CSS Button No Image Css3Menu.com
New! Search comments:
(NOTE: Relevance, Author, and Article are not applicable for comment searches)
Teams are not randomly assigned to locations, so a simple correlation does not provide any evidence about causation. There are a lot of variables to control for, which the paper (linked to below) does.
I don't have a dog in the fight regarding the tax rates, but as someone who spends a lot of time thinking about how to properly control for things to get at causation, I get frustrated when someone throws a simple correlation at a complex question and calls it evidence.
The paper itself is available online. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2946169
There's a lot of work in the paper trying to account for 'all else being equal.' It's also looking at 40 years of data (tax rates vary over 40 years).
This article, on the other hand, only does simple a correlation. Let's say I said "Life expectancy is higher in France than the US. People smoke more in France than the US. Therefore, cigarettes cannot cause lower life expectancy." You would probably think my analysis somewhat limited.
What if the manager could re-arrange his lineup to give hitters higher-leverage situations? Once or twice a game, the manager does get to move <span class="playerdef"><a href="http://www.baseballprospectus.com/card/card.php?id=59432">Mike Trout</a></span> to the RISP <span class="statdef"><a href="http://www.baseballprospectus.com/glossary/index.php?search=AB" onmouseover="doTooltip(event, jpfl_getStat('AB'))" onmouseout="hideTip()">AB</span></a>. How, if at all, would that change the way we assess hitters? (Assuming the same total numbers of <span class="statdef"><a href="http://www.baseballprospectus.com/glossary/index.php?search=PA" onmouseover="doTooltip(event, jpfl_getStat('PA'))" onmouseout="hideTip()">PA</span></a>).
I know this is the most important issue ever:
The team from Houston became the Astros for the 1965 season. <span class="playerdef"><a href="http://www.baseballprospectus.com/card/card.php?id=21185">Larry Dierker</a></span>'s season that makes you go "Who was Larry Dierker?" was in 1969. At least, I assume. That season was a good season. Thus, somewhere in the database the team from Houston is misnamed, for its 1969 season.
I realize now that I basically said what yibberat said. Whoops.
Could they have a home rotation and a road rotation? Maybe have 1-2 pitchers who are constants, and 3-4 that rotate in from the bullpen or the minors when they are at home or on the road, depending on their skill-set.
I don't see any established pitchers going for that, but with some MiLB guys and journeymen...
It probably wouldn't work! But it possibly could!
What are the historical results of the top breakout % player?
If at least 8 teams' starting catchers are below replacement... doesn't that say something about replacement? (especially since no other position has more than one). When was the last time this was adjusted? Is the distribution of catcher quality just really wonky right now? Are we underrating catchers?
I think your reaction to <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> is a good one. Unless your probability that PECOTA is right is 0%, the projection contains information you didn't have before. Even if you're only 20% sure it's right, the logical response is to incorporate that 20%. If you were at 88 and PECOTA says 76, adjusting to 86 makes sense.
As a Cubs fan I've enjoyed <span class="playerdef"><a href="http://www.baseballprospectus.com/card/card.php?id=82958">Len Kasper</a></span> and both Jim Deschais and <span class="playerdef"><a href="http://www.baseballprospectus.com/card/card.php?id=19504">Bob Brenly</a></span>. Len made his way up from the minors, and has the curiosity and enthusiasm that I expect will eventually make him into an elder-statesman sort. Plus he's numerate. You could put him on alone and he'd be able to talk about interesting things--not on the level of Scully or Uecker, but maybe someday.
The ex-players... Brenly did occasionally 'talk down,' but not as a daily occurrence. Deschais doesn't do it at all.
Of course, Pat Hughes is one of the grandpas to a fanbase.
I see where you're coming from, though. There's a bit of a vicious cycle--"low quality" announcers can be stepped on by ex-players, which prevents those "low quality" announcers from developing well. No ex-player would talk down to Vin Scully, so it's not just a 'hasn't played' thing.
I'll be honest and say that I don't necessarily want announcers to adopt "intersectional analysis," however good that might be for baseball journalism.
Is there a place we can see the bullpen leverage charts? That would be a pretty cool feature.
What's better for development, giving someone more, repeated PAs against a good pitcher, or giving them more PAs against the best pitchers on a per-pitch basis? It's not self-evident to me that being forced to make more adjustments than normal is harmful, in the long run.
I remember reading an article on <span class="playerdef"><a href="http://www.baseballprospectus.com/card/card.php?id=58350">Craig Kimbrel</a></span> showing that the NL East had started to do better against him, since they see him so much more often than other divisions do. If Russell is struggling most against relievers, isn't that who he should be facing the most?
Are these stats available for his minor league days? I'd be curious to see if his weakness is 1) Transitions or 2) Major league relievers. (and of course 3) random variation) Parsing the minor league numbers should help differentiate the hypotheses.
I enjoyed this article and learned something--I hadn't thought about this aspect of the batting order. Kudos.
Very good article, and I don't even care about the Mets!
One thing I noticed--and it may just be me--is that Granderson is picking up the ball a bit differently, maybe slowing down less when he picks up the ball. I wonder if he's simply approaching the ball in a way that gets his body in position to make a better throw. Rather than throwing with his arm, he's throwing with his body. It's hard to tell with some of those clips.
The list is sorted by AHLF, which adds .20 to AL teams and subtracts .20 from NL teams. Without the league adjustment (which you can sort by if you click HLF up top), you can see the unadjusted HLF; Oakland and Boston move to 14 and 19, while the Giants sit at 8.
The reason OAK is ranked so high, despite their record, is a result mostly of their run differential, which predicts somewhere around a 40-29 record.
I've loved watching this year, but one thing about <span class="playerdef"><a href="http://www.baseballprospectus.com/card/card.php?id=70633">Addison Russell</a></span>... He's making Baez look selective. No games yet without a K, and a 46.2 K%. I'm not worried long term, but short term...
The fact that you have the same player in many leagues might be the result of the advocacy effect, but it might simply be the result of valuing a player more highly than average in the first place.
Multiple auction drafts would not be very good at testing the advocacy effect, because being the winner at an auction means "I value the player *at least* this much, but I may value him even more." A subsequent higher auction where you pay a higher price for a player may be advocacy effect, but it may simply show that you paid less than you valued the player at in the previous draft. Snake drafts are probably better at this.
Auction drafts might be useful is showing a reverse advocacy effect. If you miss out on a player, you'll think of all the bad things about that player and maybe fail to purchase a player at a price lower than your highest (unsuccessful) bid in the previous league.
Finally, the attention 'bias' may be perfectly rational. Attention is a scarce resource, and making a 'sub-optimal' choice from the point of view of perfect rationality may cost attention that is best deployed elsewhere. Time spent researching that second catcher could instead be spent researching batter matchups, etc. The real lesson might be to find ways to decrease attention costs in making choices--using automated ranking systems or other heuristics, or by playing in leagues with similar rules.
You'll almost certainly see a much clearer assignment of responsibility of who chooses the pitch. Instead of the slow transfer of decision as a pitcher matures, it's likely that a pitcher will go to the mound having been told "throw what the catcher says" or "you get to choose." Maybe this decision changes by opponent--if you're out of contention or are facing a bad team, you give your pitcher some leniency (this also provides an excellent 'random trial' evaluation opportunity).
I've been wondering if it's possible to quantify the 'safety premium.'
In the draft, $260*#teams equals the dollars spent in total.
At the end of the year, $260*#teams equals the total value produced (e.g., if you run a retro PFM).
However, some value comes from undrafted players. If 80% of the value at the end of the year comes from drafted players, then 100% of the draft money bought 80% of the value.
That means, if you know with certainty that someone will produce $20 at the end of the year, you should be willing to pay $20/(80%)=$24 for that player in the draft.
So, my major questions are:
How much value at the end of the year comes from drafted players?
Are consistent players undervalued using a PFM style forecast?
Kudos on the highlight video thingy. Very cool.
Thoughts: Do the major sports in smaller-market countries select non-randomly from the possible major league quality prospects? For instance, my complete ignorant view of cricket is such that the skills/physical characteristics that make for a major league pitcher carry very poorly to cricket, but that characteristics that make one a good major league hitter/fielder translate well into being a cricket hitter/fielder. A young athlete with good mlb hitter skills will be taken into cricket, while a young athlete with good mlb pitching skills won't be.
I don't know how this would explain Asia or Latin America, however...
My uncle covered the Boise Hawks, and he says that Dan Vogelbach is one of the best clubhouse guys he's seen come through. Mature, natural leader, all that.
Speaking of makeup though: do scouts tend to rate players on the 20-80 scale? Or is it a fuzzier 'good makeup, bad makeup' scale?
2015 is probably more realistic, you're right. Taveras and Piscotty will be lower variance. Shortstops tend not to last long into their 30s, so the chance that Peralta is the long term solution is small; there are fewer old shortstops than second basemen.
What did I say that made you think I said that Holliday is overpaid? I mentioned I think he'll be worth 3-5M above his contract per year, on average. The trade I proposed included Martinez, another potential all-star from a position of StL's depth, and an additional low 5 - 5 starter. Try reading it again.
How much sense does it make for StL to try to trade Matt Holliday? (2014-2016, $17M/year, 2017 option vests if top 10 MVP in 2016). Holliday has slipped quite a bit in the field, so a move to a smaller park or the AL might make sense (open a spot for Piscotty?). There's probably ~3-5M/year surplus in his contract, especially given increasing salaries (plus some additional possible surplus in the option--how many top 10 MVP players are worth less than $17M the next year?) I could imagine a package, with some pitching, for a high end SS prospect (most of which are with AL teams). Would Holliday/Martinez/Cooney be an awful deal for Correa, Lindor, or Russell (for either side?)
This is really cool, thanks.
The 80% cutoff is pretty strict, though, and I worry about SSS. More categories (80%, 70%-79.9%) would be interesting to see, whether the effect is somewhat continuous.
The big thing seems to be: watch out for Tony Cingrani next year!
Also, I love the Weaver .gif. It looks like he is throwing a dart.
It would be fascinating to see the relationship between the variability of outcome (e.g., bad days due to poor mechanics) and the funkiness of mechanics. My intuition says that pitchers with unusual deliveries are more likely to get into slumps--not only from the delivery itself, but from a pitching coach's difficulty establishing a baseline to figure out mechanical problems. When pitchers that have normal deliveries go into a slump, it should be pretty easy to find the problem, but when someone with a funky delivery goes into a slump...
The draft pick compensation does not change the overall demand teams have for players. If the QO system reduces some players' salaries, it will increase others'. Money that would otherwise go into these mid-level free agents might instead go to player development or low/high-level free agents. The compensation system does not fundamentally change the demand for or supply of player talent. So I would add a 4) Players unaffected by QOs get paid marginally more.
I wonder, do different teams systematically improve at different rates? I imagine with a good clubhouse atmosphere, manager, or coaching staff, a team can be better than average at adapting to a pitcher. A bad clubhouse atmosphere leads to less attention and slower adaptation. Could this be a sort of instrument to dig further in to the effects of a manager or 'intangibles'?
How much is it OF positioning? OFers come in closer for the pitcher than any professional hitter. That would be my guess for the reason, rather than a social norm against throwing hitters out at first.
Net manager changes could also be there (but if that were the case, you would expect to see lower SB but higher SB%, unless more conservative managers just suck at figuring out when to run).
1) Big skew in terms of RS. If you throw the bottom couple teams (e.g., the Marlins and their 2 RS/G) out of the equation, is run scoring still down?
2) People are taking an extra base 40% of the time, vs 41% last year. So similar it's hard to argue a 'base-path condition' explanation.
3) SB% is about the same as last year, indicating that if they did steal as much as last year it would be lower (Assuming diminishing marginal returns). So for whatever reason the conditions for stealing bases is worse, even if SB% is the same (since there's a selection bias where people don't choose their attempts at random, instead choosing the best opportunities).
4) Catcher (net) changes from last opening day: +Flowers +S. Perez, +Cervilli, +Jaso, +J Montero, +Laird, +Castillo, +Rosario, +Kratz, +Brantley, -Ruiz, -McCann, -Suzuki, -Olivo, -Soto, -Torrealba, -Pena, -Hernandez, -Thole, -Barajas. I don't know all that much about catcher defense, so I can't instinctively guess the sign of that effect (Assuming opening day starting correlates with more play the first 20 games). There are some stinkers on both lists, to be sure.
I wonder if there could be some sort of "expected days lost" based off the injury history and age of the players; if the front office or training staff is going to get credit/blame for lost WARP, I would feel more comfortable with a system that measured relative to a risk-adjusted baseline. E.g., if you keep Lance Berkman and Rafael Furcal healthy for a full year, you get more credit than if you kept Anthony Rizzo and Starlin Castro healthy for a full year.
Obviously this wouldn't be super easy, since the injury history of players who have been with the team a long time is a function of the quality of the training staff, so there's a bit of an endogeneity problem if you were to try to measure the training staff based on projected days lost, if that projection is based on injury history... the best information would come from players who changed teams and exhibited a difference in their injury profile after the change: Do injury prone players stay healthier? Do healthy players get hurt more when coming to the team?
That 'team' age 27 to 28 is not the same as individual players aging. If you have a 22 year old and a 32 year old, the former might improve half a win while the latter declines half a win, even though their average age is 27 going on 28, where you would expect an increase in value.
Including age-squared in addition to age would probably help to capture any non-linearities. The current equation constrains age to go either up or down, which means it can't go up and then down, like we expect for players.
Including some measure of team salary might be useful, as well. There is a selection bias if the older teams are less likely to keep playing bad old players and instead sign free agents. The Yankees will probably not age as much next year as the White Sox, for instance. You might see pitching age become significant with the inclusion of either of these variables.
Has there been any variation in the strike zone?
This would have been a better sentence: "Probably pretty predictably, Paul Phillips pitched pretty poorly." Oh yeah, alliteration.
Fun fact--I had to look it up after remembering Walker's AB batting righty against Johnson--Larry Walker turned out to be very successful versus Randy.
Bbref tells me that, from 2000-2004, Walker hit .393/.485/.571 with 6K's/2BBs and one home run in 33 plate appearances. Sounds like the Denver Post guy was right...
I know there's a huge literature on pitcher injury rates at a young age, but what about position players? Do you learn anything about the 'skill' of staying healthy for position players through their 20-23 year old ages? I would hate to be the Angels if it turns out Trout has a problem with a labrum somewhere. When to clubs/players usually figure out if a position player is injury-prone?
Is there some database for "tools." I like to get excited, but I have a hard time visualizing players using 'tools' without comparisons. Plus, it would be cool to have access to data like that...
The Cubs announcers have been talking about it quite a bit, but it seems like Soriano has made some changes to his defensive game, and according to UZR has been worth 6.9 runs above average there so far this year. ZIPS still projects him to hit .246/.292/.435 the rest of the year, and with his defense (it assumes another ~+5 this year) that's 1.7 WAR. And including this game, his OBP is higher than Pujols's...
Now, I know small sample size, but when the qualitative evidence and the quantitative evidence align I'm more likely to put some stock in it.
So the questions become: what does Lahair's defense in left look like, and how much better than Soriano can Rizzo hit? If Lahair is 10 runs worse in the field than Soriano, it's probably worth making the move, but any more of a difference you're sacrificing defense for offense and no additional wins (though if Rizzo is a better defender at first than Lahair, that could make up for some of it).
And a related thought: how do you value low-inning players going into a draft? It seems to me like a pitcher that averages 7 innings but is held back for the last month to stop from surpassing 160 IP might be valued more highly than someone who consistently throws 5 innings and reaches 160 IP (and puts up otherwise identical stats), because you can slot some other pitcher in once the converted reliever has been benched. On the other hand, if you have a playoff system, the opposite might be true. Has anyone looked at this?
Quick thought: If you can predict something better than PECOTA by adding a simple dummy variable (converted reliever = 1), why isn't it added to PECOTA?
The other thing to worry about Gorzelanny is that, with Carlos Zambrano moving back into the rotation, Gorey could be sent to the bullpen. Of Wells, Lilly, Dempster, Silva and Gorey, the two obvious guys are Silva and Gorz. Silva sits at a 3.35 ERA with a 30/10 K/BB in 48 innings. His FIP/xFIP stand at 4.24/4.15, and Gorey's at 2.77/3.56. (I don't have SIERA on hand). Obviously Silva's been worse, but one bad start by either of them while Z is getting stretched out could seal their fate.
I'm not sure you can say, a priori, that replacing Mariano with an average set-up man decreases Wetteland's aLI. Let's say in one world, he has a set up man who never allows a run. He'll come into a game with a lead of either 0, 1, 2, or 3. Replace that first set up man with one who always gives up a single run. Wetteland now enters in the exact same situations - 0, 1, 2, or 3 runs up, but those were previously 1, 2, 3, and 4 run leads. So the bad set up man decreases the aLI at 0 runs (as those Wetteland no longer enters) and increases the aLI of the other situations. It seems like an empirical question regarding a team's distribution of runs scored and allowed whether not having Mariano hurts Wetteland's WPA. Maybe I'm missing something.
I would vote Maybin. Gardner's superior and you'll still have speed.
Marc doesn't take PECOTA as gospel; his subjective rankings have Ianetta performing better than the above line and (perhaps) Doumit worse. Doumit's also coming back from a wrist injury, which can be scary for a hitter.
In my Yahoo league Jeff Clement is, for whatever reason, still considered a Catcher. Would you consider him a worthwhile 3-star player as a catcher?
You didn't really just quote Gordon Gekko for serious, did you? How would signing Albert Pujols be a zero-sum game? If the Cards sign Pujols at 23-24M/year, they've determined that's the best use of their money, and if Pujols accepts he's decided he's better off in StL, quite obviously a positive sum.
You lost faith in humanity because people hold some sense of fairness? One can argue that by not letting yourself accept an unfair deal is rational, because by punishing people for trying to offer what is perceived an unfair deal they reduce, a tiny amount, the chance they'll be cheated again. Also, there's the non-monetary benefit of knowing you screwed someone out of 9.95 since they only offered you .05.
I'm going to put in a minor quibble here, in particular with the use of the phrase "These two changes may have had some effect, but chances are the Yankees' improvement probably was due to regression to the mean."
Regression to the mean implies that luck was a big factor in the rating, which is true in a lot of cases in baseball (team BA with RISP, for example), and even in some factors of defense. But I would posit that defense is much more than luck. If the Yankees defense was truly 'regressing to the mean' then we could assume that their defense would be better regardless of personnel changes - and I doubt a team fielding Abreu and Giambi would actually improve relative to where they were - it would also imply that if they were fielding the same team in 10 years (when Giambi is 80) their defense would tend towards the mean.
Now if we assume that each team is an entity trying to optimize, among other things, defense, and that all things being equal teams will bid up talented defenders' salaries so that a team with superior defense will eventually lose them due to their increased salary, and poor defensive teams will seek out players who are better defenders - then teams' defense will tend towards the mean; but if this is the case then getting Swisher and Teixeira and shedding Abreu and Giambi are part of the regression to the mean, so listing off personnel changes which provide real improvement in defense, and appealing to the regression to the mean are redundant.
I know, nitpick nitpick...
They had a 500-HR club year where they included Rafael Palmeiro even though he wasn't an all-star
Which, if I finish the article (I started jumping around to the players you listed and Feliz's BB% numbers), you looked at that and noted last year's improvement. Whoops.
I think it might be useful to look at BB% (or UBB%), rather than pure OBP, since a large factor in OBP is batting average, which we all know is highly variable on BABIP. If you look at Pedro Feliz, you will see a player who, from 2001-2007 walked 4.9% of the time. Since coming to the Phillies, he's walked 7.2% of the time - last year was a career high at 7.1%, and this year he's at 7.5%. Only last year, his BABIP was .251, considerably worse than his career hitherto BABIP of .269 (which is still atrocious). This year it's .339 - hopefully, if he's swinging at fewer bad balls, his BABIP will climb a bit higher than .269 (with the new hitFX, I would be interested to see, once we have the data, if players who have upticks in their BB% also hit the ball harder afterwards, and can therefore raise their BABIP).
His OBP will likely drop lower than it has been now, but his OBP-BA should remain a bit higher than it has been career.
Johnny Damon anyone? He's turned into quite the power hitter; his iso stands at .252, more than 80 points higher than his career high TEN years ago.
Right now he's sitting at highs on his FB% and his HR/FB% (46.6% and 14.5%, respectively).
The iso is .198 on the road, and .305 at home, which is huge. The road split would still be a career high.
Looked up Nick Esasky. Career ended when he developed an inner ear infection and developed vertigo. Hope better for Votto.
In Jamie Moyer's 10th percentile rating, he strikes out -7 batters. That's right. When Cole Hamels strikes someone out, Moyer will come out and argue with the umpire, who then decides that the strikeout was instead a walk.