Twitter
Search
« 2019 NFL Mock Draft | Main | An Overdue Apology to Jared Goff »
Tuesday
Dec042018

The Problem with Analytics – An Opinion from a Former Skeptic

“Analytics is just a word… it’s a huge asset to have information and the ability to use it.” Rocco Baldelli, Minnesota Twins manager.

Well when you put it like that, it’s a wonder that we’re steaming in on 2019 and there’s still a large part of baseball’s fan contingent that is completely resistant to the concept of analytics in baseball. But look, I get it. I’m really only a recent convert to analytics, maybe in the last two or three years. And that’s not to say I have abandoned good old-fashioned observation and traditional statistics; only that I’m basically doing what Baldelli says above – availing myself of more information.

Members of the analytics-subscribing baseball fan community have lamented to me the lack of acceptance for analytics in baseball fandom’s mainstream. I hate to be the bearer of bad news, but to a certain degree, they never will be. On a television broadcast, it’s always going to be a battle to supplant batting average and home runs as the go-to stats when talking about a hitter’s ability. Let’s face it, the announcer has a few seconds to introduce the hitter and do a bit of color commentary before the guy steps into the batter’s box. Purely in terms of fitting what you need to fit into that brief window, it’s way easier to talk about a guy’s batting average or on base percentage than it is his wRC+ for the very simple fact that the first two don’t require an explanation.

Is wRC+ a better measuring stick for a player’s offensive value than batting average? Absolutely, and it’s not close. That said, on any given broadcast, what percentage of people know (or care for that matter) how wRC+ is calculated, and what it means for the probable outcomes of a player’s at bat? Add to that ongoing efforts to shorten the length of games, don’t hold your breath waiting for MLB to create windows in its game structure for the announcers to explain it.

That’s not to say, however, that analytics can’t gain a greater foothold in the baseball fan community. I, for one, would like to see this happen, but I am also realistic that there are some pretty big obstacles to overcome if this is going to happen. Fortunately, these obstacles are less complicated than the calculations for some of the metrics. Here they are:

The people: Since I started this article with a person (Baldelli), I might as well start here with one. Baldelli’s approach to it is the right one – the fan-friendly one. He’s basically saying that they’re trying to accumulate as much information as they can in order to put their players into the best positions to succeed. With every single pitch of every single game available in stunning high definition, there is more data out there than ever. If you have access to something that shows you Pitcher A has the most success throwing a certain pitch, but only uses that certain pitch 19% of the time, you have done exactly what Baldelli said.

However, this is not the approach analytics subscribers have taken with skeptics in my experience. God forbid you mention batting average or runs batted in to one of these folks. You might as well have walked into rural Mississippi with a “Vote Hillary” headband on. Yes, new metrics are definitely more reliable and exacting indicators of a player’s offensive performance, and a statistic like batting average which weighs singles and home runs equally is inherently flawed. Yes, runs batted in are heavily dependent on how well the hitters in front of a player get on base. No, you don’t have to be a pedantic ass when talking about it.

I suspect a lot of the condescending smugness is due to residual animosity the intellectual crowd still holds against the jocks from high school who treated them like crap, and being able to demonstrate their intellectual superiority in adulthood is in some way cathartic and blissfully revengeful. I mean, can’t you just take silent pleasure in knowing your paycheck is probably twice or three times what Meathead Malone now makes at the local Dick’s Sporting Goods?

This isn’t the Old West, and this town is in fact big enough for the both of you.

The metrics themselves: Yup. Even though a much more precise method of analysis takes place with the metrics, they often provide data that is at best confusing and at worst useless. Take WAR, or VAR. Each is used to determine how much better a certain player is than a proverbial “replacement” player. Sounds like a valuable tool, right? Sure, except that there are several sources of this metric, and rarely do they align with one another on a player. How are you to know which is reliable?

Take launch angle. Sure, there’s a sweet spot where the majority of home runs are hit (25 to 26 degrees), but that’s absolutely meaningless unless combined with sufficient exit velocity. I get the appeal when Giancarlo Stanton mashes one of his prodigious and majestic dongs into the stratosphere, but in 99% of batted balls, I can imagine exactly zero scenarios where anyone should care what the launch angle was. To be fair, there are hitters out there who say they looked at their average launch angles and decided to focus more on hitting the ball in the air, with impressively positive results, but they could likely have done the same by just acknowledging they hit too many weak grounders and correcting accordingly. Want a cool metric to use instead of launch angle? Hard hit percentage. How often does a guy square up a ball? Hint. The good hitters do it more frequently.

Take defensive metrics. Wow, what a dumpster fire. To be fair, defensive stats like Catch Probability have taken hold on television broadcasts, and while it’s hard to say with any degree of confidence that they’re accurate, it’s pretty damn exciting to watch someone like Kevin Pillar make an absurd play to take a triple away from a guy, and have the graph come up confirming he just made a play with a 3% Catch Probability. Still, the majority of defensive metrics…well, suck. Mike Trout can go from being a profoundly negative defender in center field, to Gold Glove caliber the next season. For much of the first half of 2018, the Rockies’ Nolan Arenado, probably one of the three or four best defenders at third base of the last 25 years, was rated as a negative defender. Manny Machado was metrically one of the worst defenders in baseball. Manny Machado!

Even when defensive metrics make sense on the surface, red flags abound. Take Matt Kemp, rated analytically as one of the worst outfield defenders in the history of history, behind only Old Four Toes McGillicutty of the Port Philbanks Plankwalkers from 1857 to 1861. And he had a peg leg. There were some dWAR corrections this year, but before that, defensive metrics would have you believe that Kemp would need to win the offensive triple crown to have any value. In fact, he could get the game winning hit in the ninth inning of ten straight games, and never have a single ball hit to him in left field, and still over the course of those ten games come out to be a negative player because his dWAR is so bad.

My advice? Stick to catch probability, and the infield equivalent. Route efficiency is another fun one, and since the geniuses behind it decided to put the function in the nomenclature, it doesn’t require explaining. And if you really want to have a good time looking up a player’s defensive ability, Fangraphs’ Inside Edge Fielding is a superb and fun resource. And in case you were wondering, yes, a player has indeed made a play rated as “impossible.” His name is Andrelton Simmons, and seeing as you can simply observe that he’s beyond incredible in the field, it passes the sniff test.

The false narratives: Whoa boy… this one is probably my favorite. With the rise of analytics, and the increased use of in-depth statistical probability, certain truths about baseball have come to light. For example, the sacrifice bunt has all but disappeared from the game except by pitchers. Why? Well, the numbers bear out that giving up an out to advance a runner is less likely to result in a run than letting the (non-pitcher) batter try to hit. Jay Bell had like a billion sacrifices from the two-hole one year for the Pirates or Diamondbacks… some inconsequential franchise anyway…(easy, guys, I’m kidding). And you know what? You’ll never see that again. Not only does it make you less likely to score, it cuts off at the knees your likelihood of scoring multiple runs in an inning. There is one exception. You are slightly more likely to score with runners on second and third with one out, than you are with runners on first and second and nobody out. This, of course, does not take into account who is hitting. It’s merely an aggregate calculation. Well, while the bunt has all but disappeared, the narrative has become among the aforementioned “people” that if you bunt, you’re an idiot. Is that really an attitude to take if you’re trying to get adopted by the mainstream? Also, and while I realize the bottom of the ninth inning in a tied game is a very specific instance, it’s one that happens frequently enough and it’s one in which only one run is needed to score. If you can use the bunt to ensure you have one of your better hitters coming up in a situation where only a fly ball is needed to win the game, why would you not at least consider that? I’m not talking about having Mookie Betts bunt to give Blake Swihart a chance to win the game. But if you reverse the names, doesn’t it make a little sense?

Here’s another example. Not too long ago, someone came up with this characterization of the three true outcomes player. Basically, a guy would either hit a home run, walk, or strike out most of the time. Adam Dunn would be the prototype. Another guy who would fit this model, at least until 2018 when he cut down on his strikeouts (and lo and behold, became a more valuable offensive player) is Joc Pederson. Yeah. Dunn was a very valuable hitter. Pederson can be too. But for every Dunn, there are dozens of Ryan Schimpfs. Out of this three true outcomes fad came the notion that strikeouts don’t matter because all outs are created equal. God forbid you suggest that a batter hit behind the runner when there’s a guy on second with nobody out. You might as well have suggested beef wellington to a vegan. Meanwhile, the Kansas City Royals went to two World Series, winning one, with an offensive philosophy of putting the ball in play. The two most recent World Series winners, Boston and Houston, both ranked among the teams who struck out the least in their winning seasons. It’s simple. A strike out, unless the catcher fails to hold the pitch, is always going to be an out. A batted ball has a round about 30% of becoming a hit. And this doesn’t even take into account having to make a fielder make a play. Because for every human vacuum cleaner like Andrelton Simmons, there’s a statue like Corey Seager who can barely range a foot to either side.  

Lastly, if I hear one more thing about spin rate, I might blow a fuse. Sure, if you want to know who throws a fastball that rises (even if there’s no such thing), or who throws one that sinks, you can look at spin rate. But honestly, if you’re trying to determine who is an effective pitcher, something like “out rate” will give you a better indication than spin rate.

I’ll leave it there while my blood pressure is still at manageable levels. Thank you for reading.

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments

There are no comments for this journal entry. To create a new comment, use the form below.

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>