<$BlogRSDURL$>

Saturday, January 29, 2005

PECOTA And Sox Pitching

Doug Gribben, White Sox fan extraordinaire, pointed me to this post by Will Carroll about the White Sox's offseason moves. Among other things, Carroll (who writes for BP himself) notes that Baseball Prospectus's PECOTA predictions for the Sox's starting staff do not make much sense. I quote:
One other note: The Pecota for the Sox starting staff, which has all 5 starters posting ERA’s between 4.35 and 5.05 is a load. Yeah, US Cellular is unkind to pitchers, but these numbers would be career worst for almost each pitcher in the rotation. Pecota is generally a beautiful thing, but occasionally it misses the mark and I just don’t see the staff putting up those huge ERA’s.
Clearly, Mr. Carroll is an astute reader of Black Betsy. I just wish he'd give me a little credit .

Comments-[ comments.]

Wednesday, January 26, 2005

Why Does PECOTA Hate The White Sox?

Baseball Prospectus released its PECOTA projections for (essentially) all of the players that will see playing time in 2005. As far as I can tell, PECOTA projects that, technically speaking, the White Sox will be a steaming pile of dung next year. I mean, really bad. Baseball Prospectus doesn't allow people to distribute the PECOTA spreadsheet, but I do feel free enough to post the highlights.

Offense

PECOTA projects that the White Sox will have a disappointing season offensively. Although they don't compile team statistics, here's the combined line for all the current White Sox players (except Tadahito Iguchi):

.270 AVG / .327 OBP / .440 SLG

This is about the same AVG as the 2004 Sox, but a drop of .006 in OBP and a .017 drop in SLG. Given that Magglio Ordonez only played about 1/3 of the season last year and that Timo Perez was his primary replacement, the drop in slugging percentage is rather shocking. The Sox still project to hit about 200 home runs (assuming 5500 total at bats), but PECOTA suggests that no White Sox player will hit more than 26 home runs (Konerko). PECOTA projects that Jermaine Dye will largely be a flop with the White Sox, posting a .256 / .332 / .449 line with only 17 home runs in a total of 384 at bats.

PECOTA projects that Joe Crede will rebound to a .266 AVG, with an increase in his walk rate and 20 homers in 432 at bats. Over 550 at bats, that's more like 25 homers, which would be pretty good. It also projects that Aaron Rowand will maintain a lot of his surprising 2004 production.

Overall, the bottom line is that the Sox are expected to have more consistent production throughout the lineup. Only Frank Thomas has above a .360 OBP. Only Frank Thomas has a SLG over .500. For a team with a .330 OBP and .440 SLG, that is pretty consistent production over the lineup.

Pitching

So gruesome I can hardly write about it. The starting rotation is, at the top of the rotation, league-average. The bottom is not so league average. The bullpen is disastrous.

Here are the projected ERA's for the starting rotation:

StarterERA
Garcia4.55
Buehrle4.47
Contreras4.91
Hernandez4.35
Garland5.05

Yikes. I guess Buehrle, Garcia and Hernandez will have about league-average performances, but they aren't getting paid for league average performances. Contreras and Garland are pretty much no good (but an improvement over the 5th starters in 2004). It's an average rotation, notable for the lack of any horrible or good performances.

The bullpen...ugh. The lowest bullpen ERA projection belongs to Damaso Marte at 3.99. Shingo is a no-go, allowing 10 HR in 54 2/3 innings. New acquisition Hermanson stinks with a 5.09 ERA, and Vizcaino is equally bad at 4.91. In fact, no one other than Marte is under a 4.50 ERA. If this is the case, the Sox bullpen is in really, really bad shape.

One thing I've noticed about PECOTA is that it seems to push most players to a kind of league-average performance - there are not many highs or lows. People who played poorly last year are projected to play a little better. People who played well last year are projected to play a little worse. It's taking regression to the mean to the maximum. I'd be interested in comparing the PECOTA projections to 3-year averages and to see which are better predictors.

I encourage you all to get a subscription to BP. The subscriber material is excellent - it's not all statistics and many of the writers are genuinely funny.


Comments-[ comments.]

Monday, January 24, 2005

Soxtalk.com - Iguchi Signs 2-yr $4.6M Deal

The forums at Soxtalk.com contain a thread that links to a Japanese newspaper report that second baseman Tadahito Iguchi has signed a two-year $4.6 million deal with the Chicago White Sox. Iguchi, after originally asking for Kaz Matsui-type money, apparently has accepted a contract that will pay him (roughly) his salary last year in the Japanese league.

Here's a link to a Kyodo News article saying the same. Japanball.com reports the same as well.

So who is Tadahito Iguchi? He's a second baseman with decent power and (for the last couple of years) a very good batting eye. You can see his career statistics here. I have to say I'm excited by this move. Iguchi had OBP's of .438 and .394 in 2003 and 2004, respectively, and SLG of .573 and .549 in those same years. He hit .340 in 2003 and .333 in 2004, with 27 and 24 home runs. How will he do in the majors? Well, we only have other Japanese players to guide us. Here's Kaz Matsui's last three seasons in Japan, versus his first MLB season:

YearAVGOBPSLG
2001 (JP).308.365.496
2002 (JP).332.389.617
2003 (JP).305.365.549
2004 (US).271.331.396


Here are Hideki Matsui's last three seasons in Japan and first two in the US:

YearAVGOBPSLG
2000 (JP).316.438.654
2001 (JP).333.463.617
2002 (JP).334.461.692
2003 (US).287.353.435
2004 (US).298.390.522


Hey, why not Ichiro, too:

YearAVGOBPSLG
1998 (JP).358.414.518
1999 (JP).343.412.572
2000 (JP).387.460.539
2001 (US).350.381.457
2002 (US).321.388.425



Three data points aren't quite enough to draw many conclusions. However, it seems as though when players come to the US, these things happen: (1) average goes down by 30 points or more; (2) OBP declines by a similar amount - OBP-AVG stays about the same; (3) slugging drops way down - by 100 points or more. I guess that means we can't expect a .330/.410/.580 season from Iguchi. At $2.3 million per, such a season would be a real steal. However, if Iguchi hits .280/.370/.450, he'd be one of the best second basemen in the league, and would be extremely cheap at $2.3 million. I think his price was restrained very heavily by Kaz Matsui's flop last year (although Matsui's AVG and OBP lines were pretty good for a shortstop), and the risk that he'll flop like Matsui did.

The bottom line is that the Sox's options at second base weren't that great. So I like the Iguchi signing...especially at this price. With the additions of El Duque, Jermaine Dye, Dustin Hermsanson, AJ Pierzynski, Scott Podsednik and Luis Vizcaino, I believe that Kenny Williams has shored up every hole in the White Sox with above-league-average performers. He's had an A+ offseason, as far as I'm concerned. I'll post an offseason wrap up in a bit.


Comments-[ comments.]

Saturday, January 15, 2005

Thinking About Timo's Season

There is no reason to mince words; I don't think the White Sox should have Timo Perez on the roster, and I certainly do not think they should have given him a 1-year $1 million contract in this offseason. There are simply too many decent outfielders that you can pick up on the waiver wire or from someone else's AAA team for the league minimum to carry Timo Perez on the roster.

During the year, I complained about Perez getting at bats. But there's also no question that Perez had some big hits for the White Sox, especially his two-run ninth-inning home run in Florida in June. So I tried to figure out, did Timo, because of his clutch hits, actually benefit the White Sox?

There are a number of different ways to value Timo's performance. Baseball Prospectus lists Timo Perez as having created -16 runs above average. In other words, an average hitter with the number of plate appearances Timo Perez had in 2004 would have caused the White Sox to score 16 more runs. That's an enormous number of runs for a part time player. He also had a -8.3 VORP, another huge negative number for a part time player.

Another way of scoring a player's performance, however, is to judge each plate appearance by the amount of expected runs the player cost or generated for his team. My buddy Don Money explains ERV scoring over at Nats Blog. Basically, each possible offensive situation (e.g., man on first, two outs) has an expected run value. To lead off an inning, (no one on, no one out), the ERV is 0.5349, meaning that a team would expect to score a little more than 1/2 of a run in that situation. Of course, you can't score half of a run, so the number really indicates that over the course of 9 innings, a team will score, on average, about 4.8 runs.

For some odd reason, I decided to go through each one of Timo Perez's plate appearances in 2004 and use ERV scoring to see how many runs Timo cost the Sox versus an average player. This project took 6 hours or so over several days. And the results were a little surprising at times, but in the end, I don't know what it means. With no further adieu, here are the results:

MonthPAERV
April30-3.28
May451.39
June644.62
July65-3.92
August63-4.32
September51-1.68
Total318-7.19

The -7.19 figure is actually less than half the -16 Batting Runs Above Average calculated by Baseball Prospectus. But at -7.19 runs, Timo Perez was a giant sucking sound in right field, where teams expect better than average performance (by contrast, Magglio Ordonez was able to generate 3 BRAA in his plate appearances.

More interestingly is Timo's effect on a month-by-month basis. Timo's +4.62 ERV in June, if repeated throughout the season, would make Timo worth a total of 27 extra runs over the course of the year - which would have been higher than anyone on the White Sox (as judged by BRAA). But he had an abysmal July and August, and wound up costing the Sox several runs down the stretch - 10 runs over the last 3 months of the season.

Also interesting is how much Timo's bunting affected his overall ERV. Timo had a total of 9 sacrifice hits this year, leading the White Sox. According to my calculation, these sacrifices cost the Sox 1.61 runs (proving the ineffectiveness of the bunt). Because the bunting was Ozzie's decision, not Timo's, those runs should not be held against Timo. So I adjusted his ERV to account for that. By the same token, Timo reached on an error a total of 4 times, generating a total ERV of 1.14. Subtracting that as well (without assuming an out otherwise), I get a total ERV for Timo in 2004 of -6.72.

I don't quite know what to make of this. Overall, it seems as though based on his clutch performance (which Timo is unlikely to repeat, of course), Timo Perez was a better player than the stats otherwise suggested last year. As Hawk often says, it's not what you hit, it's when you hit it. But any way you slice it, Timo Perez was costing the White Sox runs, and he shouldn't have been resigned at any price.

By the way, I think I plan to keep ERV score for the White Sox in 2005. We'll see just how it all works out over the course of the season.


Comments-[ comments.]

This page is powered by Blogger. Isn't yours?