Scoring Patterns and Winning Percentage

First and foremost, I have some good news: I will be the video intern for the Everett Aquasox this summer! Thank you to all who have been so supportive over the years. I’m excited for this opportunity, and I’m very grateful for the encouragement I’ve had along the way. I wrote most of this post before I got the job, so it’s still written from a very Rangers-centric perspective, but I will start incorporating Mariners material very soon. Anyway, on to baseball things.

Deciding when to lay down a sacrifice bunt is typically very context-dependent. If the context is “Ron Washington is my manager,” then the answer is usually yes, sacrifice as much as possible. Sabermetrics people will tell you that one of the most scarce resources in a baseball game is outs, and to give one away for free is usually a bad idea unless you have a really good reason. That reason is usually when a) there are no outs, b) there are runners on first and second, c) the bunter is terrible, d) the guy(s) behind the bunter are awesome, AND e) a single run will almost certainly win the game.

Ron Washington has a much looser definition of “really good reason”. He has called for a sacrifice with the Rangers down by 4. He’s called for a sacrifice with the Rangers up by 7. He’s called for a sacrifice in every inning from the 1st to the 14th. He likes sacrificing in a box, he likes sacrificing with a fox. He likes sacrificing in the rain, he likes sacrificing on a train. Have I made my point? Ron Washington calls for sac bunts too much, and it’s annoying.

Recently, Washington called for Elvis Andrus to lay down a bunt in the top of the first against Felix Doubront and the Red Sox with no outs and a runner on first. This is one of the WORST times to call for a sacrifice bunt, because Elvis is good, the baserunner situation is less than ideal, and a single run in the top of the first will almost certainly NOT hold up against a team like Boston. The sacrifice was successful, but the run failed to score and, more importantly, any chance at a big inning was greatly reduced by giving away a full third of our outs that inning.

As they say, if you play for a single run, that’s the most you’re going to get. Thinking about this little quip made me think of the saying that pitchers should always try to avoid “crooked numbers”.* The idea is, you’re inevitably going to get scored on sometimes, but if you can limit the damage, the game will still be within reach. Both of these sayings seem to indicate that teams should try to score in bunches, and try not to let their opponents do the same. Groundbreaking stuff, I know, but I wanted to know just how true these truisms were.

I wanted to know more about scoring patterns and how they related to win probabilities, but mostly I wanted to prove once and for all that Washington’s strategy of taking a single run as often as possible was sub-optimal. To do this, I looked at the box scores from every game from 2013, and tallied the W-L record of the team that scored in the most innings, the team that had the single biggest inning, and the team that recorded the most innings with crooked numbers. For example, if a team won 5-4 by scoring 5 in one inning while their opponent scored 1 in each of four different innings, the “Most Innings” team would get a loss, the “Biggest Inning” team would get a win, and the “Most Crooked Numbers” team would get a win.

Before we start, I will offer up the caveat that this data will not be particularly useful for in-game strategy decisions, because the context of the specific game will almost always outweigh the overarching trend. This should be filed more under the category of “thought experiment” than “game-revolutionizer”. Okay. Without further ado…

As you would expect, the winning percentage in all three categories was quite high since, in a vacuum, winning any of those categories is decidedly better than losing it. There were also quite a few ties, but I just threw them out, because I think the wins and losses tell a complete story already, and it would only get muddled by trying to figure out the fairest way to allocate the ties. I only even bring it up because it explains the difference in the number of games in each category.

Surprisingly (to me at least), Wash’s strategy of playing to score at least one in as many innings as possible, at the expense of some big innings, seems to be more successful than the strategy of forgoing a single run to preserve a better shot at scoring four or five at once. However, the MOST effective strategy is to go for “Most Crooked Numbers”, which makes sense if you think about it. It’s really just a combination of the other two: you score often and in bunches, and that’s tough to beat.

It turns out it is good to score as often as possible, but it’s best if you can push across more than one at a time. So if Wash can at least MODIFY his bunting strategy to only do it in situations with multiple runners on base, then he’ll be playing to score two every inning, and he can shift from being a “Most Innings” guy to a “Most Crooked Numbers” guy. And I know, it’s only a 1.4% difference, but if you can be 1.4% better, you should do it. If the Rangers had been 1.4% better last year, they would have won two more games. That’s the difference between hosting the wild card game and hosting the play-in game for the right to visit Cleveland for the wild card game.

Baseball is a long season, and little things add up. The best teams are usually the ones who make the most of the opportunities they get.

---

 

 

*Because the number 1 is a straight line. And 2, 3, 4, 5, etc. aren’t. Get it?**

**I want to meet the pitcher that gives up 11 in an inning and then brags about avoiding crooked numbers