Wednesday, March 17, 2010
Looking Ahead to 2010: Jorge Posada
I’m going to start my player projection pieces like I usually do, by looking at Jorge Posada.
It took an injury-plagued 2008 that saw Posada have his worst season as a Yankee for a lot of people to realize just how important he was to the Yankees. Posada’s 2008 injury was the single biggest reason the Yankees missed the postseason. Given his age and the history of catchers in their mid to late 30s, it was a very strong possibility that Posada was at the end of the line, but he came back strong in 2009. In fact, of all the players in MLB who had at least 300 PAs and played 70% of their games at catcher, the only player who out-hit Posada was the second-most valuable player in the AL, Joe Mauer.So 2009 was pretty cool. In fact, I seem to recall the Yankees winning the World Series or something, but it's over. So what do the projection systems see for Posada in 2010? I'll be using the following projections this year:
- CHONE, by Sean Smith of baseballprojection.com. Sean is one of the best analysts around, and in just about every test CHONE has outperformed all other projection systems.
- Marcel, by Tangotiger of The Book fame. Although it's consider the most basic projection system, it's built on a solid foundation and it is the standard that any projection system should try to beat.
- Oliver, from The Hardball Times. Brian Cartwright has done a lot of work in building Oliver, and although it's the new kid on the block I expect it to be pretty good. While it's not completely free, the Yankee projections are free.
- PECOTA, from Baseball Prospectus. Although it had a bad season last year and so far this season seems to be fraught with problems, it's been a very good projection system prior to 2009, although their claims about deadly accuracy are stupid. The one caveat here is I no longer have a subscription to Baseball Prospectus, so the projections I have are from their February 25 update, so if they've changed since then I won't know.
- ZiPS, from Dan Szymborski of Baseball Think Factory. Dan puts a lot of work into these, and if CHONE's not the best projection system, then ZiPS probably is. I won't hold the fact that Dan's been using ZiPS for some stuff at ESPN against him.
- CAIRO, which is my own projection system, which was created for the sole purpose of making the Yankees look better than they really are.
As far as the position and replacement level adjustment. As I mentioned in the last CAIRO update, I'm going to use the methodology that is used by Fangraphs and Tangotiger. What that means is:
a) Calculate a player's runs above average, without considering position. b) Add the runs above a replacement level hitter, again ignoring position, which is 22.5 for 700 PAs. So it'll be player PAs divided by 700 times 22.5, which gets added to the player's runs above average.
c) The following are the positional adjustments that are used for 700 PAs.
So the average catcher is 12.5 runs worse than an average hitter over 700 PAs, so we add 12.5 divided by 700 PAs times the player's PAs to account for that.
2) Defense, using both standard Zone Rating and UZR. While defensive metrics are not as trusted as offensive metrics, I think they tell us something useful for the majority of players, particularly when you start looking at multiple seasons and a larger sample size. For catchers, I won't use zone based metrics, I'll use a formula that looks at SB/CS, WP/PB and errors. While there are other aspects to a catcher's defense that are important, at this point we don't have a solid methodology to look at those so I'd rather just focus on the things that we can get a decent handle on.
3) Non stolen base baserunning, which I'm projecting using Baseball Prospectus's baserunning data from 2006-2009.
4) Value, which will just be the sum of the projected offense, defense and baserunning.
wOBA: Weighted on-base average
BR: Linear weights batting runs, not position-adjusted
BRAR: Batting runs above replacement level (position-adjusted)
BRAR/650: BRAR pro-rated to 650 PA
There's a pretty wide range of projections in here, with Marcel and Oliver the most optimistic and PECOTA/ZiPS the most pessimistic. Posada's an outlier, and the systems that regress him towards other 35+ year old catchers are going to penalize him more. On the average though, he still looks like an asset with the bat, which he's going to have to be, because the defense and baserunning aren't so hot.
It's important to understand when looking at a projection that it's a baseline, it's not a prediction of what's going to happen. Even if the projection was deadly accurate in terms of gauging a player's true talent level at the current snapshot in time, a player's performance can fluctuate around that talent level in ways we just can't predict. You can think of a player's stat line in terms of a normal distribution. For example, if Posada hits his projected playing time and the average projection is accurate, we'd expect him to hit more than 13 HRs a little less than half the time, and fewer than 13 half the time. To better illustrate that, here's how Posada's CAIRO forecast looks in terms of percentiles, which is just a fancy way of showing the aforementioned normal distribution, with a bit of an adjustment for playing time (the better a player plays, the more he's likely to play).
I think Posada will be somewhere between the baseline and the 65% forecast as long as he's reasonably healthy.
Inn: Defensive innings at catcher
TE: Throwing errors
FE: Fielding errors
WP+PB: Wild pitches plus passed balls allowed
SB: Stolen bases allowed
CS: Runners caught stealing
CS%: Percentage of runners caught stealing (league average was around 23% in 2009)
SBR: Runs allowed compared to average due to stolen bases
WP+PB R: Runs allowed compared to average due to WP/PB
TE/FE R: Runs allowed compared to average due to throwing/fielding errors
RS: Total runs allowed compared to average (SBR + WP+PB R + TE/FE R)
Posada's throwing really wasn't a problem last season, as his 22.3% rate of throwing out basestealers was just a smidgen below the MLB average of 23.0%. His biggest problem was the passed balls, which has generally been a problem every year. He's not a good defender at this point, but he should still hit enough to compensate for that.
Having a catcher that gets on base as often as frequently as Posada is great. Having a catcher that runs the bases like Posada on base that frequently? Not so great. Posada led the league in SB% in 2009, but in non-SB baserunning he continued to show that he's one of the worst baserunners in baseball.
GA_OPPS: Ground advancement opportunities
EQGAR: GA_OPPS runs above/below average
AA_OPPS: Air advancement opportunities. Opportunities to advance in sacrifice fly situations.
EQAAR: AA_OPPS runs above/below average
HA_OPPS: Hit advancement opportunities. Stuff like first to third on a single, etc.,
EQHAR: HA_OPPS runs above/below average
OA_OPPS: Other advancement opportunities such as wild pitches, passed balls, etc.,
EQOAR: OA_OPPS runs above/below average
OPPS: GA_OPPS + AA_OPPS + HA_OPPS + OA_OPPS
EQBRR: EQGAR + EQAAR + EQHAR + EQOAR
So, adding it all up...
Offense is just the average BRAR from the first set of projections, and RAR is runs above replacement (offense + defense + baserunning). WAR is wins above replacement, wich is RAR divided by ten, and then the value dollars after that show Posada's projected 1.5 WAR would be worth based on some different values of a marginal win. So if a marginal win is worth $3.5M, Posada's worth around $5M.
Now, we should have a reasonable amount of confidence in the offensive projection. With the defense and baserunning it's a little more nebulous. Still, I don't think it's a stretch to say that Posada gives back a decent amount of his offensive value in the other aspects of the game. He still looks like he should be an asset in 2010, even if he's overpaid.
There's barely two weeks until Opening Day, so I'm going to try and churn through these quickly. The site's still a work in progress and we may have some growing pains so please be patient with any of those as well.
Page 1 of 1 pages: