The uncomfortable truths of analytics: Lessons from baseball and actuarial work

By Chris Fievoli, FCIA

It’s been a couple of years since we tried to tie baseball to the actuarial profession, so I thought it was time to do it again. Only now, instead of focusing on the ethical side, I want to get right to the thing that likely appeals to most actuaries – the numbers.

Baseball has always been interesting because of the sheer amount of data that the game generates. Statistics not only capture what happens in the course of a game or a season, but can also paint a surprisingly accurate picture of how a player performed or even the prevailing trends of an era. As a result, certain figures have become instantly recognizable to fans of the game. If I mention the numbers 61, .406, 56, 755 or 511, someone familiar with the history of baseball will know exactly what I am talking about.

But it wasn’t until the late 1970s that baseball statistics took their next great leap forward. That was when Bill James, a nighttime security guard at a pork and beans cannery in Kansas, began spending his downtime developing a new type of analysis of the game. That eventually led to the widescale publication of The Bill James Baseball Abstract in 1982, and suddenly the baseball world had a whole new way of looking at the sport.

“James took existing baseball statistics and developed a raft of new measures, all of which attempted to provide insights into the game that had never been seen before. It was the first genuine exercise in baseball analytics, which he coined “sabermetrics,” in recognition of the Society for American Baseball Research.”

The next big event in the evolution of baseball analytics was the 2003 book by Michael Lewis, Moneyball: The Art of Winning an Unfair Game, which led to the 2011 film, Moneyball, starring Brad Pitt. It told the story of Oakland A’s general manager Billy Beane, and how he adopted sabermetrics to find talent, which stood in stark contrast to the more traditional style that relied heavily on personal observation and gut feel. By then, just about everyone in the sport saw that analytics provided a way to gain a competitive edge, and there was no looking back.

By the time we got to the 2010s, baseball statistics had an entirely new look and feel. Traditional measures such as batting average and earned run average were supplanted by on-base plus slugging and wins above replacement. Fans had to figure out what was meant by terms such as spin rate, launch angle and exit velocity.

But something else happened along the way. The game, for lack of a better word, started to get boring.

Baseball had evolved into a matchup that was power versus power. Rather than try to manufacture runs through base hits and sacrifices, batters started swinging for the fences. Pitchers, in response, were more apt to try and blow the ball past them. Strikeouts and home runs became the order of the day. Fewer balls were being put into play, and stolen bases became a rarity. As batters started to lose the incentive to hit the other way – as opposed to pulling everything to their power side – teams started to employ defensive shifts… and the games started to drag.

Why did this happen? A lot of the blame was placed on analytics. It was becoming apparent that the traditional hit/steal/sacrifice formula had a relatively lower chance of paying off. It was actually statistically better to play the power game, with the upside of knocking the ball out of the park more than offsetting the downside of multiple strikeouts along the way.

In other words, analytics had uncovered the true formula for success in baseball. And it wasn’t terribly exciting.

So, what does this have to do with actuarial work? What sabermetrics did was reveal an uncomfortable truth about the game of baseball. And actuarial work can often do the same thing.

Take the controversy over the use of postal codes in automobile ratemaking. Residents of a particular geographic area may balk and complain about being charged a higher premium solely because of where they live, but the data provides the evidence – simply put, there are a lot of claims there. You may be a good driver who takes care to keep their vehicle safe, but you may live in an area where car thefts are more frequent, or where residential streets facilitate higher speeds, which may result in more severe injuries in an accident. And you may not like the result, but pricing analysis reiterates these territorial differences over the years.

As we gather more and innovative types of driver data from usage based insurance (i.e., telematics), we are bound to find new categories of factors that contribute to higher insurance costs, much to the chagrin of drivers whose habits directly contribute to that poor experience.

I’m willing to bet that there are several examples of this throughout history. We’ve known for a long time that women live longer than men, but the first time we figured that out, it must have been a blow to male egos everywhere. Similarly, smokers must have been unpleasantly surprised to hear that they had significantly higher mortality rates (especially if they were influenced by cigarette ads featuring Jackie Robinson and Stan Musial).

Undoubtedly anyone declined for life insurance, or saddled with an extra premium, would not be happy about the decision. But, like it or not, it would be based on something in their medical profile, occupation or avocation. Again – just because you don’t like the answer doesn’t mean that it’s not true.

“So, what other uncomfortable truths will we discover, especially now as predictive modelling takes our work to new levels of insight? We don’t know yet, but we should be prepared to defend our findings, and ensure that we have the evidence and rigour to back it up.”

The ability to explain these outcomes in a manner that the public understands will become even more important.

Partially in response to these trends, major league baseball introduced a slew of rule changes in 2023. We now have a ban on defensive shifts, larger bases, limits on pickoff moves, and a pitch clock. (Well, there actually always was a rule governing the time between pitches. It was just never enforced.) It may still be too soon to tell the impact, but it appears that the “old-fashioned” game – hitting for average, stealing bases, making contact – may be making a comeback.

I don’t know whether we would have the same ability to tinker in the actuarial world. We’ll just have to continue to do a good job explaining and justifying those uncomfortable truths that nobody seems to like.

This article reflects the opinion of the author and does not represent an official statement of the CIA.

Follow us

Contact Us

Canadian Institute of Actuaries
360 Albert Street, Suite 1740
Ottawa, Ontario K1R 7X7

Subscribe to our emails