Bunting on Mariano

You may or may not have noticed -- I probably shouldn't point this out in case you missed it -- but I picked the Minnesota Twins to beat the New York Yankees in the ALDS. I had what seemed to me like solid reasons at the time: I really thought that the Twins would be tough to beat at home, and I thought the Yankees starting pitching problems after C.C. Sabathia would bite them.

More than that, I think I once again underestimated something, something I tend to forget until I see the Yankees play again. Then I remember. That something is this: The Yankees the last couple of years (I think) have put together one of the greatest postseason recipes in baseball history:

The recipe is this:

1 dominant starter

9 or 10 good-to-great hitters to wear down opponents.

1 Mariano Rivera

On the negative side, the ingredients will cost you a lot of money ... you can't even find them at Dean at Deluca. On the positive side, this recipe is so good I'm not even sure you need the dominant starter.

Friday, again, we saw how the Yankees win in the postseason. The Texas Rangers improbably built up a 5-0 lead against C.C. Sabathia and the Yankees, and the Rangers were at home, and the crowd was going crazy, and they STILL lost. Why? Well, they lost in part because Rangers kind of lost their minds. Manager Ron Washington went a little bit cuckoo in the eighth inning as he started throwing out relief pitchers the way a spurned lover throws clothes out a window (Washington used five pitchers in the eighth inning though none of the five happened to be his best reliever).* Ian Kinsler (unconvincingly playing the role of "tying run") got picked off first by Kerry Wood in one of the more bizarre base running blunders of recent times. The Rangers players, as the air grew lighter and lighter, seemed to seize up, both at the plate and in the field. And so on.

*After the crazy eighth inning -- when the Yankees scored five to take a 6-5 lead -- announcer John Smoltz said one of the most curious things I've ever heard a baseball announcer say (and that is saying something). He was trying to make the point that the Rangers needed to put the bad inning behind them, realize that things weren't dire, they were only down one run, they could still win the game. It was a good point to make: Don't panic, don't make too much of things. Only this is what he said:

"If someone had told the Rangers they would be down only one run in the eighth inning, they would have taken that."

Huh? Or to be more specific: Huh? The Rangers would have taken being down a run to the Yankees in the eighth inning? Um, I don't think so. I think it was just a misspeak to make the above point, but I think by saying it that way John actually made several other points that he didn't want to make.

But it seems to me that the way teams continuously collapse against these Yankees in the postseason is no fluke and it's no accident and it's no coincidence. This is all part of the recipe. The Yankees bludgeon teams into mistakes the way Tiger Woods used to strangle major championships on Sundays. The Yankees FORCE teams to go out of character, force them to try absurd things, force them to believe that they had better be perfect or they don't have a chance. The Yankees force it, and teams obligingly crumble.

The Yankees mostly do this with their lineup, their non-stop, no-break, every-inning-is-a-threat lineup -- Jeter, Swisher, Teixeira, A-Rod, Cano, Thames or Berkman, Posada, Granderson, Gardner. You can start that lineup almost anywhere, and it's still better than just about any other lineup in the league starting right at the top. When Posada makes the last out for an inning, for example, suddenly the lineup looks like so:

Granderson

Gardner

Jeter

Swisher

Teixeira

A-Rod

Cano

Thames or Berkman

Posada

Great lineup? Absolutely. No, Jeter is not really a No. 3 hitter at this point in his career -- but he's a first-ballot Hall of Famer, and sure enough he smacked two doubles in the last two innings Friday night. But go ahead, play around with it ... almost any combination of those nine hitters makes for a scary inning. Pitchers can work through three innings, five innings, seven innings, but sooner or later odds are that lineup is going to score runs, especially during that modern-baseball-era gap between the starter and the closer.*

*As I write these words, the Rangers have just taken a 5-0 lead over the Yankees in Saturday's game. So, you could say that adding too much Phil Hughes could mess up the recipe. Then again, it's only the third inning.

Here's another way to think about it: Imagine your team, whatever your team is, down a run in the late innings. As a fan, you probably have a certain place in the lineup that you hope is coming up. If you're a Rangers fan down a run, you probably would hope that Josh Hamilton, Vlad Guerrero, Nelson Cruz are coming up. Something like that. A Phillies fan would probably hope to get Chase Utley, Ryan Howard, Jayson Werth up there. A Giants fan needs Aubrey Huff and Buster Posey. Even bad teams, say the Royals, would hope to get David DeJesus and Billy Butler to the plate.

But for the Yankees ... it just doesn't matter. Sure, they might want Teixera, A-Rod and Cano, but it's certainly no problem if they get Posada, Granderson, Gardner, Jeter, Every inning they have the heart of the order up. They come crashing at you like waves hitting the shore.

Then, of course, if they have the lead in the ninth inning they send Mariano Rivera out there and that's that. I don't know how much stock to put into the Mariano Effect -- that teams not only can't hit Rivera but also try too hard in earlier innings because he's always lurking -- but I do suspect that Rivera plays on the mind. I've mentioned before that before the movie Gandhi came out, there was a real push in India to have him portrayed only by a ray of light, that he was too remarkable to be represented by a mere actor. With Rivera, I keep expecting a ray of light to come trotting in out of the bullpen.

So -- a lineup that will eventually get you, the best postseason reliever in baseball history, and (as a bonus) a few hundred million dollars in starting pitching -- I'm not sure why I keep underestimating the Yankees. Even now, as I look ahead, I think the Phillies have the best team in baseball, and that remarkable Phillies starting pitching could neutralize the Yankees recipe. I keep thinking that the Rangers, if they can just get Cliff Lee out there, would have a shot of winning this series in seven games. But, you know what? Until I see someone break this particular Yankees blueprint for postseason success, I should probably assume that no team can beat them over a seven-game series ...

All of this was just supposed to be a prelude to my real question of the day which is this: Is it smart to sac bunt against Mariano Rivera?

I was thinking about this Friday night, of course, because the Rangers bunted. Man on first, nobody out, needing one run to tie, the Rangers' Elvis Andrus got down a successful sacrifice bunt (with two strikes). The Rangers, of course, did not score the tying run, did not even manage to get the runner to third base. But that doesn't mean it's right or wrong.

My first thought was: Wrong. It has to be wrong. We all know how good Mariano Rivera is ... just giving him an out, it seems to me, is like giving Usain Bolt a head start. The only thing you have against Rivera are your three outs ... they are precious, they are rare, and you have to use them absolutely as well as you can. To give them up for one base seems to me a bad deal.

But ... the more I thought about it, the more wisdom I got from wise people like Tom Tango, the more I realized that this topic is a lot more complicated than that.

First, let's state the obvious: Giving up an out for a base, the vast majority of the time, is a lousy deal. This is why you don't see it happen very often -- there were only 1,500 or so sacrifice hits in more than 185,000 plate appearances this year.

The numbers change, but generally speaking (I have been toying with Tom Tango's run calculator to come up with these numbers):

-- With a runner on first and nobody out, a team will score about 42-44% of the time.

-- With a runner on second and one out, a team will score about 40-42% of the time.

Of course, it depends on the quality of hitters coming up, the quality of pitcher on the mound and various other things, but in a mathematically precise world the gaining of second base and losing of an out DOES NOT give your team a better chance of scoring a run. At most, it's a break-even. If anything, it gives you a lesser chance. And, beyond the "how often you score" issue, it DEFINITELY hurts "how much you score." This is why the sac bunt drives so many of us crazy, especially in the early innings, especially when you waste a good hitter by bunting*.

*Interestingly enough, bunting a runner from second to third with nobody out -- a move I very openly despise -- DOES accomplish that one limited goal of scoring more often. A team with a runner on second and nobody out should score 59-61% of the time. But a team with a runner on third and one out score score 67-69% of the time.

Now, I still think this is a lousy move most of the time because your overall expected runs goes down -- this relates to the classic line about how if you play for one run that's what you'll get. But if you are a manager who wants or needs that one run and only that one run, then it seems by the numbers I've run that in many, even most situations, bunting a runner from second to third isn't as bad a play as I've always believed.

OK, so that's the general bunting scenario. But what about a specific question like this one: Is it worth sac bunting against Mariano Rivera in the ninth inning, down one run, in a postseason game. One of the problems with answering this question is that people tend to oversimplify the sac bunt, tend to turn it into a two-part multiple choice issue: Bunt works, bunt doesn't work.

But that's not realitiy. There are several other possibilities -- here are eight of the more common sac bunt possibilities:

1. The runner moves to second, batter's out, sac bunt.

2. The runner is thrown out at second, batter's safe at first.

3. The runner moves to second, batter's safe at first (a single or an error).

4. The runner moves to third on bad throw, batter's safe at first.

5. The runner is thrown out at second, batter's out too, double play.

6. The runner stays at first, batter pops up bunt.

7. The batters fails to bunt on first two tries, hits away two strikes.

8. The batter fails to bunt on first two tries, bunts again foul, strikeout.

Any and all of these are possibilities and each has its own value. A speedy runner who could turn a sacrifice bunt into a single 20% of the time would change the whole formula. A bad bunter who will foul off the first two bunt chances 75% of the time would change the whole formula. And so on.

So let's simplify the Rivera question even more -- let's make it this: is it worth it in the larger sense to give up an out to move a runner to second base against the best postseason closer ever? So for this we assume 100% success rate on the sacrifice bunt, and we also assume the batter is out 100% of the time.

As it turns out, Tom Tango put this EXACT table in The Book -- a Mariano Rivera scoring distribution table. According to the table:

Runner on first nobody one: Team will score 37.4% of the time.

Runner on second one out: Team will score 36.2% of the time.

So that seems to settle things -- your percentages go down. Only, maybe not: As Tom explains, the better the pitcher gets, the percentages get closer and closer until finally, at some point, they flip and you actually have a better chance of scoring with a runner on second and one out than you do with a runner on first and nobody out.

Tom's Rivera chart refers to the 2007 version of Mariano Rivera, when he gave up 3.2 runs per game. It does not refer to his postseason work. In the postseason, Rivera has an 0.72 ERA in 137 innings of work. He has given up two unearned runs on top of that, so he has given up .85 runs per nine innings.

So, with Tom's help -- and by help I mean "Tom did this" -- we calculated some percentages using Rivera's numbers in the playoffs.

If you use Rivera's exact postseason numbers -- this is assuming he is and will be as good as he has been in the postseason -- the math looks like this:

Runner on first nobody out: Team will score at least one run 21.8% of the time.

Runner on second, on out: Team will score at least one run 26.8% of the time.

Even if you tinker with the numbers, make Rivera a bit more hittable, the bunt still works out as a good play.

And there you go. It really does seem that if you need one run against Rivera you do have a better chance of doing it with a runner on second and one out than a runner on first with nobody out. Of course, it's not that simple. How easy is it to bunt successfully against Rivera? Where are the Yankees playing? Who is coming up? And so on.

But I do think that in a game where Ron Washington made some, er, unusual moves, well, I think my initial reaction in the ninth inning was wrong and I think Washington probably made the right mathematical call by having Andrus bunt.