I’d been putting this one off until I had finished what I had promised last week - which is get holds into the PFM.
AND THEY ARE!
This went rather better than I did last week, when I posted a new version on Friday morning and wound up with Manny Corpas as the best pitcher in the universe for several hours. They’re in, they work well through all the poking and prodded I put them through, and let’s hope they work for you.
That’s the biggest news. There are a whole slew of changes that also went in PFM and the depth charts today, as we get far enough into spring to start to make some judgments about some players - primarily people who were borderline picks who then started spring 1-20. But there are plenty of situations that have changed in the last week, like in Washington (I would not have picked Willingham to be one of the odd men out) and St. Louis (Schumaker is still playing second, but if they ever pull the plug on that I’ll have to change five positions - serious musical chairs!). I know I took Ervin Santana down a few notches, and made changes in Seattle, Kansas City, Oakland, Baltimore, Toronto, Houston, Pittsburgh, Cleveland, and the Dodgers - and probably a few more I’ve forgotten I made.
And then there’s A-Rod and his achey-breaky hip. I’ve knocked his playing time down to 60%, and I still think that might be generous to him. Last week in LABR he only went for $21 or $22, making him about the 11th ranked third baseman to LABR voters. Besides the playing time, I have to think the stolen bases are gone for 2009 (at least).
Ah yes, LABR. I’m happy with the way my team turned out, and happy also that one of my fellow players could describe it as “looking like a BP team”. It was strange for me in one respect - I don’t usually run out in front of the pack, spending-wise, but I did this time. That left me with long periods of waiting while everybody else spent their way back down to me. Still - if everyone on my team and everyone else’s team performs as expected, I show my team winning easily, leading the league in three categories, second in five, third in one and sixth in one. I’m sure other teams, with their own projections, feel differently. Here it is:
$22.00 C Russell Martin
$24.00 C Brian McCann
Because one stud catcher isn’t enough. I had my prices, they weren’t close to being reached, and I got them.
$27.00 1B Joey Votto
$44.00 3B David Wright
$14.00 CM Kevin Kouzmanoff
The corners are what Jason was looking at when he called this a BP team, because those are all BP kind of guys.
$1.00 2B David Newhan
$1.00 SS Ramon Vazquez
$1.00 MI Brendan Ryan
These guys aren’t. The downside of spending $46 at catcher. But I did win LABR two years ago with a $4 infield, so I can live with it.
$17.00 OF Justin Upton
$11.00 OF Cameron Maybin
$6.00 OF Josh Willingham
$3.00 OF Nate Schierholtz
$1.00 OF Gregor Blanco
$8.00 DH Colby Rasmus
That’s a group I really like, but not without risks.
$30.00 P Johan Santana
$13.00 P Trevor Hoffman
$5.00 P Cla Meredith
$8.00 P Rafael Soriano
$7.00 P Joe Blanton
$4.00 P Tony Pena
$9.00 P Oliver Perez
$2.00 P Andrew Miller
$1.00 P Daniel Cabrera
$1.00 P Dennys Reyes
A lot depends on Santana’s elbow. I wasn’t shooting for Hoffman, or any closer; I just couldn’t believe the bidding stopped there. Here’s hoping he’s got another year left in the arm. I’ve always liked Blanton, I had to go for Cla even if I have to ask “y?”, and everybody else has at least some potential. Soriano was an overpay, one I was cussing myself over as soon as I’d made it.
Reserves
Steven Shell
Cha-Seung Baek
Brandon Backe
Nick Evans
Everth Cabrera
Eric Bruntlett
I don’t have great excitement here, except maybe for Baek, who really should have been bought during the draft; I had him at $4, and got him with the seventh pick or so in the draft.
Cha-Seung Baek is an absolute steal there, and I'm amazed no one was willing to pony up a buck for the guy. I think he and Schierholtz are my favorite "value picks" on your squad.
Going to $3 on Schierholtz is what put me into $1 mode for the remainder of my choices - it was my max bid. At the time he came up, I had him as easily the best position player left on the board (which is a good time to go for your max), and was very happy that no one who could have said 4 did.
Thanks very much for the hardwork, Clay! I didn't see any unusual relievers jumping out with unreasonable Hold numbers. Good to see where Putz' value lands, now, for those of us with Holds in our league...
Is there any chance that we might see some kind of fielding metric added before the season starts, whether it's errors or fld%? Or are you done with tweaking the PFM for this season?
I LOVE the fact that BP finally added Holds to the PFM. Just one problem though, they're way too aggressive. Your projecting the top 10 in holds to rack up 269 holds in total; last year the top 10 only earned 264. I think the methodology needs to be refined because right now it's way too aggressive. Any chance you're going to explain how you came up with the numbers?
Also, I'm curious why you have Jensen Lewis with 25 holds and a 4.25 ERA yet Rafael Perez will only have 17 with a 3.37 ERA. Last year they had 4 and 25 respectively. Is it simply because Perez is a lefty? Do we know that Perez is going to be a lefty specialist and Lewis has the top setup job?
Anyway, thanks so much for finally getting this in.
Just wondering... your argument that he's being "way too aggressive" is based on the fact that he's predicting a 1.9% increase in holds this year over last for the top 10 in the league?
Yes. Statistics in baseball involve a large amount of randomness due to luck/chance. Projection systems are therefore quite conservative to try to filter out a lot of the noise in previous year's statistics. this is generally done by regressing an individual's statistics to the league or their career mean. This is the PFM projects only 1061 RBIs from the top 10 despite that last they had 1248. The PFM also only projects 138 wins from the top 10 despite there being 189 last year.
The more variable a statistic, the more a projection system tends to regress it. So you regress Wins and RBIs more than K/9 and SBs. Holds are such a variable statistic that until yesterday BP REFUSED to even attempt to project it (http://www.baseballprospectus.com/article.php?articleid=4739). Therefore you would expect Holds to be heavily regressed in the PFM's projections and the estimates for Holds to be very conservative.
So, yes1 i do find the fact that he projects the top 10 in Holds to have more than last year's top 10 to be VERY aggressive. I would go so far as to say simply wrong.
That being said, I LOVE that fact that they've begun trying and I am sure that they will improve the projection as time goes on.
Fair enough, but I think you're missing a point. With a stat like RBIs, there's likely to be regression, because there's more uncertainty as to the occurrence of the event itself -- that is, there could be fewer runs scored league-wide, or fewer RBI opportunities for top players, and naturally most projection systems (which, as you note, tend to be conservative) account for this.
But the hold statistic, by its nature, is something that should occur with much greater (and more regular) frequency. I'm not 100% sure how it's defined, but in almost any game -- unless the starter works up until the closer comes in, or pitches a complete game -- there is going to be at least one hold opportunity.
Whatever the case, I think the hold is a ridiculous stat, and using it for fantasy seems to me utterly pointless, but that's neither here nor there, really.
The reason that he's complaining is not the number of holds for the top 10, it's just that you can't project the top 10 holds guys for 2009 with any real skill, just like wins.
Here's last season's Top10 (and ties) pitchers with wins: Cliff Lee, Brandon Webb 22 Roy Halladay, Mike Mussina 20
AJ Burnett, Tim Lincecum, Dice-K 18
Dempster, Floyd, Lilly, Oswalt, Saunders, Volquez 17
How many of those pitchers were projected to be above 50% in w-l ratio?
It's not really about predicting the number of holds, but assigning them to the right pitchers. What's why the number should be ~80% of the 2008 stats or less.
Holds are actually quite variable as well though because they are very dependent on the manager's bullpen management. In the same game their could be one hold or six, depending whether the manager chooses to use one pitcher in innings seven and eight or six pitchers, one for each out. This is why St. Louis only had an average number of saves last year but by far the most holds (106 vs NYM's 99 and then Flo at 86). LaRussa loves to play the match ups and thus throws a ton of relievers in each game. Regardless while you can reasonably project how many holds a TEAM will earn depending on their RS/RS environment and manager proclivities, it is very difficult to project which PLAYERS will earn those holds.
In sum, I will let BP's prior REFUSAL to even attempt to project holds stand for itself. The fact that they now do so is great; the fact that they are so aggressive however is shocking.
As to their value, yes along with wins and saves they are a ridiculous statistic for baseball analysis. However for fantasy they are great simply as a placeholder to require teams to roster a mix of starters, closers and middle relievers; just like in real baseball.
Clay, can we get a position into PFM for Aaron Heilman so I can stop busting my draft tool and then having to go back and manually change "#DIV/0" to "Swing" in the TAB file? =)
I ran the latest PFM for NL only. This kicked out Newhan at 6 bucks, how can he be so high? He may not even make the team and if he does, playing time doesn't look good.
Clay-----Is it possible to get a summary of all the teams drafted in the LABR draft, and the dollar values?------it would really be a good reference tool for everybody------or at least maybe the order of players taken for NL only??
Some of the lineups have already been released (check a thread on rotoworld.com's forums), but the league owners are not allowed to release other teams until after the USAWeekly Fantasy Sports Edition from March 25th gets released. That's where the entire league drafts get released.
Something funky is going on with players eligible at multiple positions... I ran the PFM yesterday with a 15 game min and today with a 10 game min, with all other parameters the same. Most of the results were comparable, except Russell Martin's dollar value dropped significantly without any other change in stats. I'm assuming it has something to do with his 3B eligibility. On the 2nd run he drops below both Wieters and Iannetta in a standard 5x5 setup. This does not seem right. I suspect he's getting dinged for his ranking amongst 3B. If anything, a 2nd position should make him more valuable, not less.
16 comments have been left for this post.
Nice looking team, Clay.
Cha-Seung Baek is an absolute steal there, and I'm amazed no one was willing to pony up a buck for the guy. I think he and Schierholtz are my favorite "value picks" on your squad.
Going to $3 on Schierholtz is what put me into $1 mode for the remainder of my choices - it was my max bid. At the time he came up, I had him as easily the best position player left on the board (which is a good time to go for your max), and was very happy that no one who could have said 4 did.