On Liriano and Important versus Measurable

The NFL draft and it's post-spin machinations invariably amuses me. Not that I blame teams for trying, especially because the analysis of their moves can be so completely ridiculous. Somehow, a college left tackle is graded "four stars", and some other right tackle is rated "three stars", and god help the team that chooses the second guy first, even if it's the right choice. In the near term, it's a bad choice, because the grading system says it's a bad choice.



But there are several biases inherent in a grading system, and one of the biggest is obvious, when you think about it. It's that each criteria needs to be measurable. Otherwise, it wouldn't be a criteria. It would just be an opinion.



And the reason that can be a problem is because that which is measurable is not necessarily important, and that which is important is not necessarily measurable. For instance, the criteria for a quarterback that is measurable might include arm strength or height or speed, but are those necessarily what you prize most in a quarterback? Or is it the ability to feel pressure, avoid turnovers, or read the tendencies of a free safety? And how the hell do you measure those things?



The answer, too often, is that you can't. And so the temptation is to make the decision based on the more objective criteria, even though you suspect, or even know) they're not as important. Because they are measurable, they become important. And that which is important, but can't be measured, becomes less important.



Of course, the beauty of such grading systems is that they can, through hard work, gradually improve in the long term. For instance, if you suspect that that avoiding turnovers is a prime criteria, perhaps you go through every game film of prospective NFL QBs and measure the times they could have thrown an interception versus how many they did. And then you do that for the prospective NFL QBs for the last five draft classes, and compare your results to how they did in the pros.



But if you've gone through that exercise, you know just how many difficulties that really presents. First, there's the data collection, which can itself be a time-consuming and tedious proposition. But even if you get the data, what are you comparing it to? The number of turnovers they had in the NFL? Over one year? Over their career? As a ratio to their touchdown passes?



And how do you decide which players to include in the study? All QBs that were mentioned in pre-draft literature? What if they're cut before their first season begins? Then what are you comparing them to? And if you only choose QBs that ended up starting, is there some built in criteria that allowed them to start? Like the fact that the matched those same non-important but measurable criteria you started out with?



The answer to all those questions is that you make a decision based on what you really want to know. For instance, switching to baseball, say you're interested in how important K/9 rates really are for a young pitcher. A Twins fan might compare to a pitcher's K/9 rate in AAA to their ERA in the first three years in the majors, because she wants to know what they can expect from that player prior to their arbitration year. But a roto player might compare it to their WHIP the next year, because they want to know what to pay for that player in next year's draft. And a Yankee fan might want to know how long their career lasts, because its not like the Yankees will ever quit paying him if he's successful.



Whichever that person chooses, it's important to know that it means they are answering a very specific question. The general sense might be that "K/9 rate in AAA is important", but the specifics matter, especially when it's "for the long-term outlook of that player" as opposed to "for next year".



All of which is why I grew increasingly uncomfortable as multiple sources evaluated the velocity of Francisco Liriano's pitchers ad nauseum this spring. It isn't that it isn't news. It's certainly news when a player's velocity changes after they have come back from injury. It shows that something is different.



But, of course, something is different almost constantly with pitchers. Maybe they're switching to a slide step when someone is on base. Maybe they're throwing more offspeed pitches earlier in the count, or when they're behind in the count. Or versus left-handers. Maybe they're raising their elbow a little higher, or changing their arm slot, or sliding to the left half of the mound. Or maybe the break on their two-seam fastball is a little greater today than it was last start.



No, we aren't obsessing about it because it shows a difference. We're obsessing about it because we think it's important. And it might be, but I don't know a single study that says velocity has anything to do with how successful a pitcher is in the majors, either short term or long term. I'll admit, it seems like it should be. But as I think about all the pitchers who have had successful careers, only a fraction of them have the ability to throw true gas.



Let's unpack this just a bit more. Reports are that Liriano threw 93 miles per hour before the surgery and 89 miles per hour after the surgery. So how much less time would a 93 mile per hour pitch take to reach the plate over an 89 mile per hour pitch? I'll add my back-of-the-napkin figures below so my fellow geeks can call me on it if I blow the math....

I get two one-hundreths of a second difference. It's certainly possible that is significant. But it seems equally possible that there are factors that matter more, like control or movement or the ability to change speeds.

But we don't obsess about these as much because we can't, and that's because they're not as measurable. But they may be twice as important.