Let’s do a thought experiment.
Suppose that Mike Trout was injured for nearly the entire first half of the season. Say it happened on a freak outfield collision on opening day, and that he fractured a rib in such a way that his full recovery was never in doubt but that doctors told him he absolutely positively could not play until fully healed.
Suppose that Trout returned two weeks ago, and looked entirely himself, hitting a home run every three games, getting on base 40% of the time, and playing superb defense.
Should he be an All-Star?
See the other side of our DRB Debate, as John Ford tells us to stop whining.
If you said yes, that a helathy Mike Trout should be an All-Star even if he missed most of the first part of the season, because we knew who Mike Trout is (the best baseball player in the world), and the All-Star Game is about putting the best players on the field, then you maybe don’t think the Blake Snell snub is much of a snub at all.
Snell ranks 13th among American League starters by FIP, 12th by xFIP, and his 116 innings pitched put him 11th. He’s a fantastic pitcher, but there are roughly ten fantastic pitchers ahead of him in the running, and with every team getting a representative, some deserving guys get left out.
But hold up, nerd. That is a lifeless way to look at baseball. Blake Snell leads all American League starters with a 2.09 ERA.
Yes, you should care about ERA.
Statistics are evaluated on how well they are at measuring three different things, and when choosing which statistic to care about, you first have to decide, philosophically, what you want to measure.
- Description—does it accurately tell you what happened?
- Prediction—is it good at telling you what’s going to happen?
- Description of true talent—oh boy this one’s a can of worms.*
*Is a good prediction the best description of true talent? Or does true talent itself fluctuate on a day-to-day basis, in a way that looks like noise/luck to our blunt instruments? This is the most interesting philosophical discussion in here, and I’m only going to nod at it.
With only a half-season sample, FIP is better than ERA at predicting what is going to happen next, and xFIP is better than FIP. There are a few other more complex metrics that are marginally better than either of them, but if prediction is what we’re going for, why spend our time considering single-season metrics at all? There are projection systems that combine, intelligently, what pitchers have done in past seasons with what they’ve done so far this season. Steamer for instance, one of the best, has Chris Archer pegged to be the sixth-most-valuable starting pitcher in the American League over the remainder of the season.
All decry the Chris Archer All-Star snub!
No. Don’t do that. Statistical prediction is fascinating, and it’s an important part of today’s game, but description still matters. There’s a reason that teams come together on a baseball field and play a game, rather than looking at each other’s roster, deciding who’s better, and awarding a weighted percentage of a win to each side**, and the All-Star selections should reflect this.
** Although if baseball wants to huddle up and take some wins away from the Mariners right now I’d be fine with that.
ERA (or ER, if you prefer) is a fine descriptive stat, and it matters that Blake Snell was the best pitcher in the league at preventing runs*** over the first half of the season.
***Yes, I know that ERA is a combination of pitching, park, and defense, don’t @ me. Assessing the role of defense is hard, and Snell’s ERA ranking holds up when adjusted for park.
Hovering over every philosophical argument about baseball stats is the question of luck. How do you, the analyst, approach the existance of luck? How do you, the fan, process that existance? The two answers don’t have to be the same, but I think in all cases we should acknowledge that luck (or “variation” if you’re the squeamish type that needs a gateway word) exists.
I think it is possible to hold both sides in one’s head at the same time—to care about quality predictions and a deep understanding of probabilities, to attempt to decipher current true talent, and at the same time to place a high value on measureable outcomes—and I think that the All-Star game is an appropriate place to celebrate those measureable outcomes, especially in the superlative.
Blake Snell has the best ERA in the American League. His All-Star snub is ridiculous. I say so proudly.
Should Blake Snell, American League ERA leader, be on the All-Star team?
This poll is closed
Yes, ERA isn’t everything, but the league leader should be there.
No, plenty of other pitchers have done better by their peripherals.