with solid control are where the great ones come from.
I have very high expectations for him, myself, if he can withstand the (so far) unreal workload.
A story that perhaps doesn't belong here, but certainly belongs nowhere else:
Some years ago, I got into a friendly e-mail discussion with Christina about diacritical marks. She likes accents, umlauts, tildes and other markings from other languages. I hate them.
She accused me (in all good humor) of being her arch-enemy, a comic-book style villain opposing all that is good, or at least all that is Christina.
(An umlaut looks a horizontal semi-colon, and a tilde is ~.
So, we're talking about a package involving marks that look like :~.)
Christina argued (approximately) that we have a long-standing tradition involving this package, and that since this is what we started with, we should keep it no matter what changes in typefaces, use of computers, or other societal issues occur.
I averred that we had moved past 1955 into a new and different era, and that those who wanted to shed this unnecessary junk should be saluted. At the very least, the disapprobation of those who choose to embrace the new culture was deeply misguided.
There is no longer a credible argument to hold onto the old ways when such a hold makes life harder for so many. I don't want to type Alt-0223; I just want to get rid of things that look wrong and feel wrong and replace them with an appropriately cleaner look.
I did not convince her.
Still, Christina, from your arch-enemy: Congratulations and good luck. Your writing has been a joy these many years at BP.
--JRM, who concedes that some of the exact language in this comment was not present in our actual conversation, though the punctuation content is accurate to my recollection.
It's not just Trout, though Trout looks to be the most severe case. Wil Myers, who has great comps, has terrible upside ratings, too.
Colin said more than once that minor league players would get more minor league comps, and instead the use of only MLB comps for MiLB players has appeared to foul up the projections short- and long-term for young and old. (John Bowker's first player card comp is Will Clark, and his long-term upside exceeds all actual prospects, I think. His 10-year upside, for instance, exceeds that of Trout, Montero, Freeman, and Hosmer combined. By a lot.)
You've made some helpful statements on these points (as here) and in my view it's important that you continue to concede the errors that have been baked in this particular cake. Fixing this for 2012 ought to be prioritized.
Check out Mike Trout's upside. The calculation of the metric is quite wrong, or the projections are seriously wrong.
I'll go out on a limb and say that Mike Trout's 10-year upside exceeds that of Orlando Cabrera, even if Pecota disagrees.
Kevin, I think you mean to say Trout is 1-5 (83%) rather than 6-5 (44%) to be the top prospect. This has been commented on before, and you and the editors might want to become a little more familiar with how odds are displayed.
Sure! It's just that the article above appeared to me to read differently than this response. That's what confused me, because I had understood that replacement level would not be Player AAA. Glad to see it isn't.
Thanks for the work. I'm having a few issues.
1. Chris Snyder is ahead of Mike Napoli.
2. If I read this right, you're using player AAA as replacement level. Like the others, I view this as misguided.
3. Felix is expected to allow 66 total earned runs per the spreadsheet. Replacement AAA is 1.8 (league ERA) or roughly 7.30. 230 IP at 7.30 ERA is about 186 runs. That's a 120-run difference, but Felix is only listed at a value of 88. What have I calculated incorrectly?
4. Pujols is listed as 75 runs above Player AAA. That number is also quite a bit lower than I calculate.
I believe you're significantly mistaken on this. ERA controls over WHIP, and hard.
The Vogon will do things to you that will make you wish you'd never been born, or, if you're a clearer minded thinker, that the Vogon had never been born.
"Mike Ivie is a $40 million airport with a $30 control tower." - Rick Monday.
But some people are phased out due to ineffectiveness, not injury, right? If a guy's a 1B who goes from a 970 to an 815 OPS from his age 31 to age 38 season, we figure that's a likely decline on this methodology. But if he goes from 970 to 715, he's not playing any more, so that's not counted.
I'm not saying something like this for Prince Albert isn't possible; guys of this quality often age well. But I think that measurement isn't the right one, leading to the funky projected rise in numbers in Albert's late 30's.
I am pleased that the people who give negative ratings do not commonly explain why. If they did, it'd take up a lot of space.
If you're comment's at a -10 and you can't figure out why, go ahead and ask (says me). If it's one comment at -1 or -2, don't sweat it. If it's a bunch of comments at -1 or -2, look for patterns.
I like the rating system (I think rating systems are mild to moderate troll deterrents). I've certainly had comments I considered to add to the equation get negative reviews. C'est la mort.
I understand this. But the comps are supposed to be comparable players under the theory that comparable players will age comparably. Uncomparable players will not age comparably. If you think a guy mashing some in High Desert at the age of 22 is going to follow the same route as a guy mashing in the bigs at age 22, I'd like to see some authority for that.
This is also a massive change to the comp system that built the Pecota brand; the comp system was based on actual comparable players, rather than people mountains better than the player being evaluated.
If the comps are nutty (and indeed they are) for minor leaguers, the shape is based not on actual comps but on a semi-random assortment of players.
Further, if the comps are no good, the percentile rankings lose substantial value.
Further further, it seems clear that these errors have made for bad projections for older minor leaguers - Kila, Dubois, Bowker, and others have wildly optimistic projections. If you think I'm wrong, I'll make a bet (cash to charity of winner's choice) and give you 7-5 on the over on all of the aging minor leaguers.
That's not it. I've been looking at these comp lists for years with some care, and I'd bet Toyotas to Tonkas there are more HOF or active tremendous players in the comps. Four Mantle comps:
Mantle's age-20 season: 311/394/530, OPS+ of 161. In the big leagues.
Mickey Mantle should not be one of the top five hundred comps for these guys. Before, if you had Mantle or Kaline or Yount in your comps, you'd have an actual chance - if slight - of being similar to Mantle or Kaline or Yount. Rymer Liriano isn't going to be Mickey Mantle unless he gets bitten by a radioactive spider or something similar.
Dave Winfield is not a good comp for Calvin Anderson. Period. There is a problem, and it's a serious one, and while reboots are hard (especially in this case), someone should have caught this before publication. It took me twenty minutes to find the pattern and twenty more to confirm.
--JRM, welcoming our new Roberto Clemente, currently known as Jay Austin.
Agree with Dan and Ragnhild. I think P's minor league comps are broken. I'm prepared to give 7-5 and take the under on Kila and DJohnson and several other guys who are late-career minor leaguers (Hello, Jason Dubois.)
There are some pretty severe problems with the comps for minor league players.
Kila Ka'aihue: Nick Johnson, Joey Votto, Adrian Gonzalez.
Dan Johnson: Jason Giambi, Nick Johnson, Paul Konerko
Chris Carter: Chris Davis, Joey Votto, Evan Longoria.
(Note that Billy Butler's comps are Conor Jackson, Dan Johnson and Gaby Sanchez.)
So far, a problem. But then:
Rich Poythress: Adrian Gonzalez, Prince Fielder, Kent Hrbek
Seriously. I am not making that up. Next?
Clint Robinson: Adrian Gonzalez, Joey Votto, Ryan Garko.
Wil Myers is an awfully good prospect, but .256/341/410 next year?
Robinson Chirinos (comps Iannetta, Carlos Ruiz, and Todd Helton) projects at 275/360/469.
John Bowker? 268/342/457. Jason Dubois has a nice projection, and I would've guessed he was somewhere selling insurance rather than hitting pretty well in Iowa.
Gerald Sands' comps are Prince Fielder, Willie McCovey and Mark McGwire.
Jaff Decker has Willie Mays in his comps.
I think there's a systemic problem with the minor league comps. Matt Carpenter's top comp is Chipper Jones. Trent Oeltjen's comps are Carlos Beltran, Vada Pinson, and Roberto Clemente. Thomas Field's top comp is Tim Raines.
If you want to tell me these are right, it's going to take a lot of convincing. Jedd Gyorko's top two comps ought not be Eric Chavez and Steve Garvey. There are many more crazy comps out there.
Now, I know there are a lot of comps in the system, but it looks like the system has a severe bias; the system is looking for people who are one hell of a lot better than the player.
This is a return to Nate's methodology, which I believe is better than a non-rounding method.
Wow, that's an impressive breakdown.
RA, I don't think Valencia was overlooked so much as people didn't think he'd be good enough to be a top-16 pick. I don't think he's protectable in a standard format.
I'm with Jivas. There's nothing quite like this anywhere else, and it's fun. I'm Wezen-sold.
This brings back both the memories of the fun movie and of an acquaintance who sent me the box score of a game over a decade back just from my description of the game. Whee!
Maybe a tad early to worry? I was handing out pitchforks and torches last year until supplies ran out, but I strongly sense that BP was wounded by the fire and pointy parts and strived mightily to avoid a rerun, as indicated by Colin's involvement.
I'd bet heavily against a 2010 replay. Let's wait until/unless I'm proven wrong to clean the blood off the pitchforks and charge the mound, eh?
Looking forward to the spreadsheet!
Looking forward to the reload. Luck to all.
geoff and similarly situated others:
1. I walked into the draft with almost no idea how Scoresheet worked in 2003. I was competitive in 2003, made a challenge trade of draft picks at the start of 2004, and ripped off a couple of championships in a field of BPer's and other sorts. I've been a contender every year. Full disclosure: I'm pretty good at this.
1a. If you start with a bad team, Scoresheet's keeper system makes most bad teams turn-aroundable in a couple of years.
1b. You can always start with a one-year league to test-drive it. I recommend it. There are first-time discounts.
2. There are hundreds of Scoresheet Leagues. In 90+% of them, a typical BP subscriber and fantasy player is going to do just fine. There are a handful of scarier leagues out there.
Check out the Scoresheet Talk page:
Tons of helpful people there.
I got in on a fluke, and I wasn't sure I'd enjoy it. But it turned out to be a marvelous experience, not just from the game playing, but also in meeting more people who enjoy this kind of thing.
Looking forward to the coverage. Scoresheet rocks; I joined a league started by BP via draft organizer Gary "Majerus" Huckabay eight years ago, and it's been a terrific experience.
I don't know how many of my reprehensible leaguemates will read your columns and take your advice, but I'd appreciate it if you'd explain that they want to take Jeff Mathis, anyway.
[Note: Reprehensibility of leaguemates is due to their persistent efforts to defeat me. I hate them all.]
--JRM, who is pretty sure Gary is listening to his vast collection of Creed songs right now.
Or maybe the charts will come tomorrow with the AL...
Thirded. I always look forward to the big chart, and no big chart is saddening.
Kevin, you've said the bell curve is being eliminated before, but the new proposal actually reflects a bell curve.
Prospects are on the far side of the bell curve, so you *should* have fewer five-star than four-star, more three-star than four-star, more two-star than three-star, and more one-star than two-star players (though at the end they'll be off the list.)
Sickels' lists have had this pattern for years, I think - zillions of C or C+ prospects per A prospect. This is right.
Like everyone else, I'm looking forward to the Top 11's. You've done a sterling job this year; the daily updates are great (reminding me of when Jay-Del Mah did those sorts of updates, without the scouting.)
Large volume results in mistakes - and this was a mistake, not a fictionalization. It's regrettable but understandable.
But let's get the fix right: Both the original article and the first correction should have "UPDATE" as the first word, and an explanation of how those articles were wrong.
(And while we're here, the original correction has a paragraph that starts "This what," which is missing a word.)
Talking about the competition, and talking about the flaws in PECOTA is good for everyone. Colin has addressed a number of critical issues. Tango's clearly interested in furthering the field; on his own blog he often touts others' research (and sometimes critiques others' research.) It's not just a "My game is better," thing.
I have little doubt that the majority of readers aren't supremely interested in the fine details - but the ones who are are worth something to the system.
Tango's point in the comments is very well taken; I have always been curious about the failure to have bigger ranges for players with limited histories (though the issue isn't just with players with limited histories, obviously.)
As one of the pitchfork-and-torch wielders from March Madness this year, I'm glad to see the effort's being made. The proof, as ever, will be in the results (I don't subscribe to Shandler's accuracy-is-irrelevant theory).
I look forward to the remaining articles. As ever, the proof will be in the results - I think PECOTA, once in front, now has a little catching up to do, and I think CHONE is going to try to stay ahead - but this should be fun.
This is tremendous. You should be nominated for a Hucky. (Or whatever the baseball scribbling award is, or should be.) Heck, I'm nominating you for a Hucky now. Consider yourself nominated.
--JRM, Official Hucky Nomination Committee Chair
"The Hucky: Since five minutes ago, the finest online baseball award."
Huh. I think he's been wrong a lot. Let's see...
1. The Rose debacle.
2. Said he sources said Strasburg would see Andrews, then said sources didn't say that, but said he would see a specialist whom he assumed to be Andrews.
3. Implication of [famous player] on steroids, then professed surprise that this would stir people up.
4. Beltran's career over, then professed surprise that this would stir people up.
5. Claim of finding Verducci effect prior to Verducci, then when pinned on numbers, slight retreat, but continues to make claim.
6. Claim (now recanted) of being able to pick off inevitably to-be-injured pitchers.
7. Claim that multiple teams have hard 30-pitch-in-a-single-inning limit, which appears incorrect (or perhaps teams that have this policy don't follow it.)
That's off the top of my head, and doesn't include more simple, mundane misevaluations, such as predictions earlier this year that Beltran would be back much more quickly than he was. Carroll is prone to remarkable assertions of various sorts.
That said, he's certainly been a popular and prolific columnist, and some explanation would be useful. But if you think he's not been wrong, and seriously so, this strikes me as somewhat inattentive to the pattern.
Things done immediately prior to disappearance:
1. Discussion of PED producers and links to MLB.
2. Statement that "sources" told him Strasburg was to see Andrews, then recantation (saying sources just said he was going to see a specialist, which he assumed was Andrews.).
Whether either of these is connected is unclear.
I'd like to know what happened, too. Carroll claims to be legally barred from talking about something that angers him, which implies an NDA.
I'd certainly not judge BP too harshly if there was a split and we got little information other than he was gone.
Carroll's a bit of a curious character, making some intriguing claims (he found Verducci effect before Verducci popularized - when pinned on numbers, backed off a bit; claiming Beltran's career was over then backing off that; the steroid implications then the disclaiming of same toward one well-known player; the claim that multiple clubs had 30-pitch limits for a single inning on all pitchers, which is researchable if you have the tools... there's more) and of course he was the primary source of BP's most embarrassingly wrong story, the Rose debacle. Of course, he's also a highly popular and prolific outputter of injury information.
Whether any of this is relevant to the instant situation is unknown by those on the outside, like me.
I have no inside information on any of this. I'd caution against confident conclusions based on limited data. I'd hope for some sort of announcement from BP. And I'd wish all parties the best.
Kevin, it's "Wil Myers," like "Wil Wheaton," not like "Will the copy editors catch this this time?"
I disagree with all of this except for loving Goldstein.
Don't power down, Christina.
When did Jason Vargas become right-handed?
Yah, Scoresheet. But I don't think you can beat Scoresheet; a non-luck balanced/non-ERAish Scoresheet preceded the existing version and led to rioting.
I don't think you can out-Scoresheet Scoresheet.
What I'd really like is a market with payoffs. That is: Each action done by a player earns money/points/whatever, or reduces earnings. At the end of the year, it pays off. That's sort of like PROTRADE, but PROTRADE didn't have a system where trades were balanced by the market, so you end up with injury runs and vast point gains through gaming the system.
I want to compare my projections to the markets. And at the end, when I'm right, I want the others' cash in my pocket. (Please note: I express no opinion on the legality of this setup in all jurisdictions. You could probably ask the guy who wrote "Fantasy Baseball and the Long Arm of the Law," but I'm guessing he's no more help than I am.)
I note that Team Tracker didn't update today; I suspect it's related to the server switch. Hoping to have it back tomorrow morning!
Thanks for the opportunity to review.
1. The "Historical Stats" section is now duplicative and unnecessary.
2. The navigation is quite unwieldy. The search box didn't really work for me at all. What I want to do is look at the 2009 Cleveland Indians, and to do that.... well, I don't know what to do. I got to the Yankees through ARod, who was on the face page, then clicking 2009.
3. The card I looked at (ARod) looked good.
4. Are you sure those weighted means on slugging percentage are right?
Glad to see they're out.
Looks like one item not caught in testing: There's a totalling problem with at least one pitcher. Check Gio Gonzalez' projection; I think 102 K is possible in 101 innings, but not so possible in 66 IP. Projecting Gio to K 14 per 9 seems like there's some adding issue.
1. You say that PECOTA ignores dropped seasons altogether, rather than factoring them in. That appears to be true for the 2009 iteration - see Jay Bruce's 2009 card for an example. But Clay said it isn't true for 2010, right? That's why we're ending up with these very strange looking lines - excellent half-time play for guys like Montero and Heyward.
2. Mike said that the comps weren't such a big factor. That's not true, is it? For most players, comps are the backbone, not a small add. Nate made it repeatedly clear that comps were the backbone.
3. Mike said PECOTA doesn't use BABIP. Nate said PECOTA uses BABIP (and pretty much any other stat that works as a leading indicator; PECOTA looks - or looked, as the case may be - "at the interaction between every statistic and every other statistic.")
Am I right, or I have I misread Nate or you or Mike?
"Modalities"? I thought those were what I had to go through to get the $46 million in the trunk in Lagos.
The 80th/90th percentile problem existed with the hitters, too. See Jake Arrieta for a more compelling example of how the percentiles are busted.
Jivas and all:
I just looked at 15 2009 cards and 15 2010 cards; I think the effect you cite is real, and substantial. I think there's something wrong there at the low-end percentiles.
Great catch, worthwhile post, and I'd love to know the reasons this is happening. A strong piece of evidence for this Beta Blog and wisdom of at least part of a crowd (or mob, as the case may be.)
I might be wrong; this might not be as pervasive as it looks like. But it looks pretty damn pervasive; I found a handful of players with similar differences this year, no one with a long 10% tail, and lots with big 90% performances. I'm guessing this stems from the same issues as the 10-year projections - out of baseball isn't considered a performance decline, and that's one of the causes of being out of baseball.
Are we handing out Great Catch Awards? One to Jivas, assuming he's right, seems in order.
I also see the same thing; Utley's weighted mean isn't right.
I see there are some Beta changes coming tonight; that's good.
On an aside, can I suggest that things like prior OBP problems and weighted mean problems be acknowledged in a public post at some point? Some note that there were systemic problems with the 2010 PECOTA's seems like it would help regain trust, not just from the pitchfork-wielding mob, but also from those who saw what was illuminated by the torches.
The step forward in opening the drawbridge to the mob in this exercise is a positive.
Good luck on tonight's update.
I would love that.
First off, I salute the effort.
Things that look wrong:
1. Some of the Weighted Means issues I've discussed earlier look fixed (Druby, for instance.) But the TAv weighted means still look way wrong (see, for instance, Nolan Reimold.) Adding the percentiles and dividing by nine should come close; it doesn't and in fact no matched pair (90+10/2, 80+20/2) comes close. (I see tbwhite points this out also.)
2. The ski-slope playing time problem in the 10-year PECOTA's still exists. Jesus Montero's not going to be a half-time player in six years hitting 300/357/519. Jason Heyward, same deal.
3. ccmonster's right.
4. Something's wrong with the EqA translations. For 2009, St. Louis reads as a pitcher's park for Holliday (Eq's: 380/437/658 on raw data of 353/426/604) but a neutral park for Glaus (172/250/241 for both). This may be yet another rounding issue due to low PA's by Glaus in St.L, but if so, it's a good reason to not round; there's especially no reason to round to generate EqA's. Nate did not round for anything; the display of homers was decorative, but the actual Excel numbers had plenty of entries like 11.975 homers. Going to a rounding system has generated errors. If you are measuring old P with the rounding, you're going to get the projected slash stats wrong.
Good luck. Off to work.
I made some comments in the depth charts. It's a little difficult to suss out why there are differences in the projections from various PECOTA sources (two spreadsheets differ, which differ from listed projection, which differ from 50th percentile.) Integer rounding explains a lot of it, but I don't think all of it. Weighted means appear to still be miscalced. Big differences (like Andy Marte's slugging) between projection and 50th percentile are undesirable.
I'd note that any comparisons to past iterations needs to include slash stats or (or preferably and) total value stats, like virtually every other forecasting comparison ever done, including by BP. Failing to include that basic comparison renders such an exercise of little value.
Asdrubal Cabrera (First guy I tried for all of them):
PECOTA projection under Major League Hitters on the spreadsheet: 288/355/409
PECOTA projection on the spreadsheet as a "Hitter" (no playing time adjustments): 288/356/414
PECOTA 50th percentile projection: 285/353/412 (I can't figure why this wouldn't match with the no playing time adjustments number.)
PECOTA Weighted Mean projection (listed): 295/348/437
PECOTA Weighted Mean projection (calculated from percentiles): 291/359/431
They've said the projections are the 50th percentile but they don't match up.
It might be an integer rounding situation; this causes some pretty severe differences (Marte's projection: .417 Slg; Marte's 50th percentile: .451.) However, it appears there's something else in play here; Brian Roberts OBP 50th percentile is .365; his projection is .371. With his PA's, you should be able to get within four points.
This was previously noted by several commenters. What is the cause of this, guys?
Thanks very much for this information. Comparing against actual PECOTA's for 2007-2008 (under the old Silver system) would be very, very helpful, too.
A full discussion of various iterations of current PECOTA vs. 2007-2008 PECOTA vs. (say) Marcel and CHONE, with complete explanations of the methodology and a direct comparison of the slash stats would be welcomed by many.
Combined with a full explanation of how we got here, and (critically) a good PECOTA performance in 2010, and that's dirt in the hole rather than dirt out.
I wouldn't mind a true-beta; one with some bugs here and there. But it simply has to be closer than every single release so far.
Dave, if you're writing a chronology, explaining why facially defective updates were put up is important. If that explanation is, "We're idiots," that's better than no explanation or an implausible or excuse-ridden one. The latest Depth Chart snafu is one of many that are obvious to moderately mathy folks. I've articulated some of the many other problems elsewhere.
I realize I have a pitchfork and a torch, here ("Some problems are best solved by angry mobs" - Homer Simpson, and if it wasn't him, it should have been) but I think the pitchfork and torch crowd are right.
And I didn't come with an axe to grind (to mix my metaphors); you can see my number, I've been to at least a half-dozen BP events, my Scoresheet league is called AL_BP_NorCal, I've bought every book since number two, y'all have kindly paid me for a couple of articles, and I've been acknowledged in the book (which I found quite flattering). It took *a lot* to get this torch lit.
But I think it's important to shed light on the unsupportable continued releases and continued reliance on PECOTA. Further, it's important to show that what you're saying now doesn't explain nearly enough, doesn't make up for continued obvious errors, and doesn't speak in a properly meaningful way to PECOTA 2010 v. 0000011's accuracy.
No more releases of Depth Charts until wins average 81 (or slightly under, if you want some rainouts.) No more releases of PECOTA that facially don't make sense. No more using 2009 data to determine if it accurately predicts 2009 outcomes. No more tinkering. Back-engineer it to 2008 PECOTA if you can; that was the last well-performing iteration. If not, well, there's 2011. But enough of this course of conduct, which people are rightfully quite angry about.
I'd note the comparisons appear designed to show the newer PECOTA is better.
First, back-testing on one year of data is insufficient.
Secondly, was the back-testing done after filtering out the known information (the 2009 season?) That is, doesn't the 2010 PECOTA have some 2009 season information in it? If you did filter it out, you should have clarified this. If not, the data's value is damaged.
Thirdly, what bias exactly was filtered out?
Fourthly, the BA/OBP/Slg are all worse now than they were with the book's method. You used a measuring system that a cynical unpleasant person might think is designed to hide this.
Fifthly, the changes keep coming. I have to say, if I kept changing a system, I'd think I could fix it up to predict the 2009 players right well by this time.
To quote myself from the Depth Charts thread, I don't believe you when you say PECOTA 2010 is fine now. More very simple errors are being generated with each iteration, so I don't think you believe you either.
There's an ongoing discussion in the Depth Charts, for those interested in the thoughts of some who have followed the updates. Some of us are skeptical.
The depth charts now show the AL at 61 games over .500, which would be impressive, all the more if not for the fact that the NL is projected to be right around .500.
3/12 update: Looks like they've gone with 50th percentiles for the Indians. Rockies (like Fowler) still all fouled up. Minor fix; still plainly broken.
Oh, I can't pass up on the awesomeness of the Fowler projection (Good catch by objective-reality slave jberkon!). Fun all over:
1. The listed weighted mean doesn't add up.
2. The 80th percentile projection has BETTER rate stats than the 90th percentile projection.
3. The EQA's or whatever the hell they're called now have a better 90th percentile rate stat than 80th percentile. Park effects are different based on percentile! Awesomeness!
4. The projection is based on something unknown and unknowable; it's well under any projection figured out by, I dunno, looking at the percentile rankings.
Awesome. Every time I look at something else, I find something wrong. I'm admittedly good at pattern recognition, but it's fun to look at every projection and find something ridiculous!
Take two minutes and look. Find your own ridiculousness! You can do it!
Best. PECOTA. Ever, I say. With multiple projections for each player (50th percentile, weighted mean, "Projected" performance, and multiply that times multiple iterations of PECOTA).
You know who this is great for? Software QA folk. They're looking for jobs and can now say, "QA is important. Look at what happened here!"
I love how BP keeps citing PECOTA, too, knowing it's junk. Truth-hewing always did suck.
I took the actual projection at the top as the weighted mean. Apparently, they are not the weighted mean, or 50th percentile, or anything related to the other numbers at all.
BP's willingness to retain the mystery surrounding the projections is exciting for those who want to find hidden meaning in them. Anyone interested in hazarding a guess as to what the projections now mean is encouraged to do so.
And jberkon, your insistence on the projections meaning something is just a buzzkill, dude. I've been there. I've railed against many apparent errors. But when you realize the projections aren't real, man, that they're just Player Expectations Coming Out The Ether (and if I could just find an a-word substitute for Ether...) then you can chill out.
Like BP has done. Put out new projections unread by human eyes, untested by people looking at them, and unrelated to any data. Don't be a stathead, jberkon. Don't choke off innovation based on your requirements to conform to your precious objective reality. Then it'll all be cool.
The problem I note above on Marson has gotten much worse. See my comments on the Depth Charts. Clearly, no one is reading these comments who is looking at PECOTA. (The alternative - that they were aware of the issues and still put out a seriously broken update - seems worse.)
Anyone competent who looked at the last release would have seen it was broken very quickly (it took me two minutes.)
At this point, I believe someone has to know the degree of the breakage, and they believe admitting that will damage their standing with unsophisticated users more than not revealing will damage their standing with sophisticated users. When you can't do a weighted mean right, it's time to go home.
I welcome contrary views. I do not welcome assertions that everything is fine with PECOTA 2010. Because it is impossible to be marginally mathmatically competent and assert that.
I'm offended and angry, in case it doesn't come across. A broken update with no commentary and no sign that anything will be fixed should make people angry. Trying to hide the ball should make people angry. PECOTA is broken. PECOTA is broken. PECOTA is broken. Say it!
Keith, the links don't work from the main page. Find a PECOTA beta post, and go to the PECOTA beta page (Or hop there from the fantasy page, if that still works). The 2/28 PECOTA's are from 2009.
I have some new PECOTA acronyms. I believe the OT in PECOTA now stands for "Out The."
March 9 update is up.
Gentlemen, stop telling us PECOTA is fine for 2010. It isn't. The new information is worse than the old. I am trying to be amused.
Let's see what happened:
Stuff scores are wrong? Sure, but everything else is fine.
Upside scores are wrong? Sure, but everything else is fine.
Comps are screwy? Sure, but everything else is fine.
SS/Sim numbers look to be miscalced at some positions? Meh. No big whoop.
Book and website numbers are way off in some cases? Natural order of things.
Long-term projections are impossibly off? Sure, but 2010 is fine.
2009 was PECOTA's worst year? Fluke!
Weighted averages don't quite match up with the percentile forecasts? Well, we fixed that! Now they don't match up *at all*. A quick check of some Indians (look at Druby and Marson) has their weighted mean below their 20%ile projection. Other players run just a little low.
If you tell us one more time that the 2010 projections are fine, I don't believe you. More to the point, you don't believe you. Putting up this very seriously flawed update does not give one confidence that learning is big on your list of things to do.
Now, maybe I've misread something. Maybe I've gotten something wrong here. I welcome correction if so. But.... I don't see it. I don't know how to reconcile the Druby and Marson projections with a system that is working properly.
This just kills me to be this harsh to a product and brand I've been loyal to for forever. Competence problems are serious. Trust problems are worse. Now's the time to confess: It's broken.
No one on this thread has talked about what the actual number of games cost is. Nathan figures to have generated about a 3.5 WPA. A workmanlike Fuentes 2009-level year is about 1 WPA, plus there's the loss that the new closer is replaced by Shiny New Pitcher To Be Named; that looks like a one-win difference. I'd say this is a 3-4 win loss, so it's time go bet even more heavily on the Indians at the ridiculous Vegas odds available. At the place the Twins are at - they look like an 85-win team or so - this is a bad injury. Nathan's a stud.
Yet another problem:
Some of the weighted averages look wrong. Lou Marson's percentile numbers don't seem to mesh to his weighted average.
I wonder if BP will make the results of the investigation public.
That's the first step. Do these things to defeat your demons:
Step 1: I see that I have no power over BROKOTA. Me and my buddies ground PECOTA into a fine powder, and snorted 'til it was gone. I must stop using PECOTA in its current form, and stop pretending sitewide that it's delicious in its white powdery form.
Step 2: I recognize that there is a power greater than myself that could restore PECOTA to sanity and could stop me from breaking stuff.
Step 3: I am turning my will over to Nate as I understand him.
Step 4: I will make a searching and fearless inventory of our PECOTA failings. I (and to the extent I can persuade them, my fellow BROKOTA addicts) will stop saying that individualized problems with virtually every number don't affect the final product.
Step 5: I will admit to Nate, to myself, and to others the exact nature of PECOTA's current failings and of our failings in busting it up.
Step 6: I am ready to have Nate remove all of the defects so generated.
Step 7: I will ask Nate to tell me how to stop this in the future.
Step 8: I will make a list of all the subscribers I have harmed, and be willing to make amends.
Step 9: I will make amends.
Step 10: I will continue in the future to actually watch for obvious errors pointed out by smart members, and admit errors early, rather than after the pitchforks and torches are already out.
Step 11: I will ask Nate for his will, and have his will be done.
Step 12: I will try to carry on this devotion to honesty and transparency, so others can emulate my behavior.
Follow the twelve steps to serenity. The guys at New Coke, Crystal Pepsi, the Apple Lisa, and the IBM PC Jr. went, and it all got better. We're rooting for you, and the recovery program is necessary.
Excellent. I bet in Vegas most years (missed one due to work) on winners of the pennants, but I don't like to bet the over-unders because with the vig, the hold, and the necessissity of doing something that costs to collect... I'm looking for a bigger payout. But this is fun.
Baltimore 71 - 76 - 73 Over. Really like Baltimore.
Cincinnati 79 - 79 - 78 Over. Even with Dusty. Low confidence.
Cleveland 68 - 75 - 73 Over. High confidence. Over 68, over 73, over 75, over 81. Gimme some over goodness.
Florida 83 - 81 - 81 Under. Expect a selloff.
Houston 74 - 73 - 77 Under.
Tampa Bay 89 - 89 - 89 Over.
Ideally, Rob can bring this back at the end of the season for those willing to predict publicly. Mock the losers, exalt the winners. (I will not be mocked! Or maybe I will!)
Heyward vs. Snider and Rasmus? I'd just note that Heyward's not all that comparable; Heyward's really a better prospect (accentuating your point). BP has Heyward hitting 96 HR's in the next 10 years.
I'd take the over and give 2-1 odds, thanks. And even that's a sucker bet.
But you have to predict who the outliers will be. The over/unders won't have a 100-win team, because predicting that a specific team will win more than 100 is much tougher than predicting some team will win more than 100.
And Montero's 2010 projection is spectacular; Montero's viewed to peak next year at a high level at the age of 20 and then decline.
This is a radical change in what projections look like for these sorts of players. There has been a substantial change to PECOTA. It isn't clear that the people who changed it know the changes made; the issues with all sorts of numbers are pretty substantial.
The SS/Sim numbers look a little off, too - Miguel Tejada oughtn't rank below Daric Barton for next year with the projections given.
The real problem here is that the cards are pretty, but broken in some untrackable way.
I'd take the over on Cleveland. I expect them to be much better than expected. Sizemore and Druby are excellent, Choo's pretty serious, and I expect the pitching (notably Masterson) to be better than most expect.
Joe Sheehan used to call for the under on some low numbers and be wrong every year.
Where'd our 2/24 update and 2/25 PECOTA cards go?
Check out the Yahoo Scoresheet-Talk page. Lots of good advice there.
ARod's quite different in the book and on the spreadsheet.
Some of the hitter projections are difficult to understand. PECOTA uses last three years, right? Luis Castillo's last three:
Castillo's 34, and we get a projection of:
I guess the OBP is justifiable, but I have a tough time with that slugging percentage. There are a number of other projections that look very odd; Kelly Johnson's expected career year sort of jumps out. Brian Roberts is expected to to do better at the age of 32 than he did at 29-31.... this seems like a pretty serious red flag. I don't remember these sorts of outlier projections for older players before.
Maybe I'm wrong. Maybe we should project a year for Brian Roberts that's better than all of the past three years. But I'm inclined to believe a more conservative projection (CHONE has Roberts with 50 points of OPS fewer than PECOTA) that doesn't project Roberts to hit better than he has for the past three years. It seems to me that something's wrong. (Reversing the effect of the very good pitching in the AL East?)
Thanks for the fast answers in this thread, BTW.
I agree completely.
The worst part is that there's not a visible quest for truth here. To use RS/RA numbers that approach impossibility is not a good sign; to keep with them after people point them out seems unwise to me.
These are smart people running the show, and they know this is screwed up. Caring insufficiently to even note that they know its screwed up in these articles is regrettable.
BP's got enough well-earned cred that it'll take a long time to run the reputation to ground. I hope they don't.
It looks to me like its a miscalc, using a raw stat instead of a rate stat.
I wonder if the stuff scores are in the book, and were available to the authors? This seems like something someone would catch virtually instantly.
On the main subject, the conversion of PECOTA off a spreadsheet appears to have resulted in some errors, and the loss of Nate appears to have made it hard to fix those errors.
It's unclear if there are more invisible errors generated. Like sfastatsprof, I'm going to give far more credence to CHONE and Zips relative to PECOTA than in the past.
I do too much, probably, but you don't have to. There are weekly lineups, and most owners don't change them unless they make a trade or have an injury (and you don't have to change them for injury; the program's pretty nifty about replacing your players.)
They have deep discounts for first-time players, too. I got into Scoresheet on a fluke - they had a BP event where they held a Scoresheet draft in the Bay Area, and it looked like fun, so I went. Put your toe in the water, and if you don't like it, it'll be OK. You can join a one-year league (I'm pretty sure) for 2010 to get the hang of it.
I echo the praise for Scoresheet. It's a different, somewhat more complicated game. But it's really fun.
I really think it'll be about $16.5m, but everything in that territory is taken. I'll try $21,105,000.
Aaron was 0-7 against Blyleven. Pretty good memory, there.
Derek, thanks for the clarifying comment to my unclear assertion. I had thought there was no mechanism for anything other than the limited en banc the Ninth practices, but a quick check shows that there apparently is a way to do that.
Still, in 2003, Kozinski said the Ninth had never had a full en banc hearing (limited en banc having worked sufficiently well since its inception.) See: http://www.fedbar.org/Kozinski_testimony.pdf.
My understanding - and I don't practice federal law and could well be wrong - is that hasn't changed and that the Ninth has not had a full en banc hearing since they changed the rules decades ago to the limited en banc. I believe that the limited en banc decisions have been reviewed by the Supreme Court many times.
(Sorry, non-lawyers. Soon, you will be back to your regularly scheduled programming on this issue: Steroids! Steroids, steroids, steroids! Also, I think Derek knows way more about federal practice than I do.)
This was an en banc decision - this was the whole panel, as Derek notes.
It's a decidedly unusual decision in a lot of ways. I recommend Orin Kerr's posts on the subject at www.volokh.com.
I don't think the very specific procedural rules set up by the Ninth will hold up. Of course, I could be wrong.
Bakersfield is CSUB.
I made this point when the topic was announced. I salute Matt for his graceful exit, and Ken Funck for noting that it's a fair point - even though he got to the finals with it.
Me, I'd rather have the next Joe Posnanski than the next Jay Mariotti. The defensiveness by Mr. Carroll on this issue is striking.
I wonder: How many people play the podcasts vs. read the articles? Maybe it's some larger percentage, but I'd be surprised. I think the audience is smarter and better than that.
And I guess I go back to the formation of BP: Gary Huckabay - who is incredibly personable - got Keith Woolner and Clay Davenport based entirely on their written work. Maybe things have changed. Maybe I'm unaware of the media issues today. Maybe I'm an outlier.
But I'm unhappy about this predictable result, and I don't think I'm alone. The remaining contestants are very good, but I was looking forward to a Funck-Swartz finale.
I've done transcriptions. It's brutal, time-intensive work.
I understand faulting BP for not having it done in-house, but I salute Richard for doing it.
Will wrote: In fact, find me a major writer that can't do radio *and* television.
Arguably America's best feature writer, Gene Weingarten?
Greg Smith of SI?
Dave Eggers, founder of McSweeney's and brilliant writer (and former contributor to ESPN The Magazine)?
That's just off the top of my head (and maybe they've done more radio or TV than I'm aware of, but I could come up with a dozen more with a little thought and the help of Google.) Writers are often odd ducks, or are not amenable to public appearances.
I'm disappointed in the format. I think it's unwise; I know BP does a lot of co-branding and various people make appearances on a lot of shows, but to eliminate someone at this late stage for a short interview seems misguided.
I'm reminded of Bill James' occasional media appearances; he was always a disaster on TV or radio, despite the brilliant writing. (BP founder Gary Huckabay is wonderful at this type of thing.)
I've done media appearances, and I do fine. But for the people who don't do fine, for whom it's not a plus.... I'd hate to lose the best writer because he's not radio-friendly.
Finally, if you determine radio's a critical element (which it appears you have), failing to give each candidate a one-hour primer on radio appearances seems to me to be an error. The question for BP isn't whether a person can do good radio right now, but whether a person could do good radio with some advice and training.
But that's just my opinion. I could be wrong.
Jesus Montero's two dingers put him at 855 OPS in the Eastern League at age 19. That's.... that's good.
I thought this was a particularly nice, breezy stylized discussion of the game, with jumping-off points that seemed unforced. One of my two thumbs-up pieces this week.
While we're going with analogies, using Win Shares to do baseball analysis is like using a rock to drive in a nail - it'll sort of work, but it's not a particularly good idea. With the avalanche of good fielding data and methodology, using the Win Shares fielding metrics is just a disaster. Even if you use better data, the methodology outlined in Win Shares is wholly meritless.
I liked this a lot better than the BP reviewers did. It was a fun little topic. I like the link-o-rama, too.
It seems to me that the author did address Pedroia's real height with "generously listed at 5'9"."
The Silver Bullets got routinely killed by men's teams of every level. To play them off as competitive isn't accurate, IMO.
Is it so hard to believe women can play? Yes. Yes it is. I'd like it to be otherwise, but I don't see it.
Mine was on the pitch data. You can massively improve projections with it. I have, and I feel like I'm just scratching the surface.
As someone whose failed entry in the contest was improving on PECOTA with available data, I gotta like this article.
The use of the data available is a huge boon to fantasy players. Systems which don't use this data can be improved by using it. I thought this was a great piece, and I don't understand Will's visceral reaction to improving PECOTA.
Also, picking up Ludwick last *June* indicates some sort of broken league. Ludwick had an OPS of over 1.000 in April and May. That's not luck. That's taking the awesome guy. He didn't just pop in June.
That's pretty grim, I'm afraid.
Wieters is currently hitting .222. I think this sample size is sufficient to conclude that Sally Radigan will have to undergo more chemotherapy, the Orioles won't win the pennant, Wieters will be selling insurance by next Friday, Jamie Walker will still suck, and humans will continue to die.
The apocalypse begins.
TT seems to be defaulting to the projections, rather than the defaults that I've personally set.
But I'm glad to see it up early this year!
Hey, if you have problems at Penn, just tell them you'd be willing to be booted for Joe Paterno or some other university legend, but not for Dumpster divers. That should get a reaction you like.
I'm headed out to Alameda; I'll be there a truncated time myself so I can pick up a leaguemate at the airport. I'll be the chubby guy with the glasses. One of them, anyway.
"Were you there before it was in the Big 10?" - A colleague of mine, about my undergraduate college.
Usually, catchers are comped to catchers, but there are no good comps of any real sort for Wieters; Pecota\'s chosen non-catchers as the best comps. (Wieters SimIndex is zero, meaning the comp list isn\'t really similar to Wieters at all.)
Since the comps are non-catchers, it\'s projecting playing time consistent with non-catchers. This is actually fixable, I think, and I\'d guess that we\'ll see an adjustment for this by Pecota 2011.
\"Very soon\"? If like, \"within a week,\" is \"very soon,\" I have a Woo-hoo for them. If the Scoresheet values are coming very soon, I have a triple woo-hoo. Pecota Day is one of the best days of the year.
Yankees 5/155 12/12
Matt Wieters is \"arguably one of the top prospects in the game\"?
Is Albert Pujols \"arguably one of the top 1B\'s in the game\"? What\'s the counter-argument?
The 100% number looks like a mistake; the Rays never really hit 100%.
If you look at the chart for the Rays, they go from 75% to win the division on Sept. 1 to 100% the following two days, then back to right around 75% on Sept. 4. This is not possible; you can\'t be 100% to win to many significant digits, then lose 25% of that overnight.
During the same time period, they bounce to 0% to win the wild card; same impossibility problem.
Even if you look at the totals, that\'s impossible - you\'re not going from 100.00000% to 99.59320% in one day; that\'s going from less than one in a million (the last digit applies to ties only, I think) to one in 250 in a day. The chances of the worst possible outcome - a Rays loss and everyone else winning - is far more than one in 4,000.