As you'll see from our earlier post, the Nachoman and I are beginning a conversation on stats vs. gut feelings. I got the idea from this after Kansas City Manager Trey Hillman let home run prone Kyle Farnsworth face Jim Thome, even though Thome hits lots of home runs and has always had his way with righties like Farnsworth. In contrast, Thome is quite pedestrian against lefty pitchers. I agree with the Nachoman that it's not that useful to rely on limited sample sizes -- looking at how one pitcher as done against a batter during the last few years, for example. Twenty at bats simply don't give us enough information. That's a favorite strategy of Tony LaRussa; he likes to know how each of his batters has done against every pitcher on the opposing team, and vice versa. I don't think that's very useful.
I do strongly believe, however, in looking at platoon splits. That brings us back to Mr. Thome. In his career, Thome has hit .296/.431/.620 (AVG/OBP/SLUG) against RHP – against lefties, he has hit .240/.342/.442. Can you tell the difference? In that case, stats clearly show that it's a much better decision to send in a lefty to face Mr. Thome. It was the bottom of the 8th, Royals up one. Thome was hitting with two outs and two on. If you get him out, you have a good chance to win the game. If he gets a hit, you start fresh (if he gets a homer, as he did, you lose the game).
Yes, it's important to consider how hot or cold a pitcher is. Hillman decided in this case that Farnsworth was "hot" because he was "throwing strikes." Unfortunately, he was also throwing hittable strikes. And again, even if a lefty is cold, his strikes -- across Thome's long career -- have given the slugger more trouble than a lefty's strikes. See what I'm saying?
Finally, I want to address the Nachoman's "Dandelion effect." He argues that a defense will be sluggish after having to stand around and wait for the pitching change. I think that's nonsense. Sure, defensive lapses may happen on occasion after a pitching change. However, they also happen during the normal flow of the game, and I don't think they're more common after mid-inning pitching changes. Major League players are paid millions to pay attention and field their positions. Most of them know that if they stop fielding somebody else will get their spot (that is, unless the awful fielder can hit .300 and have 30+ homers a year... then all defensive lapses are excused).
Stats can teach manager a lot -- it can even teach them to take it easy and not try to do too much. If Hillman wants to completely ignore stats and manage with his gut, I can't stop him. But I am convinced he'll be hurting his team in the long-run.
Subscribe to:
Post Comments (Atom)
1 comment:
Only semi-relevant here, but perhaps interesting, is the concept known as "giss" (or "jizz," as it's sometimes spelled). The idea began with wartime aviation, but it's used today in birding: identification of a species not by conscious and specific recognition of individual field marks, but by subconscious impression of the bird's entire shape, size, coloration, and motion. It's a bizarre idea for a novice birder, but an experienced birder can in fact tell at a glance whether he's looking at a mockingbird, even if he doesn't hear a call or spot the tell-tale white wing patches; the general shape and behavior of the bird simply shrieks "Mocker!" to the veteran, though he may not be able to rationally justify the I.D.
So basically, yeah, I think a veteran manager may well have enough unconscious information to judge a pitcher's likelihood of success, even if he doesn't have the stats in front of him.
Of course, this is the Royals organization we're talking about, so they might well look at a mockingbird and I.D. it as an emperor penguin.
Post a Comment