IGT 7/30/2013: NYM @ MIA -- Abyss You Much

[Guest LeiterWagnerFasterStrongr]

smg58 wrote: Hmm. I don't sacrifice with a guy who's 4 for 4 getting on tonight, and I don't intentionally walk Q to face two better hitters with the bases full.

All the more confounding, these, since Gary and Ron praised both moves so effusively, I thought the half-inning was sponsored by the American Smallball Council.

But hey, I'll take a stupid win anyday.

[Edgy MD]

I don't know. I mean, I'm against bunting, but let's not fall into the notion that who Lagares had been earlier in the evening against other pitchers (actually 3-3, with a walk, not 4-4) define who he was at that moment.

As for walking the Q, well, there was a whole lot to be said for setting up the double play. Extra-innings are full of such stuff. That marathon against St. Louis that should have been Jerry Manuel's last game, it was going on all night. Guy reaches base, guy gets bunted over, next guy walked to set up double play.

[Guest LeiterWagnerFasterStrongr]

Leave aside the hitter that he was tonight 'til then... he's still a surer bet than Quintanilla, right? (And certainly a better bet to score runs than Quintanilla-with-one-out.)

And as to the second one... the odds that Quintanilla's doing something that isn't getting on-base safely are about 70%. That he's something of a weak spray hitter makes the odds that he won't do something useful with the ball slightly south of that, probably, right? Maybe, say, 60%? What are the odds that an average-to-slow hitter like Satin/Brown hits into a double-play with the corners in and the bases loaded? Are they higher than, say, 60%? (HINT: They're somewhere in the 20-30% range.)

[nymr83]

Edgy MD wrote: I don't know. I mean, I'm against bunting, but let's not fall into the notion that who Lagares had been earlier in the evening against other pitchers (actually 3-3, with a walk, not 4-4) define who he was at that moment. As for walking the Q, well, there was a whole lot to be said for setting up the double play. Extra-innings are full of such stuff. That marathon against St. Louis that should have been Jerry Manuel's last game, it was going on all night. Guy reaches base, guy gets bunted over, next guy walked to set up double play.

I look at Lagares as actually trying to bunt not just for a sac but for a hit there with the side effect of moving the runners up if he fails. i hate bunting, but this wasnt the worst bunt, particularly in a situation (already up 2 runs in the last inning) where increasing your probability of scoring 1 more run at the cost of your probability of adding more than that might not be the worst trade to make.

[Ceetar]

Nymr83 wrote: Edgy MD wrote: I don't know. I mean, I'm against bunting, but let's not fall into the notion that who Lagares had been earlier in the evening against other pitchers (actually 3-3, with a walk, not 4-4) define who he was at that moment. As for walking the Q, well, there was a whole lot to be said for setting up the double play. Extra-innings are full of such stuff. That marathon against St. Louis that should have been Jerry Manuel's last game, it was going on all night. Guy reaches base, guy gets bunted over, next guy walked to set up double play. I look at Lagares as actually trying to bunt not just for a sac but for a hit there with the side effect of moving the runners up if he fails. i hate bunting, but this wasnt the worst bunt, particularly in a situation (already up 2 runs in the last inning) where increasing your probability of scoring 1 more run at the cost of your probability of adding more than that might not be the worst trade to make.

I believe the run expectancy matrix suggests that bunting with first and second and no outs actually leads to more runs than not. Well, successfully bunting does anyway. But that factors in average hitters, not Quintanilla and we don't know.

[batmagadanleadoff]

Ceetar wrote: I believe the run expectancy matrix suggests that bunting with first and second and no outs actually leads to more runs than not.

Impossible. I don't even need to look at any matrix to know this. How can an out increase the number likelihood of total runs scored in the inning? Unless the batter bunts with the bases loaded and the fielder takes so much time to get the only out at first base that in the meantime, all three runners score.

[Frayed Knot]

I think what a run analysis may show is that a bunt in that situation increases the likelihood that you'll score [u:n64jz06l]at least one run[/u:n64jz06l] even if the [u:n64jz06l]average number of runs[/u:n64jz06l] decreases overall -- and the argument can be made that an extra run or two there vastly increases the odds of winning the game so it's the better move than playing for the larger inning.

[nymr83]

Ceetar wrote: Nymr83 wrote: Edgy MD wrote: I don't know. I mean, I'm against bunting, but let's not fall into the notion that who Lagares had been earlier in the evening against other pitchers (actually 3-3, with a walk, not 4-4) define who he was at that moment. As for walking the Q, well, there was a whole lot to be said for setting up the double play. Extra-innings are full of such stuff. That marathon against St. Louis that should have been Jerry Manuel's last game, it was going on all night. Guy reaches base, guy gets bunted over, next guy walked to set up double play. I look at Lagares as actually trying to bunt not just for a sac but for a hit there with the side effect of moving the runners up if he fails. i hate bunting, but this wasnt the worst bunt, particularly in a situation (already up 2 runs in the last inning) where increasing your probability of scoring 1 more run at the cost of your probability of adding more than that might not be the worst trade to make. I believe the run expectancy matrix suggests that bunting with first and second and no outs actually leads to more runs than not. Well, successfully bunting does anyway. But that factors in average hitters, not Quintanilla and we don't know.

1st and 2nd nobody out is 1.556 runs (with the chance of scoring at least 1 at 64.3%)
2nd and 3rd 1 out is 1.447 runs (with the chance of scoring at least 1 at 69.8%)

so like i said, you are increasing your chances at 1 run while decreasing your chances at more. Also, these numbers assume the bunt was successful. if the bunt is unsuccessful and they get the lead runner (or you pop it up and they get you without runners advancing) your expectancy is down to 0.963 and 42.9%... other things can go wrong too (bunt into doubleplay, bunt foul and lower your chances of getting a hit when forced to swing with extra strike)

Frayed Knot wrote: I think what a run analysis may show is that a bunt in that situation increases the likelihood that you'll score at least one run even if the average number of runs decreases overall -- and the argument can be made that an extra run or two there vastly increases the odds of winning the game so it's the better move than playing for the larger inning.

Yes, thats what the tables show and I agree (as per my original comment that this wasnt the dumbest of bunts) that the strategy may be an ok one here. but i think we can easily refute the silly notion that bunting "leads to more run than not"

here is an article with the charts (i referenced the most recent of the 3 data sets)
http://beyondthescorecard.blogspot.com/2013/04/tom-tangos-run-expectancies.html

[batmagadanleadoff]

Frayed Knot wrote: I think what a run analysis may show is that a bunt in that situation increases the likelihood that you'll score at least one run even if the average number of runs decreases overall -- and the argument can be made that an extra run or two there vastly increases the odds of winning the game so it's the better move than playing for the larger inning.

There's two different matrixes for run expectancies. One matrix returns values for likelihood of overall number of runs scored in the inning for each base/out situation. According to this matrix, a sacrifice -- an out --always decreases the likelihood of overall or total runs scored in the inning.

The second type of run expectancy matrix provides the chances of each specific base runner scoring. According to this matrix, a runner on second base with nobody out, who is then sacrificed to third base is likelier to score. This is the only instance when a specific base runner would be likelier to score after advancing one base on an out. So if you're playing for one run, (e.g. as the home team in a tie game after the eighth inning) it makes some sense to try and sac bunt the runner on second if nobody is out. But using up an out to move that runner on second with nobody out decreases the overall or total number of runs the team could be expected to score in the inning.

[dinosaur jesus]

For what it's worth, Baseball Reference give the bunt as having no effect on the likelihood of winning (which was 96% at that point). It's a low-risk, low-yield strategy, but that actually makes some sense in that situation. Normally you want to try to score as many runs as possible, but when you're already up by two in the ninth, there's no real need to. There's not much difference, as far as winning, between one more run and two or three more.

[Frayed Knot]

But using up an out to move that runner on second with nobody out decreases the overall or total number of runs the team could be expected to score in the inning.

Which is something I wasn't arguing.

[batmagadanleadoff]

Frayed Knot wrote: But using up an out to move that runner on second with nobody out decreases the overall or total number of runs the team could be expected to score in the inning. Which is something I wasn't arguing.

That's right.

[Ceetar]

batmagadanleadoff wrote: Ceetar wrote: I believe the run expectancy matrix suggests that bunting with first and second and no outs actually leads to more runs than not. Impossible. I don't even need to look at any matrix to know this. How can an out increase the number likelihood of total runs scored in the inning? Unless the batter bunts with the bases loaded and the fielder takes so much time to get the only out at first base that in the meantime, all three runners score.

You're right, but it's not that far off base. It's not a big different in expectancy and if you look at the second chart you'll see that in that case bunting makes sense. (Or, perhaps, that teams of that age bunted so much in that spot that it's skewing the numbers?) Now, 2013 is closer to the 93-'10 chart, but the last time I looked run scoring overall is down as compared to those years, which shrinks the difference even more. Add in Quintanilla/PH being below the average player and it's probably close to a wash.

The chance you score 1 run DOES go up by about 5%, which seems a little negligible in my mind. It's not a bad call really, not one I'd make, but not a bad one.

[Edgy MD]

I don't think that's negigible, but it has to be weighted against the likelihood of success, and bunting success rates are not something I've ever seen good data on. Intuitively, I'd say almost nobody bunts well enough to make the 5% bump worth it.

But what really matters isn't the expectancy for one run versus the expectancy for two or more, but rather the win expectancy. If you don't know that, it's hard to determine whether an increased chance at one is worth a decreased chance of two or more. I'd say maybe, but again, the success likelihood of the bunter is crucial to the decision.

[Edgy MD]

Yesterday was also Duaner Sanchez day. Always nervewracking to be in Florida for Duaner Day.