Comparing Like Things

While on vacation, I was sitting in a hotel lounge, with some sports show in the background playing on mute. From the overly animated expressions and the screens of stats and the chron banner though, it was pretty easy to guess that they were riled up about some game the day before, someone’s outstanding performance—or lack thereof—and debating how it stacks up in the grand pantheon of that sport’s greatest moments/players/teams/franchises. The entire format is a reflection and systemic expansion of our obsession with ranking everything, repeated ad nauseam for hotel lounges and gyms across the country.

What kind makes it work—enough, anyway, to fill up a dozen hours, every day, every week—is that there is an effectively infinite number of dimensions of comparison, and modern sports now generate enough statistics quantifying on-field play that the movement is celebrated by an annual conference. In fact, one surefire way to create debate is to espouse an opinion about whether sports analytics actually works; reliably, the data-driven and analytics advocates will math out their chosen sports’ equivalent of Moneyball, while old-school players and fans will argue for the intangibles and the immeasurability of the human condition. It makes for infinite content.

But even when everyone agrees on using numbers as the basis for measurement and comparison, the tricky part is then figuring out an amicable method to map multiple dimensions onto a singular one. That is, what we learned in kindergarten—that 10 > 4—works for all sorts of complicated things in real life; the question is then what we count and how much we count it. The simplest technique to intuitively apply here is just adding up all the criteria, weighted by how important each one is: academically, this is called the weighted sum model, and I’ve even written about using this in the context of evaluating job offers and opportunities.

With sports talk though, there is absolutely no consensus on what these weights should be. Does Jordan’s 6 NBA championships overwhelmingly weigh his case for being the greatest player in basketball, or Lebron’s 8 straight trips to the NBA Finals? Are the 49ers more impressive in the 80s and 90s or the Patriots’ sustained excellence over 2 decades? Even countries get in on this during the Olympics: is it more impressive to win gold at most events, or the most medals overall? Have people considered weights? The linguistic and psychological ambiguity in defining what is “best” appears to be a feature, not a bug.

Taking a step back from silly sports discussions, rankings and quantifications still show up in all sorts of places. Money, for instance, is an economic measurement of value, and a commonplace way to map a single number to most things in the world—even things that don’t feel like they should have a dollar sign attached to them1. Investors determine where they’ll put their money to work, again, based on this abstract representation of value that can translate to a business of selling 0s and 1s (big tech) to greasing the wheels of commerce (money market accounts) to providing semi-permanent shelter (real estate). Transmuting things into a standardized numbering system—currencies—has unlocked all sorts of innovation2.

In my world of software development, I often ask to “compare like things” when confronting gnarly prioritization problems. How do you directly compare the value of a new product feature, versus an exploratory integration that could boost sales, versus fixing the technical debt that played a role in a site outage a month ago? Do we optimize for retention, engagement, revenue, efficiency, or throughput? To state the counterfactual: if there was a consistently correct answer to these questions, we would have all read it in a business book somewhere and be following their guidelines as best practice already.

So unfortunately, although we have an instinct to quantify and compare entities, knowing that fact doesn’t help when the way to get to those numbers is a subjective and oftentimes arbitrary process. Decision-making frameworks, as far as they help provide structure, do so by minimizing and concentrating on where to inject human judgment. And perhaps that is entirely apropos; they are in service of our compulsion to compare, and rank, and coronate.