Halfway through spring training, I try to make spring numbers important.
You may think spring training performance doesn’t matter at all. You may think it matters a little or you may think it’s suggestive of the season ahead. Whatever your take, you probably will look at the spring training stats of a player who is doing well or poorly and use those stats to support your preconceived notions of that player. Or you’ll ignore their stats. People watching Tyler O’Neill this spring or watching Matt Carpenter are in some way paying attention to their spring and letting it influence their opinion of them.
They may not be wrong.
About seven years ago, FiveThirtyEight tried to measure when spring training mattered. And they found that spring productivity is statistically significant when predicting actual performance in the upcoming season, using expected performance from, in this particular case, the now defunct Marcel projection. I assume a hybrid of ZiPS and Steamer will do just fine in replacing Marcel for this purposes of this post. The relevant part of the conclusion lies here:
In other words, spring numbers can and should affect our predictions for a player’s regular-season production, but only slightly, and only after a particularly strong or weak performance.
Unfortunately, the author doesn’t reveal his method on how exactly he is raising or decreasing a player’s expected performance, but just know that there needs to be a difference of at least 17 points of expected wOBA in order to make a difference. So I’ll just show their expected wOBA through Depth Charts and what they’ve done so far.
This is everybody with at least 19 PAs. The next two players in PAs are Nolan Gorman and Evan Mendoza, two players who wouldn’t be much use in something like this. And then Harrison Bader, who missed a few games to injury, and Edmundo Sosa, who missed a few games to being a father. Jose Rondon I don’t believe missed any games and is killing it but still falls short of my plate appearance threshold. Nobody else has more than 10 PAs. At least, among the plate appearances recorded. Another reason I ignored Sosa. I know he hit a HR in a B game and well, there is no evidence of this online.
Interesting results thus far. If you’re a player who we have remotely any hope for becoming an All-Star, you’ve sucked this spring training. But if you’re a fringe starter who we hope will break out, pretty good news for those players. The catching situation looks good this spring, we seem to have a legitimate starting left fielder, and looks like our starting 2B will be fine (assuming it’s Edman, that is).
Except wait! We’re only halfway through spring training. There seems to be no plate appearance requirement – I assume players getting 15 PAs simply aren’t getting a boost just because they had a few singles find a hole. So to try to correct for that, I’ll simply assume they play exactly to their expected wOBA for the rest of spring training – up to 60 PAs. Which means I’m weighing expected wOBA more than current performance yes, but another way to look at it is that they’d really, really deserve a better or worse expectation if they’re still significantly different.
Like I said, I don’t know how the author is figuring out the revised wOBA at all. I could find no consistency myself in the examples listed. But let’s just say as a rule of thumb you get a 1 point increase or 1 point decrease for every 17 points of wOBA differential from your projection. I don’t know if this is high or low. Kolten Wong, during the spring of that post, had a spring wOBA 201 points higher, which would translate to 11.8 higher points of wOBA… and the author landed at 13. On the other hand, Castellanos saw the same increase with just 147 point difference. So I don’t know. So it seems like for extreme performance, I might be undershooting the difference, at least by whatever method this author uses?
So keep in mind that I might be undershooting the wOBA when I say that Tyler O’Neill’s 103 point difference would translate into an extra 6 points of wOBA. An extra 6 points of wOBA would give him a .314 wOBA. John Nogowski has a projected .314 wOBA and a 95 wRC+, so you can put O’Neill’s new and improved wRC+ there. And that’s if O’Neill cools down and if I’m not in fact underestimating here.
Others are more marginal with the adjustments. John Nogowski gets a 3 point bump in wOBA, which would push him to a 97 wRC+. Molina is starting from a 79 wRC+ so his boost of 2 points is appreciated. And then we enter a long period where we could effectively assume no difference at all. Carpenter isn’t as quite affected as you’d think since he’s still walking, but he still loses a couple points of wOBA, which is rather important at this stage of his career. And, uh, let’s hope Paul DeJong turns it around!
To be clear, I would advise mostly using this information as a fun exercise. Half of spring training is still left to be played and you can expect these numbers to change. But it is something that Tyler O’Neill has played so well so far in spring that he could play to his relatively low offensive projection for the season and still see a big boost. And really, the same applies to DeJong in the negative sense who would definitely have below his 100 wRC+ projection, which again, I’d prefer to not entertain that possibility.