Interpreting Oliver Perez's 2009
If you've been reading Amazin' Avenue lately, you know we're not particularly enamored with Oliver Perez. This opinion is hardly universal though, due in large part to reliance on statistics like Wins and ERA, which give an incomplete and often misleading view of a pitcher's performance. Omar Minaya and staff, having just lost a bidding war against themselves for Perez, obviously use some if not all these statistics in evaluating players. This failed method can be summarized in a chart courtesy of Paul DePo:

Minaya's offseason thus far has both reflected a bad process in evaluating players and represents a bad process in itself. That's not to say the Mets can only do well through sheer "dumb luck" next year, but Minaya is not effectively working toward a good outcome
Similarly, this chart can represent Oliver Perez's past and future performances and the statistical results of those seasons. Applying the chart to Ollie's '09, we can better predict his results and later interpret them.
Bad Process:
Oliver Perez's pitching, barring a breakout similar to his 2004 season, is a bad process. He throws two pitches, a slider and a fastball, with erratic control. He posts high strikeout rates, but also very high walk rates. This approach also leads to large numbers of flyballs and, accordingly, homeruns. Fielding Independent Pitching (FIP), a statistic that attempts to evaluate a pitcher's performance absent factors beyond his control and express a number similar to ERA, uses just strikeouts, walks, and homeruns. Not surprisingly then, FIP depicts Ollie's ERA as deceptive, scoring him at 4.35 in 2007 and 4.55 in 2008.
Dumb Luck:
The difference between his FIP and ERA, then, can be partially attributed to luck, or factors outside his control. Oliver benefits from the help of a good defense, Carlos Beltran's in particular. Beltran handles the many flyballs and linedrives Perez surrenders. This fielding help mostly explains why Perez's BABIP has been relatively low (.280-ish) in his time with the Mets. The latest in DIPS theory suggests pitchers do have some control over their BABIP, with flyball pitchers having a particular advantage. So Oliver should continue to benefit from his defense and will probably again "beat" his FIP in 2009.
Another luck factor, park, will be in flux for Perez and the rest of the Mets' staff in 2009. Will Citi Field help Perez? Probably, but not greatly, as his hittracker plot suggests the depth and location of his homers aren't particularly suited for Citi's dimensions. Still, the thought that "the Catch" would have been a routine flyball on the track in Citi is an encouraging thought.
Poetic Justice:
Most projection systems see Oliver taking a step back next year. Only Marcels projects him to match his ERA from 2008. Needless to say, if Perez posts an ERA over 4.50, many fans and media members will be calling this new contract a total failure. That would be poetic justice, but is not likely given the aforementioned factors. These projections, which assume a spike in HRs and hits, are still possible, however.
Deserved Success?:
Many people hold out hope that the Oliver Perez of 2004 will show up and hang around for the next three years. Although it's highly unlikely, PECOTA gives Perez a 14% chance of a breakout and 58% chance of improvement in 2009. We'll see, but don't be fooled next year. Focus on the process.
19 comments
|
0 recs |
Do you like this story?
Comments
No analysis is complete
that disregards market dynamics. (I recognize that you have addressed them, in some ways, elsewhere.) This is the best Mets blog going, but I don’t think you’ve been sufficiently attentive to those dynamics in this whole Omar-Lowe-Opie-Sheets saga.
After the first tier starters signed (Sabathia and Burnett), you were really left with Opie and Lowe in the next tier. I totally agree with you that the performance gap between the two is substantial and that Opie hasn’t even really distinguished himself from some of the other starters in the same tier. Opie is an above-the-median starter, and that’s probably it, barring some unlikely-to-happen epiphany.
But once Lowe was off the market, it became clear that the non-Opie alternatives in the 2nd tier all came with big fat “buyer beware” tags because of age-related decline, injury risk, and more commonly both. Perez became the dominant alternative, warts and all, once it was clear that teams are avoiding Ben Sheets like the plague.
The best possible outcome would have been Lowe. And, perhaps Omar (don’t forget the Wilpons) is rightly criticized for not throwing in the extra $ and the 4th year for him. That seems to be the major thrust of this ongoing narrative. I get that. At the same time, however, I don’t think it’s quite fair to characterize Lowe-to-Atlanta as a botched process by Omar. First, signals of Atlanta’s interest were non-existent-to-mild-at-best until just before Lowe signed. Second, and more importantly, I am inclined to believe—based on what we’ve seen this entire off-season with Manny, Adam Dunn, and others—that Omar has some hard-and-fast salary constraints despite the Wilpons’ public denials of such. I believe the team has said that salaries must be at or below last year’s totals, and nothing long-term (i.e., 4 year) under pretty much any circumstances. Given that we’re all inferring process based on what we’ve seen publicly, I think there’s enough out there to warrant granting Omar some benefit of the doubt.
In his time in NY though, Omar’s been pretty good at closing the deal for the best (or perceived best) player available. When the Mets have made their play he’s demonstrated that he’s pretty competent at that part of the job. Constructing a bench? Different story. I believe that given what we’ve seen this off-season that there’s a failrly strong circumstantial case to make that a 4th year for Lowe was an absolute deal-killer for the Wilpons. Did they order the code red? Probably not. But hell, Omar’s been a Manny guy for a long time. I find it hard to believe that the lust has waned so much that he wouldn’t even be willing to kick the tires without some strong disincentives.
As for the notion that he bid against himself for Opie, I think the lesson we just learned is that markets can come out of nowhere. But even without speculating that some mystery team out there perhaps started kicking the tires on Opie, overpaying even when you’re the only suitor is a pretty common fallacy. That doesn’t make it okay, but the fact that it’s nearly universal makes the criticism a little less profound. Yeah, he bid up at the end to scare off other possible suitors and get the guy to FL on time. That makes the final numbers on the contracts Opie and Lowe received tantalizingly close, ramping up the frustration of not getting Lowe. But I think there was a bright line Omar wasn’t going to cross and wasn’t so much about money as the commitment.
"Those who fear disorder more than injustice inevitably produce more of both." -- Rev. William Coffin
While I agree that money was clearly a concern and out of Omar's hands
That doesn’t change the fact he didn’t allocate money particularly efficiently. Like the 2.6 million for Cora, i think 1.6 million on one of the outfielders. Then there were cheaper alternatives for the bullpen, like signing Fuentes or Wood, and saving 3-5 million dollars, instead of K-rod, or possibly getting K-rod for a lesser contract since there didn’t seem to be much of a market for him either, plus overpaying Perez even though there’s no market for him. Not to mention pitchers like Smoltz, Penny, Petite and I think Randy Johnson who signed for incentive laden contracts that are much cheaper than what we’ll be paying Perez. Even though their were clearly financial restraints put on Omar I don’t think that’s an excuse for a so-so off-season.
I don't think these players were actually options for the Mets
Smoltz = Lifelong Brave, kills his rep by taking a one year, cheap deal with Mets. Instead goes to Boston, Braves fans can’t hate.
Pettite = Same as Smoltz but replace Braves with Yanks and Boston with Yanks.
RJ = Explicted stated he wanted to stay on the West Coast.
Penny = Pretty hurt, I wouldn’t back the deal if the Mets had to outpay Boston.
Same for Lowe, the Braves signed him for 15 mil, would he have gone to us for 16 or would it take 17 or more, I doubt it’s that simple as a “beating” the Braves offers.
Ollie comes with a "buyer beware" tag
He walked four or more batters in about 33% of his starts over the past three seasons. I don’t see how locking a guy like that in for three years, knowing that approximately 1/3 of his starts are going to be disasters makes sense. Sheets, Penny, Smotlz, etc… all come with big risks, but at least you can make contingency plans for injury and get off of their contracts if they don’t pan out. The Mets are now saddled with this contract, and if Ollie doesn’t improve his control they won’t be able to eat it until mid 2011.
One nitpick
I like using stats like FIP to suggest what we can expect from Ollie in the future, but Ollie was successful in 2007, even if he was lucky. And I don’t really care if he gets luckier than Marco Jaric – if he puts up good ERAs despite bad FIP’s, I’ll be perfectly happy.
Why does that make HIM successful?
And can you provide any evidence that having an ERA lower than a FIP in the past is an indicator of doing it again in the future?
And I don’t really care if he gets luckier than Marco Jaric – if he puts up good ERAs despite bad FIP’s, I’ll be perfectly happy.
In other words, is there reason to pay for potential luck?
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
A guy who wins the lotto may be lucky and probably won't ever win again, but he was still a successful lotto player
I don’t care that Ollie’s peripherals suggest that he shouldn’t have been as successful in 2007 as he was, he was successful in the real world of Major League Baseball. I’m not sure where I said that having a lower ERA than FIP is an indicator that it would happen again – I merely said that if it does happen again, I’ll be happy.
I love using in-depth sabermetrics to help predict the future, I just think it’s silly to use them to re-invent the past. It’s why I can’t help but laugh at all the hand-wringing over the MVP/HOF votes in recent years. If a pitcher goes 24-4 with a 1.94 ERA, but has awful peripherals that suggest he should have had a 4.24 ERA, I think it’s silly to say he didn’t have a great season just b/c he was lucky (I’m assuming this hypothetical pitcher would have had an insane amount of stranded runners and other assorted ’luck’).
And for what it’s worth, I was against this signing (at least for this amount of money), but now that it’s happened I will be very happy if Ollie bucks the statistical trends and outperforms his projections for each of the next 3 years.
So you'd vote the lottery winner into the lottery Hall of Fame? (Half kidding)
What about defense and park factors and offense? Do you adjust a pitcher’s record and ERA for those things before judging the stats?
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
All hail FIPS, meh
And let’s not start acting like FIPS is some perfect predictor of anything, because if it was, everyone would be using it and there would be nothing to discuss. The mere fact that PECOTA is anticipating a 72% chance of improvement or breakout suggests that at least one indicator to which many people have come to view as a better predictor than most is optimistic regarding Ollie.
As I said in a previous posted comment, this is a paid for potential deal. And, when you consider that Oliie knows and can play in NYC and that several of the other options might or might not be able to handle the concrete jungle, the more I think about this deal, the more comfortable I become with it.
And, how do we know the Mets raised the offer against themselves. I just hate that. Really, as if you know. Just stop it, please stop it. I’ll buy the FIP vs. ERA luck argument. I’ll buy the contract is not worth the predicted value wins. But, let’s not start regurgitating rumors that no one can substantiate as a fact to be used in an analysis. You’re better than that.
In closing, I agree they paid more than I would like. However, they paid for potential on a young player and I prefer that over paying for past performance on a declining skills player. As far as bad contracts go, I still think Ollie has a better chance of reaching his salary in value wins than Wagner ever did in any season with the Mets or that Castillo ever will.
Most projection systems FAIL
The stats are great to a point. But you do have to be aware of their limitations. One of the major limitations is none of them include any scouting data whatsoever. A 26 year old with two plus pitches and inconsistent mechanics is valued exactly the same as a 34 year old with good command of marginal stuff. If both guys “project” the same, I’m paying a few million more for the guy who scouts like a lot better.
In particular, with DIPS, not only is it that recent research shows pitchers have some control over BABIP, but this has been pretty well known all along. There are clearly a percentage who have significant control. It has been argued that DIPS still works pretty well because this skill tends to correlate well with the DIPS stats, strikeouts and walks, because all correlate with stuff. Nonetheless, it is clear that there are at least some pitchers who are badly underestimated by DIPS stats.
If you bring scouting into it, it isn’t too hard to recognize that the fly ball pitchers who fare better than “expected” tend to be those with plus four-seam fastballs. The four-seam fastball tends to be thrown up in the zone, which can lead to a lot of flyballs, and often a high rate of HR and extra base hits as well. But, how much a pitcher gets hurt by the extra base hits isn’t just random luck, just a random percentage of flyballs, it tends to be pretty strongly related to HOW GOOD IS THE FASTBALL.
But, a DIPS based system tends to give credit for all of that to the defense rather than the pitcher. If you believe UZR, Fernando Tatis was an average defensive left fielder, and Murphy and Evans a bit above average last year. Neither UZR, nor the CHONE ratings for the Mets pitchers, are capable of considering that this might have had something to do with a pitching rotation that included Santana, Pelfrery, Perez, and Maine. If you believe UZR, and CHONE, then Oliver Perez projects next year to a 9-10 record and 4.72 ERA on a neutral team; but, putting him in front of this amazing Mets defense, he now projects to 10-9 with a 4.34 ERA. (link)
I don’t really want to knock CHONE or UZR here, they are amongst the best statistical models we have. But, like all models, you do have to be aware of their limits. Not to get too political, but some of the over reliance on statistical models reminds me of the economists who, overly fond of their mathematically and theoretically sophisticated models, remained oblivious as the economy fell to pieces around them, claiming the economy was still fundamentally sound. Because, well, to admit otherwise, would mean the models could sometimes be wrong. At a certain point, you do have to give some weight to what you can actually observe in the real world.
So lets look at what Perez’s real world results the last two season were worth. I’m going to use stats here too, just ones that measure actual results. Here are Perez’s average wins above replacement for the last 2 seasons according to a few of those measures (assuming a replacement SP is a .380 win%, and skipping the math here to convert to this basis):
WPA: 2.9
RE/24: 2.6
tRA: 2.65
WPA gives him credit for changes in win probability for the actual results of each play, RE/24 gives credit for changes in run expectancy for the same play by play data, and tRA is a measure from statcorner.com which is meant to be defensive independent, but also includes percentages of batted ball types (FB, GB, LB, IFF) in addition to SO and BB.
There is also a regressed version called *tRA, which is supposed to be better for projections. This puts Perez at about 2 wins per year over the last two seasons. My concern here would be you are regressing away some of his talent. The point of regression is that you don’t know what is talent and what is luck, so you regress towards league averages. But, remember *tRA doesn’t have access to scouting data. If observation suggests the results are due to talent rather than luck, you might want to regress less. Maybe even a lot less.
If the Mets are valuing Perez as a 2.5 win (above replacement) pitcher, I don’t think that’s unreasonable. And, you have to consider also that costs in NY are a bit higher than some other towns. Every other professional gets a higher salary in NYC to offset higher costs of living and taxes. The going rate per win for ballplayers may be a bit higher as well.
So with all that, while Perez is hardly a bargain at $12M a year, I think the contract is maybe being criticized more than deserved (including by myself in some prior comments). I think most Mets fans, who actually watch games, will rightly believe that Perez is a better option than Randy Wolf at least, despite what some projection models might say.
by acerimusdux on Feb 4, 2009 1:29 PM EST reply actions 3 recs
Great comment
Its very easy to get too accustomed to the discreet and quantifiable nature of statistics. This is a wonderfully lucid explanation of the degree of context that often gets lost in statistical analysis.
by Mark Himmelstein on Feb 4, 2009 4:22 PM EST up reply actions
The Ollie Special
It’s not unfair to say FIP “regresses away” some of his talent. I don’t advocate FIP as a perfect representation of his talent, nor do I think fielding is the only other variable involved. Still, I’m not sure about the quality of the fastball being a huge determinant of HR/FB. In the case of Oliver, for instance, it has been the placement, not the quality, of his fastball that’s the problem. When he gets behind in the count he tends to throw fastballs either way too high or down the middle. Both results are bad, obviously.
Not sure about that 2.5 win projection. He’s more likely to be a shade under 2 which agrees with last year’s tRA. I’m also not sold that the wisdom of the crowd is good determinant toward performance, as they often rely on ERA and the like. Also, the fact that we’re even comparing him to Randy Wolf is telling.
King of the bling come to lay down the evidence//Not George Bush, L-Millz be da president
Perez is just a slightly better version of Dave Mlicki
Mlicki had nasty stuff. In ‘95 at age 27, Mlicki went 9-7 4.01era with 54 walks and 123k’s in 160 innings. Very similar to Perez’s ‘04 season for the Pirates 12-10 2.98era, 81 walks and 239k’s in 196 innings at age 22. Both the Mets and the Pirates were absolutely dreadful teams making their performances stand out that much more.
Those kind of years makes you think you have something developing with this young pitcher who has the kind of stuff most pitchers dream of.
The numbers do indicate that Perez will have a far better career than Mlicki did and he has up to this point. However, the makeup of both pitchers are eerily similar. Perez is capable of brilliance, as was Mlicki who shutout the Yankees in the very 1st subway series game, which only makes them more maddening.
The idea that Perez has a 3yr 36mil dollar deal but has the very real potential of turning out to be a Mlicki type of pitcher does scare me. I will say that Perez is showing a lot more promise than Mlicki and has turned in consecutive plus .500 seasons. Maybe the contract settles him down and he just goes out there and pitches. It is possible that Boras put so much pressure on him to have a winning season that Perez tried too hard to be perfect.
On the other hand, maybe that pressure helped him be as good as he was and that is as far as his potential goes.
not sure that I'd call
123 K’s in 160 innings with a 4.01 ERA at 27 very similar to
239 K’s in 196 innings with a 2.98 ERA at 22.
At best, I’d say they are vaguely similar, in the sense that they are both seasons by pitchers.
I said similiar in makeup,
the numbers are just the year everyone decided they were potential stars while pitching for bottom feeding teams. I wish people actually read before they defame.
BTW, if you were a Met fan in 1995, Mlicki was supposed to be a potential ace
to go along with generation K after that season. Next time I’ll just leave the numbers out since people seem to gravitate to that instead of the actual point.
no, I get the point
and I remember all the talk about how electric Mlicki’s stuff was. I just don’t think Mlicki ever had the success that Ollie has already had.

by 

























