Wednesday, April 13, 2011

Playing the Odds



Last night, the Twins defeated the Kansas City Royals 4-3 in 10 innings. During the 9th inning, the Twins faced what TV announcer Dick Bremer called a "difficult decision." The game was tied at 3 in the bottom of the 9th, and Danny Valencia had led the inning off with a walk. Obviously, the decision that Bremer was referring to was whether or not the team should bunt Luke Hughes, because everyone knew Jim Thome was going to pinch hit for Alexi Casilla, who was on deck. The assumption of course was if the Twins did successfully bunt Valencia to second base, Thome would be intentionally walked, and the top of the order would be coming up with one out and runners on first and second. The tough decision was if the team was willing to let KC take the bat out of Thome's hands by walking him. My initital assumption was that bunting was the wrong decision, and Gardenhire did things correctly.

Ultimately, Gardenhire decided to let Luke Hughes swing away. I've mentioned here countless times that I don't like sacrifice bunting, because giving up an out to move a runner one base is almost always a poor decision, unless that base is third to home. Statistically, teams score more often with a runner on first and no outs than with a runner on second and 1 out, so it was the right decision. But not because it allowed Thome to hit; an intentional walk is statistically more beneficial for the offensive team than the defensive team.*

*That seems obvious, considering you are giving an opponent a free base, but despite intentional walks usually hurting a team more than they don't, managers continue to use them to try to get out of dire situations. An intentional walk is no different than a hail mary in football, but most managers either don't realize this or refuse to admit it.

While Gardenhire's decision to let Hughes swing away likely made NO DIFFERENCE on what happened during the inning (Thome singled to center and Span came up with 1 out and runners on first and second, anyways) what got me thinking was if bunting in that situation may have been the right idea, considering a successful sacrifice would have obviously guaranteed runners on first and second with just 1 out. I wondered if the fact that the bunt would almost certainly be followed by an intentional walk might actually make bunting the right choice--basically, is giving up an out worth ADDING a baserunner?

So I looked it up. The answer, based on thousands upon thousands of stats, is that a team will score 0.95 runs per inning when they have a man on first base with no outs. They will score 0.72 runs with a runner on second and 1 out. But with runners on first and second and one out, a team will score an average of 0.98 runs per inning. So, while sacrificing a runner from first to second moves the teams chances of scoring down nearly 20%, bunting Hughes with the assumed intentional walk to Thome would have actually increased the team's chance of scoring, although only by about 3%.

So, ultimately, based on statistics, Gardy made the wrong decision. The team won anyways, and the difference between 0.95 runs and 0.98 runs is incredibly small, even over 162 games. This does not mean "statistics are useless" as I'm sure at least some of you reading this are shouting at your computers (or at least inside your head) but let me explain myself.

I like to compare the baseball season to playing blackjack. When you play blackjack, the house is always at an advantage. This is because the casinos are trying to make money, so of course the odds will be in their favor. Unless you're counting cards, you will always be a statistical underdog against the dealer; that's just the way the game is set up. However, there are specific things players should do in specific situations that can increase their odds. There are too many to name, but a common mistake made by blackjack players is what they do if they have a 12 and the dealer is showing a 2. Without knowing the count or anything else, the average player will have a positive result about 50% of the time if they hit their 12 against a dealer's 2. The other 50% of the time they will have a negative result. Anyone who understands blackjack well will tell you, quite honestly, what you do with a 12 against a 2 doesn't matter--all that matters is what you do one time is what you do every time. So if you hit your 12 against the 2 and bust, don't let it stop you from hitting it again. As long as you always do the same thing, over time it will always even out.

So, why not try to guess every time and do better than 50%? Because unless you're somehow 'randomly' guessing each choice evenly, you're likely going to lose more often than you would sticking to one strategy. For example, say you hit 70% of the time and stay 30% of the time. If it's 50-50, as it will be over time, you're going to lose 20% more of the time then you should. It really is that simple; unfortunately, so many people don't understand this part of blackjack and that is why you see so many upset gamblers storming away from tables when you sit down and have no idea what you're doing.

This is no different in baseball. It is a 162-game season. Unlike in the NFL where a hot five game stretch can carry someone's stats for an entire season, we know that over 162 games baseball players stats are likely going to come in close to what we've seen in previous years. Joe Mauer is hitting .233/.303/.267 right now, but I would feel safe guessing that by the end of the season he's very close to .325/.405/.480. That's based off of more than five seasons worth of stats that suggest Joe Mauer is a far better player than his early start this year.

So while a simple decision that Gardenhire made in the 9th inning last night ultimately had no effect on the result of the game, over the course of a full season, if Gardenhire continues to make the wrong decision it will eventually add up. I said that the difference between 0.95 runs and 0.98 runs per situation makes little difference, even over a 162 game season, which is true. However, choosing to bunt Alexi Casilla later in the year to move a runner from first to second, without an intentional walk happening behind it, is going to cost the team about 1/4 of a run each time he does it. If the team sacrifice bunts 100 times this season, they will ultimately be giving up 25 runs for the year based on the statistics that have been calculated over the last hundred years.

If a manager is going to be making these kinds of decisions on a daily basis, much like a professional blackjack player, shouldn't the manager know all the odds of each situation? If bunting and an intentional walk will improve the team's chances of scoring, shouldn't a manager do it every time? Sure, it's not going to work out every time. There will be times Gardenhire will bunt Casilla, and Span or Mauer will rip a base hit to score the runner and the Twins will take the lead or tie the game or win the game. That's the equivalent of staying on 16 when the dealer is showing a face card; statistically you have a much better chance to win if you hit, even though you will likely bust. Over time, hitting on 16 against a face card will do you more good than standing on 16 will do. That's how I think of the sacrifice bunt. Sometimes it will work out as planned and the manager will be praised and the commentators will talk about how the Twins just "do all the little things." But what won't be understood is that the manager actually hurt the team by sacrificing--unless of course he knew an intentional walk would follow.

The great thing about statistics is that they don't lie and they don't play favorites. The numbers are what they are; people can try to discredit them and argue with them, but the stats aren't biased. They aren't emotionally attached to Nick Punto or Matt Tolbert the way Ron Gardenhire seems to be, and that's why the stats suggest Punto is nothing more than a utility player and Matt Tolbert should be a AAA utlity player, for example.

So next time, just bunt Hughes. It made no difference tonight, but it might in the future, and as we all know, every single game matters. I don't think my heart can handle another game 163, so let's just give the team the best chance to win in every scenario. Gardy, if you need a copy of "The Book" by Tom Tango... I'll gladly lend you mine. But you'll need to buy a kindle first. And if you ever sit down at my blackjack table, I will get up, immediately. Although I'm sure I'd stick around to watch.