Image credit: Orlando Ramirez – USA TODAY Sports
Translated by Fernando Battaglini
SILENCE!
We need to know what the computer wants us to do. After all, baseball teams are run by supercomputers. There is no longer any human element in the game. It’s all based on what the spreadsheet says. At least, that’s how the narrative goes. If only things were that simple.
There is no doubt that baseball strategies have changed over the last century and more. The game of 2023 is played differently than the game of 2003, 1983, 1963, etc. The last 20 years, and particularly the last 10, have seen dramatic increases in strikeout rates. The touch has practically disappeared. Starting pitchers now go five innings and call it a day, and no one thinks that’s strange anymore. And you’re welcome to any opinion you want about how the new ball game is played, but there’s a more sinister accusation that accompanies it. The baseball analytics movement has been the most visible force of the last 20 years and there have been cries that they have removed the “human element” from the game. Managers are now supposedly engrossed with their spreadsheets. There is no room for “hunches” or for managers who have their own personalities and strategic quirks. Everything has become so homogenized that no matter what managers do, they’re not really running the game anymore.
Spreadsheets become an easy villain. The theory of the case is that, given the same information, 30 teams will effectively perform the same 30 analyzes that will return the same result 30 times. And if the managers have been told that they will follow the orders of the spreadsheet… or else, everyone will do the same and individuality will be erased from the game.
Interestingly, that is a testable hypothesis. So let’s try it.
Careful! There are explicit mathematical details!
If it is mean, then standard deviation. I used to say that when I taught statistics. For example, we can look up the percentage of time a team had a runner on first and second base free and their runner attempted to steal second base and graph it over the years (1950-2022).
We see that there was a tremendous increase in the 1970s, a 20-year plateau, and then the rate has been gradually falling until this year (not shown in the graph), when they changed the rules to make the incentives around robbery attempts are higher. runner friendly.
But what was the standard deviation between the teams? Were there some who showed the robbery sign all the time and others who refused or were they all involved in the action? A low standard deviation would mean much more uniformity between teams. A higher value would mean that teams would be much more individual and free-thinking. Good…
What happened was that in the 1970s and 1980s, some teams (mostly concentrated in the National League, as there really was a power/speed divide between the leagues for a while, whether by design or chance) They really got into stolen bases. The rest don’t. It drove up the overall average, but eventually both the average and standard deviation fell and then stayed fairly constant…starting in the 1990s, which, by any definition, predates the age of spreadsheets. In the 1990s, we know that teams became more interested in having bigger, often less mobile, but more powerful position players. And before the 1970s, we see that the standard deviations were low, meaning that everyone was following the same basic strategic guide. It seems like the 70s were strange.
But let’s move on. Let’s look at another area where we know there has been a change: touching the ball. We know that the bunt percentage graph (I’m using the percentage of time a team had a runner on first, with no outs and without a pitcher at bat) has decreased significantly. Most people who read Baseball Prospectus will know why (more runs are scored with a runner on first and no outs than with a runner on second and one out), but how much has the league moved at the same pace?
We see that there was greater variability in how managers/teams approached touch until the early 2000s, with a notable increase in compliance over the last 10 to 15 years. And given how hard baseball analytics have harped on that, it seems reasonable to believe that managers have been politely asked to refrain from bunting. But. Is this the tyranny of the spreadsheet?
Well, let’s keep moving. What about aggressive baserunning? If the newspapers are to be believed, some managers like a more aggressive “get ’em” strategy and others prefer to play it safe. I looked up all the cases in which a runner could have taken an extra base on a hit (first to third on a single, first to home on a double, second to home on a single) and looked up how often each team sent the runner . Over time, the “shipping” rate has decreased, leading to more play from station to station. But…
Basically, there has been the same variation between teams for 70 years. If we’re going to say, at least when it comes to pushing baserunners, that that strategy is the same today, well, that similarity has been pretty consistent for a long time.
Here’s the most important one: the time a starter pitches in a game. The average has fallen precipitously in the last decade. But what about the standard deviation? Is the spreadsheet telling managers to pull the pitcher?
This is the game-level standard deviation (everything above has been at the team season level) of how variable the definition of “start” has been. Surprised?
The definition of what is used as “an opening” used to be much more negotiable in the old days, and conformity has been creeping into the MLB slowly but steadily over the years. In fact, it wasn’t until the appearance of “The Opener” that anyone rejected conformism.
One more. Another common complaint is that baseball teams are now all boringly the same in the sense that everything is either a home run or a strikeout. Surely that’s being spreadsheet-driven when teams tell their players there’s only one approach to salvation, and it’s to hit it out of the park? Good…
The measure here is the percentage of each team’s plate appearances that ended in a home run or a strikeout. And again, the rate has skyrocketed recently, mostly due to strikeouts. But the numbers tell a different story. Teams used to be much more consistent in how much they tolerated the all-or-nothing approach. There is now more room for a diversity of approaches.
Let’s take stock of what we have seen. There is no clear pattern that the last few years (the spreadsheet era) have been the era of enforced conformity. In some cases, compliance has been reduced. There is no doubt that, whether the spreadsheet was the cause of the changes or simply the scribe who documented what was already destined to happen, the spreadsheet era has spawned a new strategic ball game and such maybe even aesthetically. But looking at the data, we can see that over time, the conformity/variability rates have been fairly stable, with one or two notable exceptions.
What if…MLB teams had consistently tried to follow what were considered (and perhaps were, given the circumstances) best practices at the time? And by having reasonably intelligent staff, everyone figured out roughly what that was and accepted it. It would be something totally reasonable. It’s not that anyone enforced conformity. It’s just… well, that’s the best way to win a game.
What if… the spreadsheet was just another tool to figure out what those best practices were? It’s not the tyranny of the spreadsheet. It’s the teams and the managers who say, “I want to win this game” and figure out the best way to do it. As they always have.
The spreadsheet says I’m right.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now
2023-11-02 10:55:20
#tyrannical #spreadsheet
