Batter Pitcher Match-Up Data
Yesterday, the Mariners faced a left-handed starting pitcher, and one who is traditionally very tough on left-handed batters. Opposing LHBs are hitting .149/.171/.216 against Harrison this year, compared to the .343/.387/.537 mark that RHBs are putting up, so he basically turns lefties into the equivalent of an average hitting pitcher. So, Eric Wedge put both Dustin Ackley and Mike Carp on the bench, and after the game, he noted that he was trying to put the team on the field that gives them the best chance to win. Giving them a day off against a very tough left-hander is entirely defendable, even in a rebuilding year, especially when you’re trying to stop a losing streak.
So, I don’t have any problem with Wedge deciding to rest those guys yesterday. It’s just that the rest of his decisions make no sense.
Miguel Olivo, right-handed hitter, was on the bench. Left-handed hitter John Jaso started at DH and hit leadoff. John Jaso, he of the career .172/.301/.247 line against southpaws, was chosen to get the most plate appearances of anyone in the line-up against a guy who is absolutely murder on left-handed batters. Obviously, we’ve been asking for more playing time for Jaso and less playing time for Olivo, so it might seem weird to point out displeasure with Wedge deciding to play Jaso over Olivo, but a major league manager should understand basic things like platoon advantages.
Instead, it appears that Wedge was paying attention to a different kind of number – batter/pitcher match-up data. Here’s Greg Johns, before the game, on his blog:
Interesting lineup today for the Mariners against the Rangers, with John Jaso leading off and Chone Figgins batting second and playing left field, while Dustin Ackley and Mike Carp have the day off.
Why? It appears manager Eric Wedge is playing the percentages, with Ackley 0-for-15 in his career against Rangers starter Matt Harrison and Carp 0-for-10.
Figgins is 4-for-14, while Alexi Liddi is 3-for-6.
Here’s Wedge’s quote about why Ackley’s not playing:
It’s just a matchup situation,” Wedge said. “And he’s still trying to find it. We’re trying to put the best lineup out there to win. When this kid throwing against us is on, he’s pretty good. So we want to give ourselves the best chance to win today. That’s that fine line that you’re walking.”
And finally, here’s Dustin Ackley talking about why he thinks he was on the bench after the game.
“I’m just playing it day by day,” Ackley said. “I know Harrison is throwing today and I haven’t had much success against him. I’m sure that’s a lot of the reason. I know [Wedge] likes the matchups a lot and what you’ve done off guys in the past.”
When a strange looking line-up was posted, Johns looked at the batter/pitcher match-up data and realized that the decisions looked like they were influenced from those numbers. When asked about it, Wedge cited the match-up. When Ackley was asked about it, he referenced that he knew Wedge puts a lot of stock into specific batter/pitcher history.
This is a problem, because specific batter/pitcher match-up data is worthless. It literally has no predictive value whatsoever. This isn’t just some gray area where Wedge is referencing information that outsiders don’t have access to that might not line up with the numbers – this is the Mariners manager relying exclusively on numbers, but not understanding which ones matter and which ones don’t.
This issue has been extensively studied. There’s an entire chapter devoted to it in The Book, much of which is available for you to read through Amazon without even going through the hassle of buying a paper copy. This is freely available information, and the conclusion is indisputable – knowing what a hitter did against a pitcher in the past will tell you nothing about what a player will do against that pitcher in the future. You can argue against it all you want, but the data is overwhelmingly conclusive – batter/pitcher match-up data is useless.
Not useless, though, are platoon splits. Left/right match-ups are probably the most significant type of split data, and a manager could do a decent job of selecting who to play on a given day by simply maximizing the platoon advantage. If you literally just started all of your right-handers against a lefty and vice versa, you’d probably do a better job of setting line-ups than a human looking at batter/pitcher match-up data. Of course, the ideal is to incorporate all information and not just select on one variable, so we’re not arguing that every batter should be platooned every night, but if you’re going to choose between an LHB and an RHB against a guy like Matt Harrison, you better have compelling reasoning for why you’re choosing the lefty. And then, if you choose the lefty, you should probably hit him ninth. Instead, Wedge hit Jaso first, essentially giving Harrison three or four free outs at the top of the order.
Miguel Olivo wasn’t going to win the game for the Mariners yesterday, so maybe this seems like a minor issue that didn’t actually have much of an influence on the outcome of the game. But, it’s impossible to look at the stated reasons for making out yesterday’s line-up and not realize that Eric Wedge’s decision making process is completely broken. There are some basic truths about the sport that he’s unaware of. Eric Wedge stated with a straight face that he was trying to put the best team on the field to win yesterday’s game, and then proceeded to put John Jaso in the line-up against a southpaw and hit him at the top of the batting order. This is Mike Hargrove bringing in Julio Mateo to get a groundball all over again. This is the kind of decision that is just blindingly wrong, and the organization shouldn’t have to put up with.
The M’s need more talent, but they also could use a manager who actually understands how to play the percentages. When they get around to trying to win games every night, Eric Wedge should not be the guy deciding who gets to play.