Thursday, December 16, 2004

Rating the Remaining Free Agent Starters

Here's my methodology. I looked at the remaining free agent starting pitchers according to I removed Carl Pavano, who has already decided on the Yankees, and Roger Clemens, who has accepted arbitration. I removed Pedro Astacio, Omar Daal, and Andy Ashby, who haven't pitched enough recently for their stats to provide meaningful insight. I removed Shawn Estes, Jamey Wright (whose 123 ERA+ occurred despite the 11th worst FIP in the league and 17th worst LD%, 60 innings minimum), Steve Sparks, Todd Ritchie, Todd Van Poppel, Darren Oliver, and Terry Mulholland because I can't conceive of them being viable options. I removed Hideo Nomo because there's pretty much no way to predict his performance (or at least, to predict a performance with some degree of confidence that would involve him not being horrific) at this point.

Next, I compiled the career data for each player. I also took each player's batted ball types from the Hardball Times for 2004. Unfortunately, this data is not available to the public outside of THT's 2004 stats, so I couldn't use pre-2004 data; I just pretended that the 2004 numbers would be close enough to stand in for every other year. Using these statistics, I computed ERA's for each pitcher with the formulas in JC Bradbury's study of balls in play. I also ran park adjustments for home runs based on the 2004 ballpark factors for HR, normalized for Dodger Stadium. This is, I know, not quite ideal, but it should be close enough. I also decided to normalize the stats for the National League and a .711 defensive efficieny ratio, which is what the Dodgers had last season (the best in baseball; all of these ERA's will look low because of that). For each player, I ran several ERA's, but the two I'll use here are one based purely on 2004 data and one based on a 5-4-3 distribution of 2004, 2003, and 2002 (what I'll call "Weighted 3-Year"). These are not projections; rather, they are interpretations of past data.

Before I start, it should be noted that every one of these players is past their nominal prime (age 27) and each one has pretty much declined over the past three years except for some of the oldest ones.

1. Matt Clement
2004: 3.41, 181 IP
Weighted 3-year: 3.26, 194 IP
Will Carroll gives him the yellow light, as he's been subjected to the Dusty Baker philosophy on pitcher usage the past two seasons and his arm bore the results in 2004. Nonetheless, everybody else on this list is a health hazard of equal or greater magnitude. Clement's best year was in 2002, and he declined somewhat in 2003 and stayed at that level in 2004. But he's still quite good, striking out a lot of batters, not allowing many home runs (at least, that is, when the Friendly Confines are taken into account), and maintaining a high but reasonable walk rate. Doesn't give up many line drives.

I've seen concern over his home/road and day/night ERA splits. I spent entirely too much time running the data to see if that concern was legitimate. While his ERA splits are completely ridiculous in that regard, his peripherals are not. I compared all Cubs pitching data for just about every pitching rate stat for day/night and home/away over each of the past three years with Clement's, and while his stats showed schizophrenic sample size hiccups in each individual year, his stats for the whole 2002-2004 only partially bear out the bizarre split benefit hypothesis. I'll use fielding independent pitching, a rate stat combining HR, K, and BB, to demonstrate. His FIP was 101% of the average for home Cubs and 91% for night Cubs, showing a relative reverse split, so there's no reason for concern there. His day/night split, on the other hand, was significant, 88% in the day and and 108% at night. His K and BB rates were both slimly better during the day, and his home run rate is substantially lower during the day even though the rest of the Cubs tended to give up slightly more home runs in the day. An interesting thing to note, and something any team that signs him should take into consideration. Ultimately, though, this is a pretty minor concern, since the preponderance of day games with Chicago was already built into his home/away splits and he fared better on the road, rate-stat wise.

2. Kevin Millwood
2004: 3.47, 141 IP
Weighted 3-year: 3.27, 187 IP
Carroll gives Millwood the red light, but I'm not quite as skeptical. He had elbow tendinitis in 2004, but he was fine in 2002 and 2003. His performance really hasn't declined all that much, and I would think he'd be available for a one year deal in the three or four million range, which I'd consider a pretty good deal. He still gets a lot of K's and is very stingy with home runs; his walks increased last season, but not by a whole lot and it was balanced out by an increase in strike outs. He gives up more line drives than is normal, but the margin isn't huge. I think he's worth the risk, at least relative to these other clowns.

3. Odalis Perez
2004: 3.75, 196.1 IP
Weighted 3-year: 3.57, 199.2 IP
A favorite "underrated" player who's actually pretty overrated. Excellent walk rate, lousy strikeout rate, lousy home run rate. He had one brilliant season in 2002; all the evidence points to that being an outlier. The longballs came back in 2003, and in 2004 they stayed. He reduced his walks in 2004, but at the expense of his strike outs. He doesn't give up many line drives, but he doesn't have magic BABIP powers either. Carroll gave him a red light, too.

4. Orlando Hernandez
2004: 3.57, 84.2 IP
Weighted 3-year: 3.53, 107.2 IP
El Duque could do a lot of different things. He's a decent choice, provided he's actually cheap (who knows) and can log some innings (who knows). His peripherals are all over the board, with good seasons achieved through various means in 1998, 2002, and 2004 sandwiching a steady decline from '99-'01.

5. Derek Lowe
2004: 4.26, 182.2 IP
Weighted 3-year: 3.95, 198.2 IP
Groundball extraordinaire who had bad luck last season but isn't all that good to start with. His 2002 was pretty good, but he's probably turned the corner on it. Good control, few strikeouts, very few home runs.

6. Aaron Sele
2004: 4.29, 132 IP
Weighted 3-year: 4.34, 135.2 IP
He doesn't have much of anything left in the tank, but none of the guys below him ever had much in the tank to start with. He can't strike folks out any more, and his walk and HR rates are precisely average. Batters aren't getting many line drives against him, so he's not total scrap heap material. Nonetheless, I'd bet on D.J. Houlton doing better than him, perhaps by a lot.

7. Esteban Loaiza
2004: 4.69, 183 IP
Weighted 3-year: 3.53, 189.2 IP
Had an amazing season in 2003. His home run rate was ridiculously low for U.S. Cellular Field, he had an outstanding walk rate, and his K rate was well above average, the only time in his career his K rate reached average. His 2004, on the other hand, saw his walk rate approach average, his home run rate sink well below average, and his K rate revert to his old balls in play self. There's no reason to expect anything close to 2003 again.

8. Eric Milton
2004: 4.86, 201 IP
Weighted 3-year: 4.36, 132 IP
He's got wins, so he'll probably have dollars. He shouldn't, though.

9. Jose Lima
2004: 4.89, 170.1 IP
Weighted 3-year: 4.59, 112.2 IP
At least, unlike Billy Beane, this guy's @!#% works in the playoffs. Reports indicate he won a raffle at a magic store, getting himself a lifetime supply of smoke and mirrors. I think he used up the whole supply last season.

10. Ismael Valdez
2004: 5.74, 170 IP
Weighted 3-year: 4.81, 158.1 IP
Even Lima's raffle prize couldn't save him now.

An uninspiring group. Clement and Millwood in tandem, though, would be pretty solid.

Comments: Post a Comment

<< Home

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