In The Lab: Player A and B Test

One of the more fun things to do in statistics is what I lovingly call the Player A and B test. It is more of psychological ploy than anything else. Unfortunately, we have feelings that get attached to every player on or off our team. So, comparing two players is next to impossible when their names are attached. That comes with positive and negative bias. For instance, comparing Jose Altuve to any historical second baseman is nearly unfair. We have all kinds of baggage attached that muddies the waters there.

I did not invent the Player A and B test. Standing on the shoulders of giants leaves me cold (bonus to anyone that gets that reference). We simply compare numbers and remove the names. One of the troubles is that some numbers become identifiers by themselves. Like if I say that Player A drove in 191 runs in a season then you automatically know I am referring to Hack Wilson.

In this edition of the test, we are comparing one Astros pitcher to a historical pitcher. We will prorate the numbers to assume a full season for this Astros pitcher. In doing so, hopefully we will muddy the identities just enough to make this a fun exercise. We will include some basic numbers and then another table with some sabermetric numbers. First, let’s start with some basic numbers.

Unfortunately, this is one of those cases where if you know your baseball history then you already know the two pitchers. Of course, I am going to reveal it at the end, so if you don’t know you can live in suspense. I should point out that Pitcher A surrendered more home runs than any pitcher in the league in that season and Player B did too. Otherwise, these pitchers look extremely similar and that is always the fun part of the Player A and B test.

I should also point out that both pitchers are in very similar stages of their respective careers at this point. Player A is 25 while Player B is 26, so both respective teams could collectively believe that these pitchers could grow and turn into something. However, this is usually where the comparisons break down. The era is different. The home ballparks are different. The teams are different. So, there is a ton that gets in the way of suggesting these two were separated from birth.

There are some numbers I like for quick reference at baseball-reference.com. I should point out that these aren’t the only numbers out there and some of the more analytically minded will point out they may not be the best ones. However, they are fairly easy to interpret, so they make a comparison like this easy. ERA+ measures a pitcher’s ERA against the league average with a ballpark adjustment included. 100 is average with everything under that being below average. Most of you are familiar with bases per out which has become a bit of a signature for me. The lower the better on that front.

Weighted adjusted average percentage calculates what a pitcher’s winning percentage would be with an average team. So, this includes average run support, average bullpen support, and average defense behind the pitcher. This immediately spills us into a neutral record. You simply multiply the adjusted percentage by the total number of decisions. Finally, we get the percentage of quality starts for the pitcher.

These are two very similar pitchers in most respects. The BPO is radically different but Player B somehow comes out looking better in the subsequent categories. This is because he had bad batted ball luck. Of course, I am being cagey in order to avoid spilling the beans on who we are talking about, but these numbers likely have left enough bread crumbs for you to figure out at least one of the pitchers.

Neither of these are good pitchers, but the adjusted won-loss records show that neither is as bad as they look initially. They pitched in different eras which could account for the quality start percentage. However, the innings totals are not all that different and even when we include the era and ballpark adjustment, we see that these two pitchers are very comparable.

Player A is Mike Maroth from 2003 and Player B is Mike Burrows from this season with his numbers prorated to the end of the season. The situations are obviously vastly different which is why this is such a big deal. Those 2003 Tigers lost 119 games. They were never in the hunt, so they were just happy that someone was there to give them some innings. As soon as the Tigers were ready to be competitive then they were ready to move on from Maroth.

Burrows is in a different situation. For one, he was supposed to be a number two starter, That obviously hasn’t worked out, but the adjusted winning percentage shows he hasn’t been as bad as the numbers would suggest. However, the team has designs on getting back in the race, so they have to weigh the benefits of getting five innings every time out with a pitcher giving up runs at a pace that has them on pace to lose most of his starts.

I’m not supposed to make declarative statements in the lab and I really can’t in this case. On the one hand, every important metric points towards positive regression. On the other hand, it is hard to be patient and wait that out when every loss hurts. This is one of those decisions where I can’t damn them either way they go. What do you think? Would you pull Burrows from the rotation when Hunter Brown returns?