23 January 2016

How A Day Off Affects Catchers

The 2016 Orioles are cursed (blessed?) with two catchers worthy of starting roles on Major League teams. While having a deep roster at one position is always a blessing, the Orioles suffer from the employment of both Matt Wieters and Caleb Joseph thanks to their salaries complicating resource allocation and because only one can be on the field at a time.

There was a longshot opportunity for Wieters to play first while Caleb Joseph served as the primary backstop, but that was eliminated when Chris Davis signed a long term deal to stay in Baltimore. Wieters and Joseph may each get some DH duties, but neither is ideal for the role - I'm sure most people would like to see better pure hitters playing the only position for pure hitters. Most often, I would expect to see one playing while one gets a day of rest, typically assumed to be a good thing, particularly for catchers.

For health, regular days off may be ideal for players in arguably the sport's most demanding position. At the plate, rest may be ineffective at best and detrimental at worst.

To test this (actually, with the thought that rest was good for plate performance),  I gathered all events from 2003 to midyear 2013 that featured a catcher holding a bat. I got rid of any player that didn't meet a very generous 100 at bat threshold so that emergency catchers and September call-ups wouldn't affect the final results. Then, according to the date of the event, I determined which happened in a game following at least one day of rest and which happened in a game following a game in the immediately preceding day. This reflects the fact that catchers can get rest from normal days out of the lineup, days in which the team is not playing, the entire offseason, and - unfortunately - injury.

Rehabbing injuries, while technically time away from baseball, are a special kind of rest because they aren't really rest at all. Rehab is a ton of work, often painful, and many players are affected by injury in the first few games that they come back and recapture their routine. Injuries have the ability to adversely affect offensive productivity when the player returns to action, so it's worth noting that the final results of my analysis could, but are not necessarily, affected by the circumstances surrounding them.

The following chart shows how catchers have batted in a game following a day of work plotted against how catchers have batted in a game following a day of rest:

There's a pretty solid correlation between batting average and batting average after a day of rest; good hitters are always good hitters, and rest doesn't make a bad hitter a good one. But we're interested in whether a day of rest adds a boost to a hitter regardless of given quality.

One thing to keep in mind regarding the Orioles' situation is that having a significant timeshare is very unusual. The vast majority of games played after a day of rest were played after a day of rest:
That's generally consistent with the thought that most teams don't have two catchers worth significant time. Really, it's hard enough to find one player who is competent at both the plate and behind it, or is competent enough in one area to make up for deficiencies in the other. It seems most catchers have one day off, whether a scheduled day off or a simple day of rest, and then get back to action. If Wieters and Joseph split the season, say 100/60, they'll both be seeing more than one day of rest at a time fairly often.

And therein lies another potential pitfall with the final results of this analysis: good catchers rarely get rest, so the offensive productivity of catchers after time off is driven equally or possibly mostly by second-tier catchers who either back someone up or aren't good enough to take the job full time.

In five statistical tests with a random sampling of catchers (21 catchers in each sample) with at least 100 at bats, the difference between their batting averages after rest and their batting averages overall is generally too small to be statistically significant. In the roughly 15% of samples that the difference is found to be statistically significant, it's because the catchers are batting so poorly after a day of rest.

When testing is applied to the population of catchers, the batting average of catchers on days of rest is so far below the rest-independent average of the same population, it is statistically significant at the alpha=0.95 level.

These tests indicate that catchers tend to do worse after days of rest, but that their performance in many cases may be a result of random variance - that they could just as easily be getting hits as they are getting out. Other catchers are really hampered by a day off, possibly breaking a rhythm, coincidentally (or not, if managers elect to play backups against a weak starter in favor of rested regulars against aces the following day) seeing great pitchers after a day of rest, or catchers regularly getting rest simply being worse players in general.

The same effect is seen in on base percentage, and is possibly even more pronounced. Players more frequently post high rates of getting on base when playing without rest. Perhaps better players play without rest more often, but also consider that a player might need a few days to get a good sense of the strike zone, or get into a rhythm at the plate:

As with batting average, I conducted statistical tests on the on-base percentages of five separate 21-observation samples of catchers. The difference between catchers' ability to get on base after a day of rest was found to be statistically significant at the alpha=0.95 level in roughly 30% of samples. Again, the difference in OBP between players' averages and their performance after days of rest was found to be significant negatively; catchers do worse (if they do worse) after a day of rest.

Because the results were not particularly conclusive, I conducted the same tests on the population of catchers as I did with batting average. And again, the likelihood that the difference in rest OBP and overall OBP in the population was minuscule; as a whole, catchers are worse after rest.

Whether these results are directly applicable to the anticipated Wieters/Joseph timeshare is up for debate. Arguably they are both better catchers than the majority of catchers seeing days off often throughout the season. However, catchers generally tend not to do well after a day of rest - surprising, considering the possibility of healing, relaxing, and studying on a day off.

For what it's worth, I accept the proposition that catchers are marginally worse after a day of rest, perhaps not as noticeably as the full population suggests. That leaves me either hoping that Joseph and Wieters are possibly unusual cases that see no detrimental effects from rest, or, as has been my position since Wieters accepted the Orioles' qualifying offer, that the Orioles would use their depth at the position to allow for trading one of the two in exchange for filling other areas of need. In addition to filling roster holes, moving forward with one true primary catcher may generate better performance from the catcher position over the course of the 2016 season.


P said...

So why didn't we keep Clevenger and trade Wieters? Once he accepted the QO, he had a definite cost, though an undetermined worth. He probably wouldn't be worth 2.5 Wins, but the O's could have sucked up a portion of the cost to average things out. And most teams would have been interested in Wieters, even if a bit expensive.
As it is, it sure looks like Wieters is going to be wasted money. I wish we'd kept Clev...

Jon Shepherd said...

Because it probably was pretty difficult to trade Wieters. He would require a sign and trade with full no trade protection. Team can deal him mid-June without his say.

Anonymous said...

I'm sorry P, dealing Clevenger for Trumbo was a genius stroke by DD especially before Davis signed. He even got a potential LOOGY in Riefenhauser if he can find a taker for Matusz. I liked Clevenger, too, but like Jon says, Wieters is untradeable until June. Now if Wieters can rebound and excel early, he might bring an excellent return in June/July. Maybe he can be traded for useful controllable players.

Right now, we might be wasting a little bit of value, but this is a good problem to have right now.