September 15, 2015

Baseball Therapy

Seven Months Has Gone So Fast

by Russell A. Carleton

the archives are now free. All Baseball Prospectus Premium and Fantasy articles more than a year old are now free as a thank you to the entire Internet for making our work possible. Not a subscriber? Get exclusive content like this delivered hot to your inbox every weekday. Click here for more information on Baseball Prospectus subscriptions or use the buttons to the right to subscribe and get instant access to the best baseball content on the web.

Subscribe for $4.95 per month Recurring subscription - cancel anytime. Subscribe for $39.95 for one year a 33% savings over the monthly price! Purchase a $39.95 gift subscription a 33% savings over the monthly price! Already a subscriber? Click here and use the blue login bar to log in.

It's September. People are starting to use words like "crisp." Christmas trees are already out. Kids are back in school, and the world smells like nutmeg. Everyone's getting ready to tweet that Green Day joke at the end of the month. And major-league rosters have expanded to roughly 68 players each. On top of that, a lot of minor-league seasons have ended, meaning that teams are free to call up the veteran guys they had signed as cover during the offseason, their top prospects, their not-so-top-prospects who just happen to be left-handed or fast, and Larry, the guy who is about to pitch the only three innings of his major-league career in garbage time … but he wore a major-league uniform!!!

Somewhere in baseball, there will be a call-up guy who has a video-game month (.370/.420/.500 in 38 PA) completely out of nowhere. We all know that's a small-sample fluke, but what about the top prospect who is known to be a good hitter? He comes up, plays most days, and hits.290/.340/.470. It's the sort of line that is reasonable to expect a good hitter to put up. Even throwing in the usual #SmallSampleSize warning though, can we trust his stats? September is such a weird month, after all. There are all these guys floating around who belong in Triple-A and then there are a bunch of teams that are out of any meaningful playoff chase, so they may not even be trying. Is our top prospect really putting up that line against real major-league competition or is he just feasting off guys who either aren't good enough or just going through the motions?

Warning! Gory Mathematical Details Ahead!
Let's start with the idea that players who are out of the pennant race are just going through the motions. For the guys on the fringe, that doesn't make sense. This might be their one chance to get an audition for next year, and while they may not have much (they are fringy for a reason), they're going to give it their all. For guys in the middle of a pennant race, everyone will be pulling hard. Less so for the guys who are already established in the big leagues. Do some of them take days off here and there—even though their name is on the lineup card—when there's nothing to play for?

I used data from 2010-14, and only regular season games played from September 1st on. I looked at all players who had 250 or more PA (or batters faced in the case of starting pitchers, 100 for relievers) within that season. This will capture our regulars, as well as guys who have at least been hanging around for part of the season, either on the bench or with some amount of time as a regular guy. For all players, I looked at their overall season stats for several outcomes (such as strikeouts, walks, homers), and using the log-odds ratio method, created a control variable for what the probability was that a given PA would end in that outcome.

For all teams, I looked at their position in the standings on that day. I considered a team to be playing a "meaningful" game if they had neither mathematically clinched a playoff birth nor been mathematically eliminated from one, and they were within three games of a playoff spot. That is, they were either within three games of a spot or fewer than three games in the lead. Since a game has two teams (and the schedule-maker is not psychic), games can involve zero, one, or two teams that are playing for their lives. I had codes for whether the pitching team was playing in a "meaningful" game, one for the batting team, and one for when both were. For the last one, maybe there's something to neither team having much to play for and everyone just trying to get it over with. The control variable and these three "meaningful" game indicators went into a binary logistic regression.

It turned out that the control variable remained a very good predictor of whether the PA would end in a strikeout, or whatever outcome was being tested. It's designed to be a very good predictor, but if it had ceased to be one, that would tell us something interesting about September. Seems that the regular corps do pretty much the same in September as they do in June and July.

There were zero cases in which any of the "meaningful games" variables made a difference. There weren't even any borderline significance calls. Whether teams were in or out of the playoff chase, their regular players did pretty much what would otherwise be expected of them. Even the guys who are out of the race still seem to play as hard as they usually do. (In the aggregate, of course; there probably are some individual cases where guys dog it.) I don't think that this should shock anyone. There's professional pride to think of. There's individual stat-padding. There are guys who may have been up for a good portion of the year but still need to think of their job security. There are guys who want to set a good example for the young kids. If nothing else, there's not wanting to have the reputation as the guy who quits when the going gets tough.

Well, what about the guys who aren't regulars (again, defined as fewer than 250 PA in the season or 250 batters faced for a starter or 100 for a reliever)? Maybe when Clayton Kershaw, with nothing to pitch for, sees a guy fresh up from Triple-A walking to the plate, he decides to take it easy and isn't his usual self.

It turns out that looking at fringy guys who got between 100 and 250 PA or batters faced—lower than 100 starts to produce a completely unreliable sample size—in the season (at any point) facing a player who was not on the fringe that year, none of it matters. When a fringe batter faces an established pitcher, the results were pretty much what we expected from their overall season numbers, and it didn't matter whether the game meant anything to either one. Same thing when a fringe pitcher faced an established hitter, or when two fringy players faced each other. And it didn't make a bit of difference whether the game was "meaningful" or not.

So we can throw away the idea that in September, stats don't mean as much because players aren't trying as hard. But all we found above was that players perform to what we would otherwise expect of them. With all these Quad-A guys floating around, it might be a little easier for a player to put up some good numbers. How much should we worry about that?

For that, we need to look at how often those Quad-A guys are really part of the equation. Below is a table of the percentage of plate appearances, by month, that are staffed by a player who has either fewer than 250 PA in that year or a pitcher (and I used starters only for this one) who had fewer than 250 batters faced in the season.

Month	Percentage of Fringe Batters	Percentage of Fringe Pitchers
March/April	17.5%	9.2%
May	16.8%	6.1%
June	14.9%	6.0%
July	14.7%	4.7%
August	18.0%	5.6%
September/October	23.6%	12.0%

Fringe guys have their heyday in September (and October), particularly pitchers. Teams start giving their regular starters an extra day of rest or they might skip a start, but someone has to toe the rubber that day. Chances are that it will be a fringe guy, but let's keep in perspective how often that happens compared to the rest of the year. There are always Quad-A guys floating around the league and no one says that we can't trust stats from August.

But wait, there's more! It turns out that while more Quad-A players are playing in September, they might not be as bad as you think. These fringe guys aren't as good as the regular guys (they're on the fringe for a reason), but the September call-ups seem to be a cut above the normal roster-filler. Here's a month-by-month breakdown of how batters and starting pitchers who are around for fewer than 250 PA did during their appearances, when facing off against a non-fringe guy on the other side of the encounter (more than 250 PA in that season) from 2010 to 2014. I used OBP as a marker, but other stats tell the same story.

Month	OBP of Fringe Batters	OBP Allowed by Fringe Pitchers
March/April	.276	.366
May	.278	.377
June	.270	.367
July	.280	.380
August	.287	.357
September/October	.286	.345

In September, there's a quirk in the rules that says that a minor-league call-up does not accrue MLB service time while he is on the roster. Teams are at liberty to bring up top prospects without starting their service clocks, and in some cases, why not? Before September, teams have to contend with those pesky service-time issues, so it's more likely that, needing some cover, they'll bring up a washed-up veteran or a lesser prospect to fill a void. In September, we see a classier group coming up. Then again, looking closely at those numbers, the effect seems to start in July and August. Maybe it's just that fringy guys at the end of the year have several extra months—perhaps an entire professional season—of development under their belts before they get the call to The Show.

There's Only One September!
Can we trust September stats? The numbers suggest that for the most part the answer is yes. There are more fringe guys who are up getting reps with the expanded rosters, but there are always fringe guys hanging around, and the ones you find in September are better than the ones you find in May. The idea that guys who are out of the pennant race are just going through the motions finds no evidence here. If you're going to hesitate for any reason, it should be that even a guy who plays 28 games in September will only notch 100 or so plate appearances, and that's a small sample and should be interpreted accordingly; the evidence here suggests that other than that, you may overreact to your heart's content about how amazing that guy is. Or will be. Next year. We'll get 'em next year.

Russell A. Carleton is an author of Baseball Prospectus. Follow @pizzacutter4
Click here to see Russell's other articles. You can contact Russell by clicking here