Geoff from Ducksnorts wrote a piece breaking down Trevor's appearances in save situations vs non-save situations over the course of his career. I thought I'd piggyback on and run a little more analysis to see if there really are statistically significant differences with how Trevor handles the situations. Read Geoff's piece first and then come back here.
Back? Good. OK. Somebody check my numbers...
Working off of Trevor's career splits, I thought I'd first look at individual plate appearances and how often Trevor let guys get on base. Here are some numbers:
Given those numbers and adding the number of batters Trevor beaned in saves and non-saves (6 and 2 respectively), we see that Trevor allowed 568 runners in 2424 plate appearances in save situations and 364 runners in 1387 non-save situation plate appearances. Now this kinda takes into account the fact that some of those plate appearances in the non-save situations were in the same game as a blown save so you might argue that obviously Trevor would be more rattled after blowing a save, but we'll take that as a given and hope that Trevor's not the type to be so soft.
In any case, we see that Trevor allows some sort of baserunner in 23% of plate appearances during a save situation and 26% of the time during a non-save situation. With a t-test, we're 95% confident that Trevor allows more baserunners in non-save situations. One should note that this doesn't mean that the the .043 difference in on base percentage is a definitive number. It only means that the difference between the two numbers is statistically significant.
The only thing this doesn't really take into account is the number of intentional walks that Trevor gives up, which probably shouldn't count against him. There were 15 total IBB in save situations compared to 36 in non-save situations, which actually changes the numbers significantly. Taking intentional walks into account, we are NOT confident that Trevor allows more runners in non-save situations.
So we're 95% sure that Trevor allows more baserunners in non-save situations, but that's partly due to the fact that he's called on to intentionally walk so many more batters. What about the number of actual runs?
Running our t-test again... We are not very confident that Trevor allows more earned runs in non-save situations. However, closers really shouldn't be letting anybody in at all and we are 85% confident that Trevor allows more runs in non-save situations.
What about the big ones? The home run? Trevor seems to blow these things in somewhat spectacular fashion, right? He also seems to save these games in spectacular fashion. Strike three swinging on a ball 3 feet in front of the plate for the win! Let's look at these numbers:
Quick check of our numbers shows that Trevor does not give up more home runs in non-save situations. However, Trevor strikes out 27.52% of the batters he faces in save situations and 24.87% of the batters he faces in non-save situations and from those numbers we are 95% confident that Trevor strikes out batters at a greater rate in non-save situations. Only trouble with that is the f_ckin' pesky difference between plate appearances and at-bats. When we use at-bats instead of plate appearances, our confidence drops to 85% that Trevor strikes out batters at a greater rate in non-save situations.
WHEW. So what did we learn? I hope I'm getting my language right here in case my statistics professor is reading. (Hi Prof Lackritz)
- We're 95% confident that Trevor allows runners to reach base at a greater rate during non-save situations.
- If we don't count intentional walks against Trevor, the difference in runners allowed is not statistically significant.
- We're 85% confident that Trevor allows runs at a greater rate in non-save situations compared to save situations
- We're not very confident that he allows more earned runs in non-save situations.
- The rate at which Trevor gives up home runs in non-save situations vs save situations is not statistically significant
- We're pretty confident (85% or 95% depending on how you run the numbers) that the rate at which Trevor strikes out batters is greater in save situations than in non-save situations
So there you have it. It was a little more effort that I normally do, but hopefully you've learned something from it. Taking into account a few significant factors, it looks like Trevor might be better in save situations, but it's somewhat difficult to prove that he's worse given the context of the situation.
Somebody check my numbers.