by Keith Glab, BaseballEvolution.com
August 22, 2007
I don't have the exact quote, but earlier this week, Ozzie Guillen said that
no one could have predicted that the Chicago White Sox would have a season like
this one. That is a load of malarkey. The folks at Baseball
Prospectus saw this coming with their 72-90 predicted record. As it turns
out, those PECOTA projections may have been a little generous.
The White Sox carried a .448 winning percentage into play on August 22nd
after winning two games in a row. PECOTA had them pegged for a .444
winning percentage. Nail-on-the-head stuff, right? The thing is,
BP's system predicted the Sox to score 765 runs while allowing 857. Their
current numbers project to 696 runs scored and 839 allowed. So in Ozzie's
defense, even PECOTA did not believe the Sox would be this bad. However, I
seriously doubt that the Ozeroo was talking about anything besides the team's
record when he made his comments to the media.
Several White Sox players were made aware of Baseball Prospectus' predictions
during spring training, and they all laughed them off. Admittedly, so had
I and the rest of us here at Baseball Evolution. Of our
six prognosticators, Asher put
the South Siders the lowest with 79 wins.
Anyway, the folks at Baseball Prospectus appear to have correctly dubbed the
Sox' pitching, but overestimated their offense by nearly 70 runs. That's
significant. Let's take a closer look at some of the Sox' offensive
Tadahito Iguchi - Predicted: .288/.357/.437 Actual:
The two interesting things about the Iguchi prediction are that the OBPs are
reasonably close despite major discrepancies in AVG and SLG, and that Gooch is
now approaching his projections with the Phillies.
Jermaine Dye - Predicted: .286/.354/.525 Actual: .245/.309/.485
The Sox had faith in Dye after an abysmal first half, whereas they did not
show the same patience with Iguchi. Dye has rewarded them with a stellar
second half, and the Sox rewarded Dye with a contract extension.
This does not change the fact that Dye's overall numbers merit serious Dave
Kingman Award consideration. It was obvious that Dye was going to drop off
from 2006's numbers, and when I first looked at the Sox' PECOTA projections, I
thought it odd that they had predicted a relatively gentle decline for Dye, yet
still pegged the Sox for 90 losses.
Darin Erstad - Predicted: .238/.294/.325 Actual: .268/.325/.351
Here's a case where the Sox underestimated a player's offensive
contributions. Another reason I didn't buy into these predictions back in
March was because they assumed that Erstad would get 308 plate appearances
despite putting up abysmal numbers. The abysmal numbers I could see, but
with Brain Anderson, Jerry Owens, and Ryan Sweeney as viable alternatives, it
just didn't make sense to me that the Sox would stick with such a player.
But in truth, Erstad's numbers aren't that much better than BP's projections,
and he would actually have had a lot more playing time had he been able to stay
Joe Crede - Predicted: .268/.318/.477 Josh Fields - Actual:
PECOTA definitely overestimated Crede, who was coming off a career year, but
had severe back problems. Fortunately this system, Josh Fields came in and
exceeded Crede's performance enough to approach the projection for Crede.
But the veteran's 167 dismal at bats at the start of the year don't just go
away; they are a big reason why the Sox have not scored many runs this year.
Juan Uribe - Predicted: .266/.310/.459 Actual: .222/.277/.362
If PECOTA had predicted Uribe's actual rate stats, I would have launched the
same protest that I did with Darin Erstad. No way a team sticks with
someone who has a sub-.280 OBP, right? But the Sox had no one ready to
replace the struggling Uribe in their farm system, and Alex Cintron seems to
have spent most of the season on the BL. As it turns out, PECOTA was
generous to Uribe, as we might expect from a system named after a weak-hitting
shortstop, and he is a big reason why the overall runs projection was a little
Pitching-wise, PECOTA greatly underestimated Javier Vazquez, Mark Buehrle,
and Bobby Jenks, but actually overestimated Jose Contreras and much of the Sox'
bullpen. Obviously, things evened out to a very nice overall prediction
for the staff.
PECOTA seems to do better overall with a team's projection than it does with
each individual player. This makes sense, since it is based on historical
trends, and because player performance is more volatile than team performance.
I'm not saying that I will drastically change any of my pre-season picks for
2008 based on what PECOTA puts out, but if we do have a 20-win disparity on a
particular team, it may give me a little more pause now.
Disagree with something? Got something to add? Wanna bring up something totally new? Keith resides in Chicago, Illinois and can be reached at firstname.lastname@example.org.