Oakland and Detroit in a very good ALDS game 2

Posted: October 6, 2013 in In-game strategy, Managers, Post-season

If you like 1-0 games, that is. You would think that in a singleton, mundane 1-0 game, there would be few controversial managerial decisions. And you would be wrong. And you might think that I think that they got all of them wrong, which they are usually wont to do. And you would be wrong again! (Only Maddon – Joe – does that.)

Here are a few that I happened to notice as I occasionally passed by the TV on my way to and from the kitchen looking for something to eat – to no avail (I did find – and eat – some disgusting gluten free frozen pieces of cardboard frozen waffles):

Disclaimer: If you read my posts this morning on The Book blog, you might find this redundantly redundant.

In the bottom of the 5th with the game still tied, Reddick comes up with runners on first and second and no outs. Melvin was clearly determined to have Reddick bunt since he squared on the first pitch and bunted on the second, and even tipped his hand each time Verlander spun towards second I think. Now, regardless of what you are going to do, tipping your hand to the defense is absurd. It allows them to play back or up accordingly, or at least more back or up then they might if you DIDN’T send them a text explaining what you were going to do. So, that IS a mistake, one that most managers make. Some, like LaRussa, actually have the batter square early, in a potential bunt situation, then take it back, and then either square again (when bunting) or don’t (when not bunting). I like that. You can also just do nothing until the last possible instant, or however long it takes to get into position to bunt, if that is what you are going to do.

Now, was the bunt itself correct or reasonable? I have no idea, and neither do you. And neither does Melvin, or Leyland, or Ripken, or anyone else. I could possibly figure it out – to some degree. They can’t.

What factors go into it? Read my 40 page chapter on bunting in The Book if you really want to know. Here is a summary: Is the batter a good and fast bunter? Is the pitcher hard to bunt against? What is the expectancy of hitting away, vis-a-vis the batter and the pitcher? How good are the defenders, especially the third baseman and pitcher? Where is the third baseman playing? And with a runner on second, as in this case, is the batter RH or LH? The last point is because a lefty has more of a chance to move the runners on an out when NOT bunting. And that, my friends is why neither I, nor Melvin, nor Stephen F. Hawking (an example of a really smart guy) knows what is correct there, without at least going some serious research and number crunching. My guess is that is was NOT correct because Reddick is a lefty, he doesn’t bunt much, and Cabrera was playing up a fair amount. If you say anything in the comments section about how it turned out, you will be the first one banned from this blog!

Now, after the first pitch is a ball, if a bunt were a marginal play to start the AB, then it has to be incorrect to keep bunting, unless the defense changes position. That should be obvious. At a 1-0 count, hitting away is much more profitable and the bunt attempt is probably only slightly more profitable (bunts are not that sensitive to the count, at least until 2 strikes).

OK, done with that.

Curiously, in the bottom of the 8th with the game still tied, Crisp comes up with a runner on second and no outs and makes no attempt to bunt whatsoever. Of course the decision is on the manager, Melvin. I say curiously because Crisp is a weaker hitter, quite weak in fact from the right side, is a very good bunter with good speed, and he is hitting from the right side which makes it harder to advance the runner on a batted ball. And like Reddick, when the count changes, he should probably change his strategy. Again, if it is a marginal no-bunt at first, when the count goes to 0-1, the bunt is probably in order, since, as I said, the bunt attempt is not that sensitive to an 0-1 count, but hitting away certainly is.

Oh, and BTW, if you cite any RE or WE numbers with no bunt and after a “successful” bunt, I will ban you too. I mean you can estimate fairly accurately the hitting away WE or RE with each of those batter and pitcher combos, but good luck trying to estimate the WE or RE of a bunt attempt. You better have some idea, at least, of the percentage of time the batter does A, B, C, D, E, F, and G when he attempts a bunt (and what happens when he gets to 2 strikes and has not gotten the bunt down).

Speaking of bunting, here is a short treatise on bunting and Game Theory:

One of these decades managers on defense and offense are going to think of bunts like playing a poker game (or rock, paper, scissors – a game they can actually understand) – trying to establish the Nash equilibrium point, or more accurately, trying to start with Game Theory Optimal (GTO) strategy and then, if your opponent does not do the same, or you suspect that he will not, exploiting him with some other strategy.

Any time there is a potential bunt situation, whether it be in a sacrifice situation or not, the defense must play in a way that makes the bunt and the non-bunt exactly equal in WE, OR, if they cannot do that, minimize their opponent’s WE by playing as far back or as far in as possible.

The offense meanwhile should do the same. They should bunt or not bunt in a certain ratio such that it does not matter where the defense plays! That is the definition of a GTO strategy – one that makes your opponent’s decision irrelevant to the expected outcome.

In practice, if the defense does not play in a GTO manner, i.e., they are playing too far in or back, then you have to decide how much you want to deviate from the GTO ratios. You cannot deviate all the way to bunting or not bunting 100% of the time, even though it is technically correct in that one situation to do so, unless that is the last time you are ever going to play in that situation. You can however, deviate more or less depending on the importance of the game!

Teams also need to stop thinking of the sacrifice bunt as any different from any other bunt. The goal is NOT to get out! The goal is to bunt the ball in such a way that you maximize your WE, if you are going to bunt. That may or may not mean that you should square early (losing the surprise element), it may or may not mean that you MUST bunt to the first or third baseman (depending on where the runners are), it may or may mean that you don’t try and get of the box quickly, and it DEFINITELY does NOT that you try and run harder when you are bunting for a hit!

Here are a couple more controversial decisions (in MGL’s mind at least):

Classic, and common, mistake by Melvin. Most people don’t consider it a mistake or even a decision. Mediocre starter is cruising through 6 or 7 innings, so you let him continue. Why not? As long as his pitch count is good and you don’t think he is tiring, right?

One, starters who cruise through 6 or 7 pitch no better in the 7th or 8th than those who don’t or at any other time during the season. Two, you have, in addition to that, the serious “times through the order” penalty. So, they actually pitch worse in the 7th or 8th. Three, you have several excellent relievers ready to go, who are a lot better than Gray.

One of the keys here is that Gray is likely not an ace or even nearly so, Yeah, he pitched a good game and had a good season. However, all the forecasters think, at least at this time, that he is a 4th starter or so. It may be defensible to have your ace pitch late into a game if only because if he starts out as 3.00 ERA pitcher, he is probably 3.5 the 3rd or 4th time he faces the lineup, which may not be much worse than a fresh reliever. But a pitcher like Gray? He is a 4.00 ERA pitcher who is 4.5 in the 7th or 8th innings. As opposed to a late reliever who is a 3.00 pitcher or better. That’s a big difference!

Finally, the last inning was interesting.

1 and 3 and no outs in the bottom of the 9th. Leyland IBB’s the lefty batter Reddick to load the bases. Is that correct? Seems a little unusual, although not spectacularly so.

We know that the only advantage is that it creates is a force at home rather than a tag play.  Plus, you also get the home to first DP. However, the infield has to play up and come home on a ground ball anyway. So how often would you get a ground ball that you cannot get the runner out at home on a tag but you do get him on a force, and how often do you get the home to first DP where you would only get one out at home?

We know the downside. A walk now wins the game. Then again, if you walk the batter with 1 and 3, you probably lose the game anyway. So, really the walk off walk is not as bad as you think. If the batter walks with the bases loaded, you win 100% of the time. If he walks with first and third, you still lose 93% of the time. So you gain 7% when the walk occurs. That happens around 10% of the time, so the walk only gains .7%!

I don’t know the answer, but it seems like it might be close. It would take some time to figure it out, if it is even possible. Now whether bringing in Porcello was correct is another story which I will not go into (if they bring in a lefty, Melvin pinch hits for Vogt, right?)…

  1. Derek Florko says:

    Why would you ever Pinch it for Vogt the guy is a DUDE!

    Very excited for your site, Being a basesball coach/former player my favourite part of sabermetrics is the in game theory looking forward to reading your work!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s