On Luck

[This is part two of what is probably too meandering to be called a series, but anyways, here's part one.]

We have a way of diminishing the role of luck in sports. Not in any grand, obvious way -- when it comes to broad topics like injury or draft fortune, the great unknown is regularly evoked -- but a more subtle one. When all else is equal, success and failure is chalked up to effort and skill and talent. When a team wins, when a player loses, the presumed inputs are matched to the observed outputs. Generally, if we are confident that the players are (a) healthy and (b) not breaking any rules of the game, we feel safe mapping these things (effort/talent and success) one to one.

The observed outputs are the tangible things: the wins, the losses, the division titles, the lottery trips, the championships, everything in between. They are unequivocal, inarguable Things That Happened.

Identifying the inputs and explaining how the inputs created the outputs: this is sports analysis. Today, it comes in highly varied forms. This is Joe Morgan saying that grit and consitency are essential "inputs" in the creation of the win output soup. This is equally Kevin Pelton explaining that "inputs" of 109 points per 100 possessions on 29.6% usage in the 2012-2013 season create an "output" of 8.0 wins over a replacement player.

I'm not claiming here that one approach is more valuable than the other. For today at least, let's transcend the whole "stats vs..." thing. I point out two radically different approaches more to say that this input -> output translation is fundamentally what all sports analysis is. It's simply the identification of the Initial Things That Happened that eventually caused the ultimate, unequivocal, inarguable Things That Happened.

One problem here is immediately clear. This first strain of happenings, the inputs, lacks entirely the verity of the actual happenings, the outputs. The purpose of statistical analysis as opposed to its predecessors is largely to ground the identification of inputs in something approximating the "reality" of the outputs. "Anthony Davis scored 29 points on 32 isolation possessions at Kentucky" offers more of a universal foothold for both player comparison and reference's sake than "Anthony Davis was pretty good at getting his own shot, but sometimes struggled." The former is just as "true" as saying Kentucky won 32 games (a final output, an ultimate Thing That Happened), where the latter might be an accurate assessment but falls short of being a statement of fact.

The larger point is that, regardless of how we choose to identify and then translate inputs to output, we generally feel comfortable doing this mapping process in some fashion. And with that, we revisit luck.

The fellow in the picture's either flipping a coin or entranced by a levitating coin. Let's go with the first, or this becomes a different post entirely. The coin flip is obviously the simplest case there is. Over an infinite number of flips of an unbiased, non-levitating coin, the heads and tails outputs will be identical. It's a perfectly defined scenario, but even still, as soon as the number of coin flips becomes non-infinite, the output is no longer certain. Reduce it down to, say, one coin flip, and confidently guessing the result becomes ridiculous.

And so the concept of luck permeates basketball writing, analysis, journalism in a very important way: the analogue of the coin flip. Even if we knew with absolute certainty the exact contribution of a specific initial input to the final output tally (which is crazy), uncertainty must accompany every prediction. This is critical because the inarguable Things That Happened were ultimately all single outcomes plucked out of a probability density function. And this is exactly the place where we undervalue luck.

The outcome that actually happened is obviously the one that matters. This feels almost too stupid to type out. But it doesn't negate the fact that the others still very easily could have happened. We do a decent enough job of identifying and expressing uncertainty about the future. Most good basketball writers will have expressed, for example, the idea that the Lakers could range from anywhere from the low 50s to mid 60s in wins.

The issue is that when we concoct and develop our storylines and narratives about the past, we discard the odds of the things that didn't happen happening at all (the things that, and this is critical, are governed purely by luck, the uncertainty of reality, the idea of the perpetual unpredictability of the coin flip). As the sample size is whittled down, the impact of legitimate, unpredictable luck grows, and this is problematic.

Over the course of a full Lakers season, this randomness will largely even out. If the Lakers win 50 games, they'll rightly be chastised for not gelling or for not preventing Kobe Bryant from stealing all their shots or for employing Antawn Jamison. Take it down to an individual series or handful of games though, and when things outside of team and player control do not even out, the media will tend to throw the probabilistic nature of sport, and indeed reality, right out the window. It's the curse of Things That Happened; we endlessly pursue the mirage of certainty no matter how many times it eludes us.

The +6 differential goal is no different. Plenty of teams have hit +6 in the past and lost in the Finals or missed them altogether. Many fell short because of poor matchups or inferior talent. In more than one case though, teams that were good enough to be champions, that won it all in parallel universes where their opponents' threes didn't fall at statistically unlikely rates, where the series of infinitely complex coin flips that comprise the sport of basketball landed slightly differently -- those teams also lost. They fell short in the universe that matters, and as a result, they won't be remembered.

Hemingway's "create your own luck" ideal is one that harmonizes smoothly with the overall sports psyche, but as we assess the past and explain how things came to be, we'd be wise not to overindulge. Life isn't that easy, and neither is basketball.