Eric Gagne
Posted by Andy on May 26, 2009
I heard a bit on ESPN radio about how Eric Gagne was historically much better in save situations than in non-save situations.
That's easy to see from his splits.
In 216 games comprising save situations, he had an ERA of 2.31. In 138 games comprising non-save situations, he had an ERA of 3.06. While both numbers are good, that's a pretty significant difference. Note that the non-save situations excludes game starts which, although clearly a situation in which a save can not be earned, doesn't really apply to what we're looking at. Early in his career, Gagne was a starter, and it's helpful that the "non-save situation" already excludes game starts.
As a reliever, his career OPS against was .573, which is excellent. In save situations, that figure is .524 while in non-save situations it was .639. Again, both figures are good but that is a very large difference over so many innings (over 370 total.)
Eric Gagne's highest similarity score is to Bobby Thipgen, who had an even more stilted set of splits in favor of save situations. Thigpen's ERA was more than 1.5 runs higher in non-save situations.
I wonder why this is. I've heard managers reference that their closers do best when pitching one inning (the 9th usually.) But is this cause or effect? Clearly Gagne and Thigpen both did better when pitching in save situations, but was it their mentality or the hitters' mentality, or something else?
For a bit more data, I checked Mariano Rivera's splits, and he's much closer, although still a little better in save situations. Ditto for Trevor Hoffman.
I also notice that the K/BB ratios for all these guys are significantly higher in save situations. I wonder if that has to do with eagerness of the batters, knowing that they need runs and perhaps are more likely to swing, and therefore more likely to swing and miss.
Thoughts on this?
May 26th, 2009 at 10:35 am
It's partly selective sampling. Typically the best reliever is the closer. When these guys were at their best, they were closing. Gagne has not been a regular closer the last few seasons because he's obviously not the pitcher he once was. Therefore he now puts up his worse numbers in more non-save situtations. It might be more interesting to look at his splits only in those seasons when he was a top closer (although then of course you have a much smaller sample).
From '02 to '04, Gagne had a 1.70 ERA and 0.766 WHIP in save situations, 1.98 and 0.944 in non-saves. Still a gap (and it's probably enough innings to be statistically significant, though I'll let any statisticians correct me on that), but narrower.
May 26th, 2009 at 2:08 pm
Also, shouldn't we compare the types of hitters the pitcher was facing in each situation? Probable match-ups could be very different in save vs. non-save situations.
A closer is generally brought in to pitch the final inning, regardless of who is batting, so they are just as likely to face the 7-8-9 hitters as the 3-4-5 hitters. When they break from that role and they are being brought into a non-save situation (tie, trailing, extra innings, etc.), it's more likely that the closer is being brought in specifically to match up against tougher hitters.
May 26th, 2009 at 2:42 pm
That's an interesting thought Spycake. It's my sense that closer use in non-save situations is more often based on the other circumstances (extra innings, needs the work) than on the actual batters coming up, but you may be right.
All that said, there probably is something to the thought that a closer will not focus as much when up by 4 runs in the 9th as if up by only 1. Intuitively it makes some sense. I'm not aware of any rigorous studies to try proving it, however.
May 26th, 2009 at 5:23 pm
You are right -- the "quality of the hitters" argument works both ways, so it might cancel out.
I think every pitcher "pitches to the score" a little bit, or at least they should. No sense pushing your arm to make the perfect pitch when you've got a generous margin for error. The difference in approach may be even more significant for closers and other short relievers than starters, as relievers generally don't have to pace themselves for short appearances.
May 27th, 2009 at 12:31 am
Frank Francisco is facing Teixeira, Rodriguez, and Cano in a NON-SAVE situation right now!!!
/HUGE sample size
May 27th, 2009 at 10:00 am
I agree with JohnnyTwisto's observation about pitchers only being used as the closer when they're at the top of their game, and thus their non-save stats are inflated by work in less-effective years. I disagree with Spycake's match-up arguement, however. Generally, the manager will manipulate the bullpen to the relief pitcher's *advantage* - leftie vs. leftie, slider pitcher vs. batters who struggle with the slider, etc. On the other hand, the closer is brought in to face *any* hitter, whether or not they're good at hitting the closer's out pitch. While it is true the closer will equally likely face the 7-8-9 as the 3-4-5 hitters, as a non-closer, he is LESS likely to be brought in to face a batter who mashes his out pitch on a regular basis.
It would be curious to see collective split data from all primary closers during their prime (closer status) years, to see what the general trend is. Then we could look at indivisual pitchers to see whether they have a larger or samller than normal differrential.
May 28th, 2009 at 10:07 am
"as a non-closer, he is LESS likely to be brought in to face a batter who mashes his out pitch on a regular basis."
Is he really, though? Closers (the good ones, anyway) are generally thought of as the best relievers on their staffs -- I doubt Eric Gagne in his prime had ANY batter "mash his out pitch on a regular basis." Good closers are always seen as "above the fray" and not really subject to manager's bullpen manipulation. A clear example of this is that closers finish most of their games -- they don't get pulled from a game unless it goes to extra innings and they reach their workload limit, and even then they almost always finish their own innings. If they were subject to a manager's manipulation, I suspect we would see more closers pulled before the conclusion of a game or inning.
It seems more likely that a good closer would be brought in to non-save situations either (1) to match up against the other team's best hitters, or (2) simply to "get his work in" if he hasn't appeared in a game recently. Maybe those two types of non-save situations cancel each other out in regards to hitter talent, but it would be interesting to see some data on the cumulative opposing hitter talent in save vs. non-save situations.
May 28th, 2009 at 10:10 am
This is a tough argument to decide without looking at the actual data in detail. Keep in mind that closers (when in save situations) never face really weak hitters, such as pitchers...those guys would always be pinch-hit for in the 9th inning. But closers also get who they get...sometimes they don't need to face the 3-4-5 hitters. Guys brought in during the 7th or 8th often (but certainly not always) are coming in to face the best hitter on the opposing team--but sometimes they also get to face very weak hitters. Who knows how it all averages out?