Dave, you say that the Yankees’ 92% win probability was probably higher because of Rivera’s presence. Because that win probability is “average win expectancy for a major league club,” is that Rivera factor countered in part based on the fact that the Phillies have an above-average offense?
Also, the seven percent increase in the likelihood of scoring a run with a successful sacrifice is indeed important, but that doesn’t factor in the possibility of a failed sacrifice, which is exactly what turned out to happen.
Is the win probability with Derek Jeter hitting, no outs, a 2 run lead, and men on first and second really lower than with Johnny Damon hitting, one out, the same 2 run lead, and men on second and third? Jeter is one of the Yankees three best hitters and one of the better hitters in baseball. Maybe I’m wrong, but I’d rather have the first situation than the second.
Also, if Jeter gets a hit and the Yankees end up putting a crooked number on the board, maybe you don’t need to ask Rivera to get 6 outs. I know he hasn’t pitched in a while, but he is almost 40 years old.
Fully agree, some times people need to remember who the personnel is on the field. Win prob charts show that bunting late is actually not a crazy idea when the differential is -1 to +1. It’s even less crazy when a Hall of Fame reliever is pitching.
I know discussing what might have happened is worth very little, but lets not forget: supposing Jeter gets the bunt down and Cabrera ends up scoring, the score sits at 4-1. Should Matt Stairs not hit (and look) like Gary Busey at the plate and take Mo deep, you’ve still maintained your lead at least in small part to Jeter’s sacrifice. The process, to be sure, is fair game for criticism, but it isn’t a stretch to argue that championships are won on equal parts process and outcome.
I think that Rivera is leading the pack of the discussion when you talk about the best postseason PLAYER of all time. I feel like I could really only make an argument against him with Mantle, and even that is pretty hard to determine.
I’d say that as a Mets fan who just wants the Phillies to lose, I felt more assured that the Yankees would win this game with a 2 run lead and Rivera than I felt that the Mets would win any game with K-rod coming on with a lead of 4 or less runs.
What about when a Hall of Fame hitter is hitting? –or in this case, bunting.
Comment by Richie Abernathy — October 30, 2009 @ 12:38 pm
That’s my big problem with it. With Jeter up, the Yankees are squandering a chance to put the game away. Is the difference between 3-1 and 4-1 really so great that you’d throw that opportunity away? I’d rather go for 5 or 6-1, but that’s just my opinion.
It also fails to account for the how the Yankees’ likelihood of winning changes as a result of Jeter swinging away. I’m not crazy about the bunt attempts but it was defensible given the situation and the 2-strike attempt was beyond horrific. I am glad, however, to see Dave take on “conventional” sabermetric wisdom by defending the bunt attempt. That’s what makes for thoughtful debate.
Yeah if the yankees busted the game open there, perhaps Rivera wouldn’t have been used for 40 more pitches. He is almost 40 years old and is getting used way more this time of year than in any recent years. This could catch up with him eventually.
Also, Jeter was facing Chan Ho park. I always want Jeter to swing against Chan Ho park.
I consider this an example of a case where there is not so much a right or wrong answer as there are a variety of reasonable options.
I am among those who dislike sacrifice bunts but am not dogmatic about it, at least not opposed in all instances. My first reaction when Jeter came up was “don’t bunt”; it is Jeter against Park and this could be the inning that puts the game away so that Rivera is either not needed or can be given just one inning of non-stressful work.
When he seemed ready to bunt at the first pitch, I was not happy, but also was not irritated or angry. Instead, I recognized it was a defensible judgment on his or Girardi’s part. Of course the effort after strike 2 made no sense to me, but up to that point I considered it simply a matter of my preference, not right or wrong.
In fact, in the top of the 8th Utley came to bat as the go ahead run and had the play been correctly called, Howard would have also come up as the go ahead run. No matter how great Mariano has been and is, the top of the 8th presented a serious threat to any 2 or even 3 run lead, so I think it perfectly valid to prefer going for the big inning against Park and the Phillies’ bullpen. I just don’t think it is clearly the only right move.
I refuse to call a closer or any relief pitcher the “best” at anything besides comparing him to his contemporaries.
Comment by Brian Recca — October 30, 2009 @ 2:03 pm
Steve – And the next sentence was, “I’d rather go for 5 or 6-1, but that’s just my opinion.” So I’m saying I disagree with it the conclusion of the article. Is that really so hard to see? (And no I’m not trying to sound like a jerk either.)
What about the fact that Jeter hit 334 with an over 400 OBA and was facing a not very good pitcher? Even in the era of “small ball” did mgrs instruct a hitter with such numbers to lay down a sacrifice?
I can’t say I agree that there’s no difference b/w a 3-run or 5+ run lead. One substantial difference as noted is that Rivera wouldn’t have had the stress of throwing 40 pitches. This outing could have carry over effect.
Baseball Prospectus has a post up about this as well. Their stats show that bunting in that situation isn’t a good idea. This is from their unfiltered section so it’s not behind their paywall:
Why even risk bunting in that situation? According to run expectation, with runners on first and second and no out, a team is expected to approximately 1.5 runs in an inning. When there are runners on second and third with one out, the run expectation actually goes down to 1.4. There’s always a chance that Jeter hits into a double play—Jeter ranked 20th in the league in DP% with 17 percent—and he is a ground-ball hitter, but when you’re in the World Series and have one of your best hitters at the plate, it’s better to take the risk.
Slightly off topic, but how cool was it to see Melky attempt a bunt (McCarver thought it was a fake, I kinda doubt it), get to an 1-0 count, then rope a single on a pretty fat pitch with the infield in after reading through MGL’s epic the other day?
I agree. Furthermore, if the Yanks had blown it open that inning they might have been able to avoid using Rivera for 2 innings. Or, considering the fact that Rivera threw 23 pitches in the 8th, he could have tired and run into some trouble in the 9th. It’s no guarantee that 1 additional run was going to be enough. And your point about the poor umpiring call was correct as well.
And if Girardi is intent on pitching Mo for 6 outs in close games and you have 3 games in a row coming up, you’re taking a significant risk of run him down, not having him available or having him be ineffective. I suspect he’ll be fine for Game 3. But if he has another long-ish outting in game 3, I’m not sure what to expect from him in game 4.
Come on. He could reach the base. It absolutely wasn’t interference and the only reason for complaints like these is to distract from the very valid comment that the win wasn’t necessarily in the bag when Rivera came in.
I don’t think any sabermetrically minded fan believe it’s a bad idea to bunt with runners on first and second and no one out. I think it’s pretty widely accepted that this is one of those rare situations where a successful bunt increases the expected run total, and, more importantly, the probability of scoring at least one run. I think the criticism of the bunt stems from Jeter bunting with two strikes (obviously a bad idea) and Jeter being such a good hitter (perhaps negated by the situation and his DP tendencies), but there’s no personnel specific discussion here, i.e. no discussion of the probability Jeter lays down a successful bunt, the Phillies relief pitching, the Phillies offense, or any attempt to measure how much having Mo as your closer affects decision making about whether to play for one run (my guess: not very much). Pretty lame analysis.
Both pictures are extremely poor evidence since the angle is so bad. It’s taken from a camera looking toward left center field rather than one taken straight across from first base. Regardless, the only purpose of this futile argument is to distract from the main point of the article or from the comment to which the original complaint about interference referred.
Like the post and enjoyed the discussion, especially regarding the statistical theory on bunting. I can’t get beyond the opinion sacrificing Jeter was a bonehead decision by Girardi. Even before Derek struck out on the foul bunt attempt with the count 0-2.
Given the game situation, where the Yankees were in their lineup, and Chan Ho Park, Jeter should have hit away.
Park had just hung a pitch to a cold Posada, on a 1-2 count no less, and watched it be sent to CF on a line. Odds were good he’d hang something to Jeter too.
But that’s really not what you’ve done, because the second run was on first and the first run would have wound up on third. Since the bunt only makes sense if you overvalue the first run in this situation relative to other situations, it’s still moving the guy from second to third. Point on taking out the double play, though.
I disagree that the 2 strike attempt was so monstrously bad. Jeter is a career .231/.250/.331 hitter with an 0-2 count; given that the defense would be playing back with 2 strikes and not prepared for a bunt, the chance of him reaching on a bunt were actually much higher at that point than earlier in his at bat. Good bunters generally reach on 40-50% of their bunts for hits (with the infield presumably pulled in several feet further than they were playing for Jeter), so Jeter could probably equal his OBP when swinging away by getting down only half, or perhaps slightly more, of his bunt attempts – and this is before even beginning to consider the benefits of the bunt as a sacrifice.
The dumbest thing about the whole situation was actually Jeter coming out and saying that it was a dumb play – he should have said that he thought it was a good play that he didn’t execute, so that regardless of what he thought, the Phillies (and other teams) would have to take that into consideration when positioning their defense should the situation arise again (as I’m sure MGL would argue).
I have to disagree with the analysis here.
I don’t believe that you can simply look at a run matrix and compare expectancies based on historical data.
If the overall probability of scoring a run with 1,2 0 out is 62% that includes all situations.
It seems to me that the probability of a .340 hitter driving in a run while intentionally making an out is about 0%.
Therefore, the probability of a .340 hitter driving in a run given that he is swinging away must be significantly higher than 62% to bring down the overall % to that rate.
Obviously, I know it is not reliant on him driving in the run, but if he does drive it in that certainly a positive.
I just don’t see how this move is defensible.
But aren’t the run expectancies still misleading?
That 62% includes situations where much worse hitters are batting against much better pitchers.
It also includes situations where bunts are laid down and the odds of that batter driving in the run is essentially 0.
Doesn’t it stand to reason that by letting Jeter swing away it may nullify the difference is run expectancy you see when comparing the 2 situations (1,2 0 out and 2,3 1 out) in a vacuum?
Dave says:”The cost of the sacrifice bunt is in the reduced chance of a multi-run inning, but in that situation, there really wasn’t a tangible difference between a three run lead and a 10 run lead. Those additional runs that could have scored in a big rally would have been essentially worthless.”
Bollocks. I say. Excuse my language. But there is a world of difference between having to use a 40 year old Rivera for tense two inning save and being given the opportunity to find out if Hughes or Joba can fix themselves (say with a 10 run lead as a buffer).
Plus, he didn’t have to get a hit. You can still get the runner to third with less than 2 out by swinging. Plus the added upside of him actually getting a hit. It seems that you are giving up more to avoid the potential downside of not advancing the runner or a DP.
From MGL’s article on bunting (that Dave links to):
“Obviously if I told you that batter A was so good at bunting and so fast, even in a sacrifice situation, and that inexplicably, the defense was not expecting a sac bunt, that he got a hit or an ROE 40-50% of the time (about the same percentage a good bunter bunts for a hit), you would realize that a sac bunt was an excellent play.”
Not sure what his source is, of course, but I see no reason to doubt it.
Even accepting the valid point that Park is a good reliever with the platoon advantage, I still maintain that the options are both reasonable, that sacrificing or hitting away both make enough sense that we basically come down to preferences or judgments, not right and wrong decisions.
This is not meant to disparage the use of statistical analysis in the least. But that analysis leaves room for interpretation and for placing stress on different factors-Park’s advantages, Jeter’s advantages, Rivera’s excellence, Rivera’s workload, chances for the unexpected to happen (bloop hits for example), the opportunity to blow the game open, the opportunity to give Rivera a slightly larger cushion and many more factors a manager may weigh.
The fact of the double play is not relevant. Not only was it unlikely given Utley’s history (among other things), but whether it should have been one or not does not obviate the point that it could have been called the other way, an example of the unexpected if you will. And in any case, the Phillies did have Utley at bat as the go-ahead run which is probably an even better situation for them than having Howard up.
“The dumbest thing about the whole situation was actually Jeter coming out and saying that it was a dumb play – he should have said that he thought it was a good play that he didn’t execute, so that regardless of what he thought, the Phillies (and other teams) would have to take that into consideration when positioning their defense should the situation arise again (as I’m sure MGL would argue).”
Jeter always says the right things in the media….not this time. The Captain goofed. Shocking.
Should the same situation present itself again this series, is he now obligated to repeat bunting with two strikes?
I think you misunderstand the meaning of expected runs. 62% is the percentage that the offensive team scores one more run in the remaining inning, so even if the batter lays down a bunt, the percentage to score one more run thereafter is certainly not 0%. In fact, as Dave pointed out in the article, if the bunt is successful, with runners on 2nd and 3rd, 1 out, the percentage to score one more run jumps up to 69%.
I’m not sure the Captan did goof. Maybe he was saying it was the wrong play to make sure the Phillie defense stays back on 2 strike counts in possible bunting situations again?
One thing I wonder about Preston/MGL’s comments about a good bunter getting a hit or ROE in 40-50% of his attempts takes into account (usually) 2 attempts per PA. What I’m thinking is that a good bunter probably only reaches on 40-50% of his bunt attempts that go into the field of play and not foul or missed completely. I just really doubt that stat is taking these other factors into account. Which then means if this same good bunter where to try and bunt with a 2 strike count we’d see his fail rate go up as fouls or misses are no longer just a change in the count and you essentially get a do-over, but outs. I could be wrong, maybe that data MGL was siting was for any particular pitch, but again, I doubt it.
I actually understand completely.
I am just speculating – and I may be completely wrong – that if teams never bunted with runners on 1st and 2nd and 0 out, the chance of scoring a run that inning would be higher than 62% and maybe even higher than 69%.
I say this because because you now have only 2 outs to get a run home rather than 3 since the possibility of the next batter getting a run home is essentially 0.
It should be possible to test this on historical data by checking how often a run scores in an inning following a 1,2 0 out situation given that the next batter does not bunt and compare it to time when he does.
Has this already been examined?
I think so, although not as ill-advised. First, Rivera only had to pitch one inning to preserve the lead and second he was to face the bottom of the order. I still don’t like it, but the fact that 2 errors led to 2 extra runs does not mean it was the right thing to do.
In fact, it kind of emphasizes the foolishness of it. How often do major leaguers muff plays the way the Angels did? Had usual results occurred, the Yankees would have scored no runs that inning after getting the lead runner on. In fact, even after the first error, had things gone normally, the Yankees still would not have scored any runs.
Naturally such speculation is idle. We have no idea what would have happened had one factor been changed. But the point is that playing for one run in that situation was probably ill-advised regardless of the actual outcome.
As I understand MGL’s use of the statistic, if a batter is bunting for a hit, he generally reaches on 40-50% of bunts that he successfully puts in play; my conjecture is a similar figure would apply here given that, on the one hand, the infield would be less optimally positioned for a bunt than they would be against a hitter who frequently attempts bunt hits, while on the other hand, Jeter would be primarily looking to successfully execute a sacrifice bunt and therefore would be less concerned about ideal placement for getting a hit.
The million dollar question, of course, is how often Jeter will get the bunt in play in that situation. Given a .231 BA and .331 slugging with an 0-2 count (it strikes me that OBP is actually irrelevant here, as you can walk while bunting just as well as while swinging away), it seems to me like he’d probably need to put the ball in play at somewhere around a 50% rate for it to be worth it – my suspicion is that he’s perfectly capable of doing that, but there I have absolutely no statistics, so it’s a total guess.
Not likely. The frequency of scoring exactly one more run increases with a successful bunt, so as long as a failed bunt is rare enough an event, the overall frequency would increase when the batter chooses bunting instead of swinging away. Therefore the latter choice must lower the overall frequency.
Comment by tradewind — November 1, 2009 @ 12:03 am