The Tigers Dirks Problem

Andy Dirks is a fairly easy Tigers target to pick on this year. As the strong side of the Tigers left-field platoon, Dirks has mustered only a .242/.307/.344 batting line – a far cry from his 2012 .322/.370/.487 line. Dirks was never a blue chip prospect and isn’t regarded as a tremendous “tools” guy, he was the kind of prospect who had to prove himself through production time and again in order to earn a roster spot or playing time. That’s probably still the case, and one would imagine that Dirks’ job will be in extreme jeopardy come next March no matter what happens over the season’s final months.

Jul 4, 2013; Toronto, Ontario, CAN; Detroit Tigers left fielder Andy Dirks (12) doubles in the first inning against the Toronto Blue Jays at the Rogers Centre. Mandatory Credit: John E. Sokolowski-USA TODAY Sports

We have mentioned here before, and will continue to mention, that Dirks is actually adding value with his legs and glove this season. Overall, he has been worth (according to Fangraphs) 1.1 wins over replacement so far in about 60% of a season’s worth of plate appearances. The bat has been barely above replacement level this season, so almost all of that value has been provided by legs and glove. Dirks put up a .703 OPS in 2011 and an .857 OPS last year, so this season’s .660 has been disappointing. Though his strikeout rate is up a tad, most of this “diminished production” can be chalked up to a low BABIP – and the fact that last year’s numbers were just not representative of who Dirks is as a hitter. His BABIP in 2013 is actually higher than in his rookie year, though his OPS is lower. Strikeouts play a bit of a role there, but mostly that’s hit trajectory. Dirks hit more fly balls in 2011, which meant a higher slugging percentage. More fly balls tends to mean lower BABIP, though it can still be desirable if it means a lot of bombs (as for Prince Fielder). More line drives means a higher BABIP, as does more grounders – at least relative to flies. And that is the puzzle: based on hit trajectories, Dirks BABIP should be high. Liners are up, grounders are up, flies are down. He’s getting more infield hits on those grounders than before, too – and his HR/FB rate hasn’t taken a dive. According to his “expected batting average on balls in play” or xBABIP, Dirks should have had only a .292 BABIP in 2011 and a .334 BABIP in 2012 (when in fact it was .375). His BABIP this year should be .338 – not .287.

Anyway, the idea is that – since xBABIP is a good measure of what a batter is actually doing (power aside) and predictor of future performance – Dirks’ is getting “unlucky” and should have numbers that look better than they do right now. The difference is pretty big, Dirks “should have” an extra 13 or 14 hits right now. If that doesn’t sound like a big deal, it would elevate his batting average to .285 and his on-base percentage to .349. And, although BABIP ignores home runs entirely, some of those extra hits would have been doubles or triples. If you crunch the numbers, a 10 point increase in BABIP tends to lead to about a 3 point increase in isolated power. So… Dirks’ ISO should be .116 rather than .101. Tack that on to his shiny new .285 batting average and he’d have a slugging percentage of .401 and an OPS of .750. That doesn’t look like such a crappy hitter – and it’s more like what we expected from Dirks going into the season and what we should expect going forward than what we have seen thus far in 2013.

Those are reasons to feel good about Andy Dirks as the Tigers left fielder. We can reasonably expect league average sort of offensive production from him, not replacement level production. Whew.

Unfortunately, Dirks bad OPS isn’t really “The Tigers Dirks Problem”. The problem is that Andy Dirks doesn’t seem to be able – this year, last year or the year before – to perform in clutch situations. I mentioned a week ago that Torii Hunter and Omar Infante also seem to have this problem, over long enough careers to suggest that something is genuinely off. I should clarify that this is not a simple difference in batting average with RISP and the like – what I mean is that there are consistent differences between the batter’s “win probability added” and his “leverage adjusted win probability added” which means that the batter does the sorts of things that are beneficial for the team when it doesn’t count much but does not do those things when it does. That could mean that he struggles to bring runners home from third with fewer than 2 outs, it could mean that he can’t hit closer-caliber stuff, it could mean that he only seems to hit into double plays that end innings with his team down a run. Dirks is not a great batter to begin with, he has a WPA/LI of -1.00. His raw WPA is -3.74. Even when he was hitting well last season, his WPA was negative suggested that he hurt the team with his performances a tad more than he helped. It’s normal for guys who aren’t great hitters to have negative WPAs and WPA/LIs: Brandon Inge had a career WPA/LI of -11. The statistic is, in essence, calibrated to 0.00 because of the nature of a win. If an average batter has a zero WPA, a zero WPA/LI and contributes 2 WAR over a full season’s worth of plate appearances – Dirks -1 WPA/LI (in a season and a half worth of PAs) is equivalent to being worth something like 1.4 WAR offensively, which is OK since he adds to that with good defense and baserunning – it would make him basically what the normal WAR calculation shows, very close to an average left fielder.

WPA/LI is a much better measure than WPA of how well a guy is actually hitting, especially in small samples. However, WPA is a much better measure (the measure to end all measures, in fact) of the value that the player is actually contributing to the team at the plate. If you use WPA, putting Dirks’ contribution on that WAR scale, he is half a win below replacement level at the plate which basically cancels out his average value in the field to make him a replacement level player overall. Whether or not Dirks can, in fact, hit when it counts is HUGE. And let me clarify another thing – it isn’t so much that Dirks is worse in the clutch than Infante, Hunter or many other guys – it’s that he isn’t as good of a hitter to begin with, so the relative drop in the clutch matters more. If you want to frame this using stats other than the “pure” win-probability added, Dirks over his career has a .552 OPS in situations defined as “high leverage” compared to an OPS of nearly .800 in all other situations. His numbers are bad with 2 outs and with runners in scoring position, they’re bad in tie games and “late and close”. They’re bad against power pitchers and bad against relievers. In the 9th, his OPS is .432 and it’s worse in extra innings. Compare that to Victor Martinez, a guy who over the span of more than a decade in the bigs has a better WPA than WPA/LI: Overall career OPS, .830. Vs. power pitchers, .859 – vs. relievers .819, in the 9th inning .842, in extra innings .936. 2 outs with runners in scoring position? An .824 OPS. An OPS of .898 in high leverage situations compared to .813 in all other situations. Martinez is a better overall hitter, sure, but that isn’t what this is about – it’s that Dirks seems to wilt under pressure whereas Martinez does not.

To give Dirks the benefit of the doubt, he hasn’t played nearly as long as Victor Martinez. Even Martinez has had seasons in which he didn’t hit well in the clutch. All sorts of splits, including these, are small samples of a small sample. It’s possible that Dirks only looks like a bad clutch hitter due to bad luck or that he’ll improve this aspect of his game over time. But… if he doesn’t, and if what we have seen from Dirks in the field and on the basepaths as well as the career .741 OPS is representative of the true Dirks then because of the atrocious clutch hitting Andy Dirks adds no value (over replacement level, for what that’s worth) whatsoever to the team.