Behind the Numbers: Do Blowouts Predict Champions?

Posted by KCarpenter on November 24th, 2010

Kellen Carpenter is an RTC contributor.

It’s been a hard-fought game. Every shot contested, players from both teams diving all over the court for loose balls, the crowd whipped into a frenzy. Bitter rivals, ranked teams, and it all comes down to one buzzer-beating shot. This is a good game. This is why we watch college basketball. Who wins? Who cares? Rhetorical questions can cut two ways, and this one is particularly vicious in matching it’s semantic zig with meaningless zag. On one hand, we gladly affirm the joy-of-the-game sentiment: Close games are their own reward, and if both teams played hard, we can walk away with a zen-like comfort at having seen a beautiful competitive spectacle. That said, there is a colder reading, one that isn’t as much fun, but may be just as important: Close games between evenly-matched teams don’t mean much of anything, it doesn’t tell us anything about how good either team is. No matter what is marked down under the team records, win or lose, the game is basically a tie.

Teams That Win Close Are Nice, But Teams that Dominate Are Better

Maybe you are preparing a protest: pulling out a tough win shows force of will, late-game execution, grit and guts. These are the best wins, you argue, because it shows you that the team knows how to win when it really matters. A team that can win when the going gets tough can win no matter the situation. Well, okay, hypothetical protester, time to drop the biggest open secret that everyone not on television knows: margin matters and, in terms of predicting future success, a win by one means about the same as a loss by one.

This makes sense with a little bit of thought. Basketball games are a long series of events each contingent on probability, and probability is a fickle mistress. Consider our hypothetical one-point game: In the second quarter, the power forward for one of the teams manages to take a barely-contested ten foot jumper and he has a 50% chance of making the shot. Half the time it goes in and half the time it doesn’t, the rest of the game unfolds in the same way and because of that one event, the team either loses or wins. In every other way, the team is just as good or just as bad as we always knew it was. One shot doesn’t actually change that. Did Butler lose the National Championship because they missed the final shot? Maybe, but they also missed eleven other three-pointers over the course of the game. Hitting any one of those would have won the game.

Is this an oversimplification? Yes, but it all speaks to a simple idea: That the difference between a one point win and a one point loss is really almost white noise. The idea isn’t just a philosophical buzz-kill, but rather a philosophical buzz-kill with implications: Winning by a little doesn’t mean much, but totally dismantling an opponent means a lot. If a team wins by twenty points, it means that they are very likely better than that opponent. Instead of just one lucky shot, they had ten lucky shots, and if you have that many lucky shots, then chances you aren’t so much lucky as good. There is meaning in blowouts. This is why just about every serious advanced-statistical ranking system takes into account margin of victory. That said, we don’t need to tear wide open Ken Pomeroy’s magical tempo-free machine to fully understand this principle, and instead, for an illustration we can turn to a different sort of numbers game.

Read the rest of this entry »

Share this story