### Saturday, September 11, 2004

## Odalis Perez' "luck"

So almost any Dodger fan or anyone who read Joe Sheehan's column today will tell you that Odalis Perez has been plagued by bad luck this year. He's racked up a ton of no-decisions and has a 6-4 record despite a superb 3.22 ERA. He certainly hasn't had much luck in getting credited with wins this season, and Sheehan notes that he has the seventh-worst run support in the NL. Some of that lack of run support should be attributed to Perez himself-- he's a terrible hitter, and his OPS ranks 327th among the 356 NL players with 25 or more at bats. Still, that's pretty bad luck in the win-loss column.

But is luck in the win-loss column for an individual pitcher that significant? It depends on who evaluates the pitcher. It wasn't long ago that many general managers appeared to base personnel decisions on stats like RBI and pitcher's win-loss records. Today, however, GM's who use those methods are a dying breed. So if you're wondering how his paltry 6 wins will effect his chances of getting top-dollar in free agency this offseason, the answer is probably between not much and not at all.

But luck

Let's first consider runs created versus runs allowed. Using the most basic formula for runs created (((Total Bases)*(Hits + Walks))/(At bats + Walks)), Odaliz Perez has let batters create 77.02 runs. Since he's given up only 66 actual runs, Odalis has had some significant luck. If you count those extra runs as earned runs, Perez' ERA increases to 3.81. Now, this might not be entirely fair since Perez is one of the best pitchers in baseball in forcing double plays. The Baseball Prospectus Double Play Rate Stats indicate he's created a net 7.57 double plays, 15th best in baseball. But if we subtract 7.57 from his hits + walks total, his runs created total only drops to 74, and those 8 extra runs, if earned, would increase his ERA to 3.65. He hasn't been particularly good at holding runners, either, with his opponents stealing bases at a 72% clip (71% is league average).

Now, some of you might look at the above and think Perez is just really good at pitching out of jams. To a certain extent that's accurate: this year, his opponents hit .269/.308/.452 with the bases empty but only .203/.259/.370 with runners in scoring position. But there's no data I know of indicating that a pitcher's ability to pitch out of jams in one year is a good indicator of future performance with runners on. In fact, from 2001-2003 Perez' opponents hit .238/.284/.359 with none on and .279/.324/.452 with runners in scoring position. In evaluating Perez' future, it doesn't make much sense to assume he'll continue to pitch as well as he has this year with runners in scoring position.

What's more is that Odalis Perez has probably also been lucky in terms of his defensive converting balls in play into outs. According to The Hardball Times, Odalis Perez' Fielding Independant Pitching, a statistic which measures a players expected ERA based on plate appearances that don't result in balls in play, is 4.24. Since the frequency at which a pitcher allows home runs, issues walks, and records strikeouts is established as the most accurate predictor of future success, this means Perez has been quite lucky. In fact, I compared the ERA and FIP of all 19 of the prominent free agent starting pitchers mentioned in Peter Gammons' column on the class of 2005.Perez easily had the largest FIP - ERA at 1.02 (Eric Milton was next at 0.79, with Aaron Sele in third at 0.72).

In fact, the trend for Perez in this area of his game is probably negative. Using the Davenport Translations at Baseball Prospectus, which adjust for a number of factors like park, era, and opposition, there's a pretty clear downward trend for Perez:

His strikeout rate this year is an all-time low, his home run rate is an all-time high, and his walk rate is spectacular but not close to the level of his super-season in 2002. This data illustrates that 2002 was probably his fluke season (as was suspected by many at the time), rather than his poor season in 2003.

Now, walks, home runs, and strikeouts are not the end-all be-all of pitching statistics. Some pitchers are more adept than others at creating outs on balls in play, though this skill is not nearly as significant or reliable as BB/K/HR. The Hardball Times keeps track of what percentage of balls hit in play against a pitcher are line drives and what a pitcher's groundball to flyball ratio is. JC Bradbury of The Sabernomics blog did a study based on those numbers and came up with a formula for a pitchers expected batting average on balls in play (BABIP). Since Perez' batted ball types are all better than league average, one could expect a better than league average BABIP. Plugging Perez' numbers into Bradbury's formula, I came up with an expected BABIP of .279, only slightly better than his actual BABIP of .272. If we adjust our expected ERA using Runs Created to account for that handful of extra hits we should expect, we get a 3.81 ERA. That's pretty good, but it still has a whole lot to do with the Dodgers defense. If we adjust it again so that the formula uses a league average defense, Perez' BABIP becomes exactly league average, .298, and his expected ERA becomes 4.22, and if we decide that he would have half as many net double plays with an average defense that becomes 4.30. That's not a bad ERA; it's better than the NL average of 4.34.

Odaliz Perez is certainly a valuable pitcher since he can be expected to be somewhat above average and log a good amount of innings. But he's definitely not the ace his 3.22 ERA would have you believe. His teammate Jeff Weaver would have an expected 3.49 ERA with the Dodgers defense and 3.88 defense adjusted ERA using the same formula, and he's not often called an ace. Perez always could get better, but there's no evidence that suggests a substantial improvement is imminent. So if Perez has a 3.22 ERA this year and a 3.55 ERA over the past three years but can reasonably be expected to provide an ERA of 3.70-3.90 with a Dodgers team with the best infield in baseball or an ERA of 4.15-4.35 with average surroundings, isn't it likely that he'll be overvalued in the free agent market? Kevin Millwood posted a 4.86 ERA this year and a 4.01 ERA last year. Using the same formula for expected ERA, Millwood would have had a 4.09 ERA with an average defense and average home park this year. Between Millwood's 9-6 record, his 4.86 ERA, his elbow injury, and his perceived old age (he turns 30 in December; Odalis Perez turned 27 this past July) Millwood will almost certainly be available for a cheaper and more low-risk contract than Perez. Now, Millwood is just an example, and some actual scouting evaluation would certainly help in determining who's likely to perform next year, but if his elbow injury only requires some rehab time before being back to a normal there's a good bet that he'll be a good value in free agency and a better value than Odalis Perez. And if, say, Victor Santos could be acquired from the Brewers on the cheap, as I suspect might be the case, than the Dodgers could get great value in rebuilding their pitching staff after Penny and Weaver.

Odalis Perez' luck isn't nearly as bad as it's made out to be, meaning that if Paul DePodesta sees this the same way I do there's a good chance that next year the Dodgers will be lauded for getting lucky on letting Perez go and picking somebody else up off the virtual scrap heap. Branch Rickey's famous quotation-- "Luck is the residue of design"-- might not be accurate in talking about pitchers, but it probably

But is luck in the win-loss column for an individual pitcher that significant? It depends on who evaluates the pitcher. It wasn't long ago that many general managers appeared to base personnel decisions on stats like RBI and pitcher's win-loss records. Today, however, GM's who use those methods are a dying breed. So if you're wondering how his paltry 6 wins will effect his chances of getting top-dollar in free agency this offseason, the answer is probably between not much and not at all.

But luck

*is*pretty important in some areas. For instance, in evaluating a player's expected contribution to a team you have to evaluate how much of the player's past productivity derived from performance and how much derived from luck.I think that, while Odalis Perez' win-loss luck is pretty bad, his overall luck is quite good.Let's first consider runs created versus runs allowed. Using the most basic formula for runs created (((Total Bases)*(Hits + Walks))/(At bats + Walks)), Odaliz Perez has let batters create 77.02 runs. Since he's given up only 66 actual runs, Odalis has had some significant luck. If you count those extra runs as earned runs, Perez' ERA increases to 3.81. Now, this might not be entirely fair since Perez is one of the best pitchers in baseball in forcing double plays. The Baseball Prospectus Double Play Rate Stats indicate he's created a net 7.57 double plays, 15th best in baseball. But if we subtract 7.57 from his hits + walks total, his runs created total only drops to 74, and those 8 extra runs, if earned, would increase his ERA to 3.65. He hasn't been particularly good at holding runners, either, with his opponents stealing bases at a 72% clip (71% is league average).

Now, some of you might look at the above and think Perez is just really good at pitching out of jams. To a certain extent that's accurate: this year, his opponents hit .269/.308/.452 with the bases empty but only .203/.259/.370 with runners in scoring position. But there's no data I know of indicating that a pitcher's ability to pitch out of jams in one year is a good indicator of future performance with runners on. In fact, from 2001-2003 Perez' opponents hit .238/.284/.359 with none on and .279/.324/.452 with runners in scoring position. In evaluating Perez' future, it doesn't make much sense to assume he'll continue to pitch as well as he has this year with runners in scoring position.

What's more is that Odalis Perez has probably also been lucky in terms of his defensive converting balls in play into outs. According to The Hardball Times, Odalis Perez' Fielding Independant Pitching, a statistic which measures a players expected ERA based on plate appearances that don't result in balls in play, is 4.24. Since the frequency at which a pitcher allows home runs, issues walks, and records strikeouts is established as the most accurate predictor of future success, this means Perez has been quite lucky. In fact, I compared the ERA and FIP of all 19 of the prominent free agent starting pitchers mentioned in Peter Gammons' column on the class of 2005.Perez easily had the largest FIP - ERA at 1.02 (Eric Milton was next at 0.79, with Aaron Sele in third at 0.72).

In fact, the trend for Perez in this area of his game is probably negative. Using the Davenport Translations at Baseball Prospectus, which adjust for a number of factors like park, era, and opposition, there's a pretty clear downward trend for Perez:

Age | H/9 | HR/9 | BB/9 | SO/9 | K/BB | |

1999 | 22 | 9.3 | 1 | 4.7 | 6.8 | 1.45 |

2001 | 24 | 9.6 | 0.6 | 3.5 | 5.5 | 1.59 |

2002 | 25 | 7.9 | 1 | 1.1 | 5.1 | 4.48 |

2003 | 26 | 9.1 | 1.3 | 2 | 5.7 | 2.85 |

2004 | 27 | 9.6 | 1.3 | 1.8 | 4.7 | 2.61 |

His strikeout rate this year is an all-time low, his home run rate is an all-time high, and his walk rate is spectacular but not close to the level of his super-season in 2002. This data illustrates that 2002 was probably his fluke season (as was suspected by many at the time), rather than his poor season in 2003.

Now, walks, home runs, and strikeouts are not the end-all be-all of pitching statistics. Some pitchers are more adept than others at creating outs on balls in play, though this skill is not nearly as significant or reliable as BB/K/HR. The Hardball Times keeps track of what percentage of balls hit in play against a pitcher are line drives and what a pitcher's groundball to flyball ratio is. JC Bradbury of The Sabernomics blog did a study based on those numbers and came up with a formula for a pitchers expected batting average on balls in play (BABIP). Since Perez' batted ball types are all better than league average, one could expect a better than league average BABIP. Plugging Perez' numbers into Bradbury's formula, I came up with an expected BABIP of .279, only slightly better than his actual BABIP of .272. If we adjust our expected ERA using Runs Created to account for that handful of extra hits we should expect, we get a 3.81 ERA. That's pretty good, but it still has a whole lot to do with the Dodgers defense. If we adjust it again so that the formula uses a league average defense, Perez' BABIP becomes exactly league average, .298, and his expected ERA becomes 4.22, and if we decide that he would have half as many net double plays with an average defense that becomes 4.30. That's not a bad ERA; it's better than the NL average of 4.34.

Odaliz Perez is certainly a valuable pitcher since he can be expected to be somewhat above average and log a good amount of innings. But he's definitely not the ace his 3.22 ERA would have you believe. His teammate Jeff Weaver would have an expected 3.49 ERA with the Dodgers defense and 3.88 defense adjusted ERA using the same formula, and he's not often called an ace. Perez always could get better, but there's no evidence that suggests a substantial improvement is imminent. So if Perez has a 3.22 ERA this year and a 3.55 ERA over the past three years but can reasonably be expected to provide an ERA of 3.70-3.90 with a Dodgers team with the best infield in baseball or an ERA of 4.15-4.35 with average surroundings, isn't it likely that he'll be overvalued in the free agent market? Kevin Millwood posted a 4.86 ERA this year and a 4.01 ERA last year. Using the same formula for expected ERA, Millwood would have had a 4.09 ERA with an average defense and average home park this year. Between Millwood's 9-6 record, his 4.86 ERA, his elbow injury, and his perceived old age (he turns 30 in December; Odalis Perez turned 27 this past July) Millwood will almost certainly be available for a cheaper and more low-risk contract than Perez. Now, Millwood is just an example, and some actual scouting evaluation would certainly help in determining who's likely to perform next year, but if his elbow injury only requires some rehab time before being back to a normal there's a good bet that he'll be a good value in free agency and a better value than Odalis Perez. And if, say, Victor Santos could be acquired from the Brewers on the cheap, as I suspect might be the case, than the Dodgers could get great value in rebuilding their pitching staff after Penny and Weaver.

Odalis Perez' luck isn't nearly as bad as it's made out to be, meaning that if Paul DePodesta sees this the same way I do there's a good chance that next year the Dodgers will be lauded for getting lucky on letting Perez go and picking somebody else up off the virtual scrap heap. Branch Rickey's famous quotation-- "Luck is the residue of design"-- might not be accurate in talking about pitchers, but it probably

*is*accurate in talking about personnel decisions.
Comments:

<< Home

This is an excellent blog. Keep it going.You are providing

a great resource on the Internet here!

If you have a moment, please take a look at my average nfl salary site.

Have a great week!

Post a Comment
a great resource on the Internet here!

If you have a moment, please take a look at my average nfl salary site.

Have a great week!

<< Home