Wednesday, April 18, 2007

Raging Against the Serene

Q: What kind of dumb-ass baseball writer stays up until midnight on a weeknight, writing notes about a West Coast game, and still manages to miss the most talked about story?

A: A Twins Geekish one, apparently. Because yesterday, after a 11-2 thrashing of the Mariners, the most talked about story wasn’t my favorite, which would have been how the Twins signed Ramon Ortiz at half the salary level of Jeff Weaver. (No, I didn’t think that would be anyone else’s story. I’m a geek. We’ve established that.) It wasn’t Luis Rodriguez showing Twins Territory some glove. It wasn’t even Torii Hunter, despite a grand slam and a shoulder/wrist injury.

No, the biggest talker yesterday (as judged by random chats I received) was Joe Mauer’s bunt in the fifth inning. If you missed it, there were runners on first and second base and no outs in a game that the Twins led 3-1. Weaver had started the inning giving up three straight singles to Piranhas, and given up a run. Mauer laid a bunt down the third base line which died in the Safeco rough, Seattle catcher Kenji Johjima raced up the line, and his throw just beat Mauer racing to first base.

Later that inning Hunter hit a grand slam that essentially decided the game, so nobody was arguing with the result as much as the decision. The debate raged in part because Mauer did something similar last week, bunting when it seemed like the Twins really needed him to swing away.

Judging from the feedback I received, the anti-bunt side has introduced some rather strange bedfellows. On the one hand, there are those who are pretty damn convinced that bunting is usually stupid, and that the Twins do too much of it. They’ve found themselves aligned with another crowd who just wants Mauer to start hitting for some power, or at least drive in some more runs, dammit.

Last week we talked about how a person’s philosophy frames their reality and this bunt seems to be another example. Those fans whose pet peeve is watching the Twins bunt themselves silly are livid because they think this is another example. And those that feel that Mauer’s popularity needs to be justified by some offensive fireworks are equally perturbed. But objectively, this was a completely justifiable decision, whether or not Hunter hit that home run.

Mauer’s bunt wasn’t necessarily meant to be a sacrifice. There is a reason that that Johjima had to play that ball, and it was because the third baseman was playing back. If the ball would have rolled a foot farther, Mauer would have been safe, meaning the bases would have been loaded with no outs. The actual outcome also really wasn’t so bad – runners on second and third base with one out and a very good chance to add to the lead with Michael Cuddyer and Justin Morneau up next.

I’m especially surprised that anyone who claims they’re baseball philosophy is steeped in sabermetrics should argue about this. I’m pretty sure we can do a quick study that shows that sabermetric theory practically DEMANDS a bunt in that position. Let’s find out.

One tool that is really handy for talking about in game strategies is something called Thorn and Palmer’s Expected Runs Matrix. It is just a cute little grid that shows, given a certain number of outs and people on base, the average number of runs that should score, based on 75 years of major league games. It was published in The Hidden Game of Baseball by Pete Palmer and John Thorn. It looks like this.

OUTSNone1st2nd3rd1st & 2nd1st & 3rd2nd & 3rdFull
0 Outs0.4540.7831.0681.2771.381.6391.9462.254
1 Out0.2490.4780.6990.8970.8881.0881.3711.546
2 Outs0.0950.2090.3480.3820.4570.4940.6610.798

- When Mauer came up to bat, the Twins could have expected to score 1.380 runs.
- Even though he failed to get a base hit, the Twins could have been expected to score 1.371 runs because he at least advanced the baserunners. And that was the (essentially) the worst case scenario.
- With the third baseman playing back, Mauer was trying to bunt his way onto first base, which would have put the Twins in a position to average 2.254 runs, which adds almost a full run to what they could have expected.

Now there’s plenty of details that one can quibble about here. For starters, I only highlighted the two most likely scenarios, and some worse stuff could happen, like Mauer popping up to the catcher. Of course, that works both ways, and Mauer often hits the ball hard enough to ground into double plays.

One might also argue that if you’re going to have roughly 1.3 runs sitting out there for your team to pick up, you would rather have Mauer batting against Weaver than Cuddyer, and I agree with that. That’s a good argument to not have Mauer sacrifice if the third baseman is anticipating the bunt, but not when he’s playing back and opening up a real opportunity for a bunt base hit.

Finally, one additional point that further supports his bunt is that up two runs with Ortiz on the mound and a helluva bullpen, the probability of scoring one or two runs was more important than trying maximize the total number of runs scored in the inning. Even a straight sacrifice is justifiable in that scenario.

We can be pretty sure that while he was in the on-deck circle, Mauer wasn’t sneaking time with an abacus to figure this all out. And that decision might have been made for the wrong reasons, so maybe it will bite back in a future game.

But Mauer’s job at that point wasn’t to drive Weaver from the game, and it wasn’t to justify the state’s man-crush by flexing his muscles. His job was to help his team score some runs, and almost any way you look at it, he accomplished that. We should probably save our rage for an event that deserves it.


Adam said...

He was bunting for a basehit. He's done this before when the third baseman creeps to far back. He should keep doing at as it is a good way to pick up a few cheep hits and it probably keeps some third basemen honest, resulting in a few more cheep hits.

As he said, he just didn't execute the bunt very well.

GMader said...

Is there a similarly handy grid for probability of scoring (one or more runs) as opposed to expected runs produced? That is, something that breaks down the cells of the above-displayed grid by telling us how the 1.38 comes about (18% chance of 0 runs, 40% chance of 1 run, 24% chance of 2 runs, 15% chance of 3 runs, etc. (this distribution of results is completely made up by me).

Corey Ettinger said...

Masterful article.

John said...


Sorry, but if there is, I haven't found it. I'd love to see it if it is. I have the Thorn and Palmer book - I wonder if it's also in there.

There's something similar, which is the Win Probability Added grid. It's much more extensive (because it includes innings), and predicts the probability of winning a game, not scoring a run.

Anyone else have that resource?


JimCrikket said...

I guesss since I'm one of those who got embroiled in this topic on at least one blog, I'll respond to a couple of your points.

First, I don't doubt the statistical validity of the table you referenced. However, when you consider that we had a #3 hitter facing a terrible pitcher, the likelihood of greater success by swinging away would just seem to increase significantly. Give me Mauer facing Weaver in that situation over and over and I have to believe more we'd see more than 1.38 runs per event, on average. Maybe if he lays that bunt down 100 times, he reaches base safely most of the time too, I suppose. Made for an interesting debate, at any rate.

Your comment about those with a "man-crush" on Mauer being in the camp opposing the bunt surprised me, however. The discussions I participated in included a lot of "man-crushes" who seemed to feel the diety known as Joe Mauer was above even questioning his decisions.

WV said...

I know I'm going to sound like the run of the mill anti-stats guy, and trust me I'm not, but there's something to be said for momentum, whether he was sacrificing or going for a hit (which, since he's not exactly a track star, seems questionable). Hitting seems to be contagious, as it was for the Twins in the 1st inning of last night's game, and an out, whether it be a 'good' one to advance the runners, I imagine there's some way to quantify this assertion as well, and since I don't know it I may be easily proven wrong. The figure of 1.38 or 1.37 runs is just an average, meaning that there's higher figures and lower ones that go into its formation. You have to think that Weaver vs. Mauer has a great probability of being an outlier on that chart. But that's my impression, and I'll leave it at that.

One thing's for certain, every time Adrian Beltre tries to advance a base, the expected run output drops significantly.

Anonymous said...

You know there a couple of points that everyone seems to be overlooking.

* Mauer is a good hitter to left field even though he is a left handed batter. By laying down a bunt Mauer for the rest of the year will force third basemen to play up closer and with shortstops having to play closer to second to turn a double play it increases the chances for him to get a hit to left field.

* Mauer is a good bunter. I never heard Rod Carew let alone Kirby Puckett being criticzed for laying down a bunt to try get on. Why the criticism of Joe?

* If Mauer had swung away and hit into a double play will those same people say why didn't Mauer play it safe by bunting the runners over? You can't have it both ways.

Walter Hanson
Minneapolis, MN

JimCrikket said...

With all due respect, Walter, I don't believe we overlooked those points, we simply may not feel they are relevent.

First, nobody is criticizing Mauer for occasionally laying down a bunt for a base hit. It's simply doing so in THIS SITUATION is, to some of us, worthy of critical review. In fact, if I'm managing against him and feel he tends to bunt in that situation, I'll intentionally position my 3B deep enough to tempt him to keep doing it rather than risk giving up a run-scoring hit.

Second, I defy anyone to find a similar scenario where either Carew or Pucket laid down a bunt. By similar situation, I mean while they were hitting #3 in the order (Carew was a leadoff-#2 hitter much of his career), with men on 1st and 2nd, no outs, and a clearly inferior pitcher who's already against the ropes. Those guys typically went for the jugular.

Finally, no, if he had hit in to a DP, I still would not have suggested he should have been bunting. I might suggest (and still do) that he tends to hit in to too many DPs for a hitter of his calibre, but doing so would have simply meant he failed to deliver, not that should have bunted.

Anonymous said...

Is there a similarly handy grid for probability of scoring (one or more runs) as opposed to expected runs produced?

I've seen it at Baseball Prospectus before. Bunting to move two runners over with no outs increases the odds of getting one run, but decreases the odds for more than one run. Navigating their stats is sometimes a difficult chore, but they do provide the run expectancy chart here and it is actually this year's (or other years, take your pick) data.

As for Puckett, it used to drive me absolutely crazy when he bunted, even when it was successful. He used to tell the story about how he planned to bunt in the bottom of the eleventh inning in game 6 of the 1991 World Series. Chili Davis was standing by him and told him not to bunt. Instead, he went up and hit the greatest homerun in Twins history.

Mauer, over the last eight games is hitting 0.417/0.548/0.583. Swing the bat, you are hot.