Tuesday, September 20, 2011

Live-Blogging: Terrible Sports Analysis

Presumably because they're folksy, likable, and/or have good anecdotes, TV broadcasters choose to hire sports analysts who rarely understand how to analyze the sports that they used to play professionally. They can tell you, for instance, how to identify a curveball but they can't actually explain why it is or isn't advantageous to bunt or not. (Which wouldn't be so bad, except that they pretend that they know and their reasons are nonsensical, if they're even offered.) And this is hardly a contentious claim to make - every decent sports blog makes this complaint about their local announcers. (And, in fact, there was once a blog devoted entirely to critiquing/mocking bad sports announcers: Fire Joe Morgan)

So, here I am, sitting at home and 1) reflecting on the depressing fact that The Shopping Channel has found someone more qualified than me to write copy for them, while 2) my daughter refuses to go to sleep and calls my name, and I'm watching the Blue Jays play the Angels. And it occurs to me that I should blog about the errors, misunderstandings, and absurdities that the commentary team, Buck Martinez and Pat Tabler, are sharing with the audience on this particular night. (And honestly? The broadcaster, Sportsnet, has no excuse - I emailed them months ago to beg them to hire a stats guy who could offer them good stats and nudge them away from the bad ones that they love so much.)

1st inning

Buck: “[Howie Kendrick]’s had a consistent season.”
One of the annoying things that sports announcers do is confuse the words "good" and "consistent". Because Kendrick has had a very good season, (about 20% above average on offense, and well above on defense) but hasn't been at all consistent. During the 6 months of the season, his monthly batting average has ranged from a bad .231 (September) to an amazing .348 (May); he's hit 6 home runs in two months, and only 0 or 1 in three others; his wOBA shows that his bat has been excellent in three months, almost exactly average in two, and below-average in another. Point is, for a long period he was excellent, then he was very mediocre, and now he's playing badly - that's consistent? (I'm kicking a dead horse at this point, but I wanted to be thorough.)

Pat: “When you have Jose Bautista in front of you, you’re gonna have baserunners.”
This isn't really a dumb statement in isolation. What it does reveal, though, is that these guys are capable of saying really stupid stuff even though they're conscious of the fact that sluggers get RBI - in part, at least - because the players in front of them are good at getting on base. So when they say that sluggers who get out way too often are "producers" because they get RBI - as if the people on base had nothing to do with it - there's really no excuse.

2nd inning

Buck: “I think the most misleading statistic fielding percentage.”
Pat: “The only thing it tells you is how many plays he makes.”
This one could have been good. Buck actually makes a fine point, and one that statsheads have been making for decades: fielding percentage doesn't tell you how many balls the fielder failed to reach but should have, so a good percentage can mask a terrible fielder. And then Pat shows that he doesn't really understand that at all - because what he claims is the "only thing it tells you"? It's actually the exact opposite.

Pat: “Forget about the average, [Mark Trumbo] has been producing.”
Ah, yes - the "producing" that I mentioned earlier. If a good fielding percentage can mask a bad fielder, then good "production numbers" - home runs and runs batted in - can mask a mediocre hitter. The Blue Jays' J.P. Arencibia is a great example, but Trumbo works, too - lots of home runs but a terrible batting average and a low walk-rate. Because, in a league where the average player reaches base 32.5% of the time, Trumbo's 29.6% is worse than about 90% of his peers. Sure, Trumbo hits for so much power that he's a marginally above-average hitter, but "forget" that he reaches base at one of the league's worst rates? That's terrible analysis.

Buck: “[Brett] Cecil’s been much better than his win-loss record.”
This, again, is a reasonable statement that is routinely undermined by Buck's typical expression of love for the pitcher's win-loss record. Buck has, in fact, said in the past that wins are the most important stat for a pitcher. Which is absolutely ludicrous, and here's why: a typical starting pitcher usually only throws about 2/3s of his team's pitches (and, of course, none of the opposing team's), does almost none of the fielding, and either none or very little of the hitting. I don't quite know where to find the number crunching, but i recall that Win Shares credits all pitchers with only 34% of the game's result - and your starting pitcher, again, isn't eating all of those innings by himself. And yet assigning him a win or loss gives him all the credit.

Pat: “[Colby Rasmus] becomes a very dangerous breaking ball hitter.”
This one caught my attention because Pat often makes statements like these - "analysis" that can be, and is, quantified but only with great difficulty, and probably too much difficulty for Pat to bother with. So I looked it up on Fangraphs: Rasmus has been a below-average hitter against breaking balls over his career, while this season he's been awful against sliders but marginally above-average against curves. But "very dangerous"? Not by any measure.

Pat: “They don’t have a stat of plays that should have been made. Yet.”
Well, this is just plain wrong. TotalZone, DRS, and UZR all measure this, though none of them can do so perfectly. Pat doesn't need to agree with their methodology or results, of course, but he should probably know that they've existed for more than two decades. That's right - these things existed when Pat Tabler was still playing.

Pat: “The Angels are a good defensive team. They have athletes all over the diamond.”
The Angels may well be a good defensive team, but their athleticism is really neither here nor there. No one disputes the athleticism of, say, Torii Hunter, but plenty of analysts will point out that his age and declining skills mean that he's become a below-average fielder; likewise, the Jays have some pretty good, fast athletes in Rajai Davis, Brett Lawrie, and Eric Thames, but they're all demonstrably poor defenders. (That said, while Pat's reasoning is faulty, his conclusions are sound - the Angels grade as the 5th best fielding team in baseball, according to UZR, and are much closer to 2nd than they are to 6th.)

4th inning

Pat: “[The Angels] go 1st to 3rd better than anybody in baseball.”
This is tough to verify, so I can't imagine that Pat actually knows this. Like most analysts, he's probably guessed this based on what he's seen - and 'traditional wisdom' - and no other evidence. I couldn't find any stats, but evaluations of base-running on the whole are easy to find: the Angels are slightly above average on both the Speed Score and Baserunning metrics. That doesn't suggest to me, though, that they're likely to go 1st to 3rd better than anyone else.

5th inning

Pat: “At the Major League level, you fail 70% of the time and you’re a super-star.”
Pat's surely pretending that walks don't exist, here - or, maybe, thinking that they count as a "fail" - because this is patently wrong. Of the 25 best hitters in baseball, (according to their wOBA) none reach base less than 35% of the time. And, remember, failing 70% of time means that you're not only not a superstar, but you're one of the very worst at not getting out.

6th inning

Buck: “[Peter Bourjos]’s is just a triple shy of doing something that no Angel has ever done. [Have double-digit doubles, triples, and homers.]”
Pat: “Impressive.”
Well, not really. It's a quirky bit of trivia, I guess, but it certainly doesn't indicate that Bourjos is an "impressive" hitter. (He's slightly above average, and his overall offensive performance is almost identical to that of the aforementioned Trumbo.) What would be better for the player and the team, though surely less "impressive", would be if those 9 triples had actually been homers.

End of 7th inning

Gregg Zaun [between-innings and after-game analyst]: “You can’t put a price on guys like [Mike McCoy] in the organization.”
McCoy is a career .200 hitter with a below .300 on-base average, and he plays multiple positions competently. Basically, he's a replacement-level utility guy and there are dozens just like him. And you can put a price on him - somewhere just above league-minimum.

No comments: