Should Tracy Have Pinch-Hit for Mora?
Last night, the Rockies, pushing for a playoff spot and just out of the lead in the National League West, suffered a tough loss to the Arizona Diamondbacks, a team that’s been “playing for pride” for months. The loss dropped Colorado to 2.5 games back in the divisional race with only 12 games left to play. In the top of the ninth inning, down 3-1, the Rockies’ batters due up were the red-hot Troy Tulowitzki, veteran Melvin Mora, and Rockie (Rocky?) for Life Todd Helton. Wait, what? Melvin Mora? Surely the right-handed hitting Mora wouldn’t bat against the right-handed pitcher Juan Gutierrez? Maybe the Rockies’ bench is a bit thin due to injuries, but while Helton has had a down season, he has the platoon advantage… Mora is only playing because of Ian Stewart’s injury. Colorado manager Jim Tracy wouldn’t let Mora bat, not with this being the Rockies’ last chance in a key game, right? After Tulowitzki struck out, Mora did indeed go up to bat and struck out swinging; then Helton fouled out to end the game. Did Tracy make a mistake in not pinch-hitting for Mora in such a crucial game?
Mora has actually been fairly useful for the Rockies this season, offsetting his poor fielding with slightly above-average hitting (104 wRC+), which is pretty decent for a somewhat cheap bench player. Of course, for this situation we need more than current performance; we want “true talent.” For that, I’ll refer to CHONE’s August 29 updated projections, which I’ve translated to wOBA (using last season’s weights and leaving out SB/CS as they aren’t relevant for potential PH/platoon situations). Mora is currently a .330 wOBA hitter according to CHONE. Mora’s career (well, 2002-present) platoon split is not that big — .354 wOBA vs. LHP, .351 wOBA vs. RHP. When estimating his platoon skill with reference to his .330 wOBA CHONE projection, his expected wOBA vs. RHP is .326, and vs. LHP .339. Tracy would want to pick a batter expected to have a better than .326 wOBA in this situation, obviously.
On the Rockies bench at the time were two obvious candidates (the others all being either right-handed and/or terrible, and Ryan Spilboroughs having already been in the game): Jason Giambi and Seth Smith. I say “maybe three” because one could make a case for Chris Iannetta, as he’s a pretty good hitter, but a) he’s right handed, b) I’m not sure he was available given recent injury problems, and c) the pinch-hitting penalty eliminates his chances of likely being helpful in this situation anyway.
The key is c): the pinch-hitting penalty. Research has shown that hitters typically have a much more difficult time hitting off the bench than normal, so we expect their wOBA to be about 10% lower in these situations.
So in Giambi’s case, while CHONE projects a .357 wOBA, off the bench, we’d expect him to be a .321 hitter. After estimating his platoon skill, we’d expect him to be about a .328 hitter versus RHP. CHONE projects Smith projects at .368 wOBA, which is .331 after the PH penalty, and after adjusting for the platoon advantage, comes out to .333.
While the “penalty” means that pinch-hitting for the platoon advantage is rarely as beneficial as fans think it is, in this case, it does appear that the Rockies would have had a better chance by having Smith or Giambi hit for Mora. But let’s not got crazy — over 700 PA, a .328 wOBA is worth less than two runs over a .326, or barely a thousandth (~0.0017) of a run per PA. Given the margin of error in projections, not pinch-hitting Giambi for Mora is hardly a ridiculus decision. There is a bigger gap between Smith and Mora: over 700 PA, a .333 wOBA is worth about 4 runs than a .326, but that’s still less than a hundreth of a run (~0.0058).
It is fair to say that given the game and season situation that Tracy should have pinch hit Seth Smith for Melvin Mora. Once we take a closer look at the numbers, however, the advantage gained wouldn’t have been as big as one might think.
Matt Klaassen reads and writes obituaries in the Greater Toronto Area. If you can't get enough of him, follow him on Twitter.
I wish Tracy had pinch hit for him with Smith. I’ve never understood how baseball-oriented statisticians can be so careful about small sample sizes, but then glean theories about events that can only be determined by small samples. Pinch-hitting is so context dependent that it’s hard to see the value in the numbers beyond the most raw of data. I’m sure I’m missing something, and this is not to be dismissive.
I think the article should also consider that there is some value in saving the pinch hitter for later in the inning, especially considering that they needed at least 2 runs and had their catcher and pitcher due up if they got two batters on.
Jay Payton (LF) and Miguel Olivo (C) were due up. Considering there is only one out at the moment of decision, that’s at least two average batters hitting behind Smith if he can get on base.
If the Rockies were to win the game, it was certain that they would have had to pinch hit for their pitcher. They needed to score 3 runs, there was only one out and the pitcher was the fifth guy up. Anyways, you at least have to consider whether you would rather pinch hit for Mora or Olivo since you know one of the pinch hitters will be used for the pitcher.
Also, considering the situation, I don’t think using wOBA is the best way to measure a hitter. When down by 2 with 1 out, OBP is much more important compared to power than it usually, perhaps you should have used different weights.
Down 2 in the 9th with the bases empty power doesn’t matter much. I’d probably go with Giambi for his willingness to take a walk but as you said it doesn’t make a very big difference.
I know I shouldn’t nit-ipick like this, but for some reason I can rarely resist. I think the Reds would be very surprised to find out that the Rockies were just out of the lead in the National League Central as they were planning on clinching the central this week.
No, that’s perfectly fair. Thanks.
[Bangs head on desk.]
Pinch hit Giambi for him and probably have to use another bench guy to pinch run for him. He can’t run very fast.
But why? They’re down two. Giambi’s run doesn’t matter. He’s slow but if he can’t go station to station he shouldn’t be on the roster.
For such a small difference, I’d assume you would want to use more exact numbers (without rounding) that have been proven to be accurate. Yet you toss out an “about 10% penalty” and use it.. With numbers coming out so close- a penalty of 10.5/9.5 let alone a bigger +/- in penalty could change your final answer.
How accurate is that 10%? Is it 10.0%? Have different types of players shown different levels of drop off? With it being so close, yet also inexact -your whole article probably could have just said: “Should Mora have been ph for? Maybe yes, maybe no, but due to pinch hitters performing worse than starters it wouldn’t have made much of a difference either way.”
Seth Smith probably doesn’t have a pinch hit penalty, and might be the best pinch hitter over the last two years.
We often hear the “pinch hit penalty” being bandied about, which is nice, but…
I think that it came from our book, The Book. I don’t know of any other research on the subject. We explain in The Book that some or all of that penalty (as well as the DH penalty) could be because pinch hitters are sometimes/often regulars who are injured or tired. We also explain that the pinch hitter is always facing the pitcher for the first time even if the pitcher has been in the game for a while.
My personal opinion is that at least half of the penalty, and maybe more, is due to the “injury/fatigue” factor. So if you have a healthy player on the bench he may have little or no penalty. And of course in this situation, any player would be facing Gutierrez for the first time in the game.
Matt also forgot about the true platoon ratio of the pitcher. For example, if he had a small or even reverse platoon ratio, then using a lefty would be of no advantage, at least as far as the L/R matchup was concerned. As it turns out, Gutierrez’ estimated true platoon ratio is 1.06, which is a little less than the average platoon ratio for a RHP.
So we have, in my opinion, less than a 10% penalty for the pinch hitter. We have a pitcher with a smaller-than-average platoon ratio, which would favor leaving Mora in, or using a better hitter who is RH (and one who has a small platoon ratio himself, although all RHB tend to have the same “true” platoon ratios).
And we have the legitimate issue of making sure you have good pinch hitters left.
Plus, as someone mentioned, OBP and not wOBA (or OPS or lwts, etc.) is most important in that situation. I don’t know what Mora’s projected OBP is compared to the other options (and then you have to use OBP platoon ratios for the various actors and not wOB or “runs” platoon ratios).
So, you have a fairly complicated situation on paper.
However, if things appear close (one option is about as good as another), as they seem to appear in this situation, given the complexity of the issues, and given the uncertainties of our estimates, then that becomes our answer, and a fine answer it is:
Flip a coin or let the manager decide, and don’t worry about it…
Thanks, MGL. You are right about about my simplification of the issues. Believe it or not, when working on this post I considered sidebars on both the OBP/wOBA and Gutierrez split issues, but didn’t want to make the post any longer and more cluttered than it already was. Really, I did. But I’m also a bit lazy when it comes to this stuff.
Probably “no big deal” would have been a better conclusion, then, but given the information I was working from (including the not-quite-as-solid-asi_-thought 10% pinch hit penalty), I didn’t think I was too emphatic. Which, of course, makes things less interesting…
But let’s not got crazy