I have been enjoying baseball a lot recently. My wife and I watched most of the 2020 post-season, especially the AL and NL CS and the World Series. We even bought two gloves and a ball (used equipment, of course, to save money) and have been playing catch during the day. It’s quite enjoyable and good exercise – and my wife can now catch, field, and throw with accuracy and power, which is pretty exciting.
But I got into baseball because of data analytics.
Business analytics (as a field of research), certainly has a lot to teach, but it still uses so many biases. Meaning, business analytics still privileges things that are not useful for understanding a business’s performance, AND there are increasing – often politically motivated – demands on business to measure non-revenue generating or cost decreasing activities.
The objective theory of business analytics should be the same as baseball: to measure the performance of a company/employee as it directly relates to profit – or – to measure the performance of a team/player as it directly relates to wins.
But business is still very crowded with biased thinking and preferential, Moneyball-comedy-esque, ways to measure staff or company performance.
“She looks like a true professional.” “He shows up on time and leaves 10 minutes after everyone.” “Always has a great attitude.”
If you’ve seen the movie Moneyball, you’re probably picturing the scene with all the old school scouts talking about the attitude and appearance of players. Noting their “pretty swing” and even claiming that the attractiveness of their girlfriend says something about their confidence as a ball player. These same types of arguments happen in business. This employee is good because of some reason that has nothing to do with profit; this month was better than that month because of some personal feeling about it… etc. What’s worse is when people use some analytical claim but round all the numbers up and include indirect data points that conform to their opinion, while rounding down and excluding indirect data points from the model that they don’t prefer.
Be honest with yourself here (if you’re a business owner or executive especially). How often have you done that? Not trying to be overly critical. I get it. We all do it. We have a plan and we’d really like it to turn out good. So we look for numbers to suit that plan we had. But it’s not helpful for your business. Paradoxically, owning your “wrongness” leads to business success. There are a lot of people who are right all the time and not making much money.
So I went to baseball. Baseball gets it. Everyone is using analytics. Teams use them, players use them, commentators use them, and fans use them. And the models are very sophisticated.
As a side note, I’m reminded of a statement from Noam Chomsky about the political intelligence of the average American. It is often argued that Americans (or any population, really) is politically dumb. “Low information voters” and the like. Noam Chomsky disagrees, using sports as his defense. “Listen to any morning sports radio show and you’ll see that the average American understands very complex problems and can conduct sophisticated analysis.” I’d give you the same challenge if you doubt it. Baseball analytics proves this as well.
The most popular of the modern breeds – SABRmetrics – is well understood by the average baseball fan. The models are complex but the meaning is what really matters here. And the objective philosophy of SABRmetrics is fucking obvious. Is the performance of X player contributing to more wins, or not!
This is the “WAR” stat, called Wins Above Replacement. WAR uses a reasonably complex calculation to measure the productivity of a player’s hits, total bases, outs, times caught stealing, defensive productivity, errors, etc… To produce a number that says one thing: “Here is how many more (or less) wins the team had because they had THIS player rather than [Insert Statistically Average Player].
The biggest stat for measuring the quality of a player (particularly for offensive production) was batting average (BA), home runs (HR), and runs batted in (RBI). But, of course, this classic model missed so much and wasn’t ultimately calculating the stat that mattered – how many more wins is a team getting because of the overall performance of the player over the course of 162 games.
Another great stat is for the pitchers: WHIP and FIP.
WHIP = walks and hits per innings pitched
FIP = fielding independent pitching. Which I think is a poor sentence crafted so we could say “FIP” instead of “PIF” because the meaning REALLY is “Pitching effectiveness that is Independent of the Fielding.” But nobody wants to say PIF… So, yea say FIP but understand the sentence in reverse.
The old guard for understanding pitching was ERA (earned run average). ERA is still a nice calculation, as the runs you give up as a ratio of time spent on the field matters, but it misses some critical measurements of pitcher quality. WHIP tells you how many people this pitcher is letting get on base. Runs come from people getting on base. So, oftentimes you’ll see variance between ERA and WHIP. If your WHIP is a lot higher than league average but your ERA isn’t, then it is understood that you’re pitching much worse than your ERA is showing, and we can expect your ERA to rise very soon.
This variance is also where FIP comes in. FIP is similar to the WAR stat, in that it measures performance using a league wide average. The unique element in FIP is that it only takes into consideration the amount of base runners / hits given up by the pitcher that HAVE NOTHING TO DO with fielding (e.g. the defense behind the pitcher). Thus, the home runs, walks, players hit by pitch, and strikeouts. These are the results that can occur where the rest of the defense could be theoretically taking a nap and it wouldn’t change anything. The reason for the calculation is that good fielding vs bad fielding influences a pitcher’s ERA. You still want to know that calculation, but you’d like to compare it to a more “raw” metric of the pitcher by themselves.
There are many more measures of performance, but WAR, WHIP, and FIP are some good examples.
The point in all this is that we often fail to measure our businesses and employees with such rigor and philosophical discipline. We still talk about all these intangibles (which sometimes matter with respect to culture). But at the end of the day, what does your appearance, organized desk, or curated social media stream, have to do with your results? Very little.