A (54:37)
You know? Yeah, sure, they'd get some ice on it, get a little rub down, maybe put a jacket on or something. But yeah, that's, I don't, I mean, just kind of looking at Wilbur Wood, I, I, I doubt that he was doing an intense arm care, you know, shoulder prevent prevention strengthening regimen, not that anyone really was at that time. I guess if we go by fan graphs, war it was perhaps just for him not to get the Psy in 71 because let's say he was sixth overall in Fangraphs War that year, but Fergie Jenkins was number one in Fangraphs War with 11.2 that year. And you know, Wilbur was not a huge strikeout artist. Obviously those were not as strikeout heavy days. But yeah, he was, you know, four to five strikeouts per nine. But you know, knuckleballer, low babip, weak contact and all the rest of it and good control and just incredible durability. So. Wilbur Wood, what a guy. Wilbur Wood also wanted to shout out a couple interesting posts published on your website Fangraphts.com in the past week or so, which were paired posts by Ben Clemens and Davey Andrews. And Ben Clemons wrote a post entitled they Don't Make Barrels like they used to to. And Davey Andrews did a follow up entitled they Don't Make Pitch Models like they used to and they both hit on something interesting. Ben Other Ben noted that barrels are not as good as they used to be in terms of the type of production that they denote. Like if you have a barrel, you know, it's a stat cast stat and it's based on your exit speed and your launch angle. And it's basically like barrels are classified as batted ball events whose, you know, have produced a minimum.500 batting average and 1500 slugging percentage since Statcast was implemented. But this was several years ago when that definition was made. So I've always, you know, I've never really referred to barrels that much. I, yeah, I talk about barrel rate maybe sometimes, but it's, it's sort of squishy, you know, because it's just like, like these are batted balls that typically produce an outcome at least that good. Sort of imprecise, but you know, a useful analytical concept, but perhaps a little less useful these days because as Ben noted, the weighted on base average on barrels has declined pretty significantly over the past decade. The, you know, 10 years or so of the Statcast era that we have. And so early in, in that era, in 2015, barrels were worth something like a 1450 Woba. And this past year they were like a 1200 Woba, which is still really good. It's still really good to hit a barrel, obviously, but there's been a significant decline in how productive a barrel is, and there are many possible reasons for that, which Ben detailed in his article. Could be the ball that the ball is not carrying as well as it was in some of those seasons, of course. So that's part of it. But that's not the only thing. You know, part of it could be outfield defense continuing to improve. We've talked plenty about that. But it also could be something having to do, as Ben noted, with Goodhart's law. So this is. This is interesting. So Goodhart's law is when a measure becomes a target, it's ceases to become a good measure, or you can phrase it some other way, but it's basically like once you start aiming for some statistical benchmark, some metric that has proved predictive in the past, once everyone starts then tailoring their performance to that measure, then it becomes less useful as a measure. And that could be happening here. So maybe now that barrels are all the rage and people talk about barrels and, you know, players get paid based on barrels or some equivalent proprietary metric, that maybe they are aiming for barrels and they are hitting batted balls that satisfy the criteria for barrels but are not as productive. Right, because you could say just, you know, hit a ball really hard to center field or something and it'll just die out there. It. It's less likely to go over the fence, but you still get your barrel barrel because you hit that ball hard enough and at the specific launch angle that you get credited with a barrel. But it's not actually as good as barrels used to be before anyone knew what a barrel was. You know, before that was a stat when people were just trying to do things that led to productive batted balls. That was one thing. And now it's like, oh, let's just hit a barrel. And whether that leads to actual good outcomes or not, usually it does, but maybe a little less so. So that's an interesting idea because if you start just aiming to produce barrels rather than the outcome that a barrel historically generated, then maybe it's not quite as useful. And so, inspired by that, Davey looked to see if the same finding applied to stuff models. So, you know, pitching bot and stuff plus and all these many models that are available publicly or privately just to extrapolate from the characteristics of a pitcher's repertoire to say how good the stuff is. And it does look as if something similar has happened there, that there's been a compression in the range of possible stuff measurements basically like this standard deviation of stuff has declined. And Davey likened this to catcher framing. Once everyone appreciated how valuable that was, was they started training for it or guys who weren't good at framing weren't getting to catch anymore. And so the, the range of framing in the majors compressed and the best guys weren't as good relative to the average because the baseline increased. That seems to have happened with stuff as well. And so there might just be less variation in stuff because everyone is just trying to get great stuff all the time. And if that's what you're aiming for, just to have great stuff as opposed to to get strikes or to get outs or whatever, then you know, part of it is just like guys who don't have good stuff, they're going to lose their roster spot, they're going to lose their role, they're not going to throw as many innings and other guys are going to aim for stuff and they're going to come up and take those jobs. But also maybe sometimes you'll, you know, try to throw your, your hardest at the expense of something else, movement or command or whatever it is. And, and these models can measure command and other things too, movement also. But like if you're just trying to game the model basically, then maybe that will take a toll. And it seems like the correlation has decreased that the correlation between stuff and actual performance by pitchers is a little lower than it was a few years ago. And we only have these stats for, you know, four or five years or so. So it's not, it's not a huge difference, but it's perceptible. And Davey demonstrated that there's a real difference here. Or it could just be that there's so much compression in the stuff just because everyone kind of has good stuff these days, which you always hear, you know, it's like the mop up man today would have had like mind blowing stuff however long ago or whatever. You know, every reliever comes out of the pen, throw in 98 with a nasty slider. And so if everyone's doing that, then the correlation might decrease just because it's like there's just not as much separation among players basically. And so you can't really rely on the stuff plus purely to predict say a pitcher's WOBA allowed. So that's really interesting. And I just, I wonder, you know, you could constantly update your models for those metrics maybe or if you're a team, how do you handle that? Because you know, teams even if they're not showing someone their publicly available metrics, then players are going to be looking at their Fangrass page, but they're also going to be looking at whatever readout they get from their team telling them to do this or that. And I wonder how you guard against that. Just trying to juice those stats, basically, as opposed to trying to, you know, actually get good outcomes, which, you know, you're doing too. But there's, there's like misaligned incentives there, basically.
