Math led me to sports and almost ruined my life

Long read ~2000 words

It's not an uncommon story about the nerd who can't play sports, so the nerd plays with sports analytics. Charles Barkley even said this -- alongside saying we are unattractive geeks who could never play sports.


The list goes on: fantasy sports are just Dungeons & Dragons for jocks, analytics take the fun out of sports, no one cares about advanced statistics other than the basement dwellers who have Cheeto fingers...

I don't take offense. It's pretty funny at some points -- at other points, entirely just clueless. But this time of the year, Opening Day for MLB, I am always reminded of how I got into a sport I literally hated. And ultimately, how that changed my life.

I'd like to share that with whoever is reading right now. This post is mostly for me to have my thoughts concrete -- I have thought about every year for the last 5 years -- so this is not very informational beyond knowing more about me. I'm not sure how unique my story is, but maybe if you read this then you may see beyond the Cheeto finger, neck beard, basement dweller and maybe pick up interest in a sport you once hated.

It's not surprising that the intersection of sports statistics and sports fandom meets at fantasy sports. Fantasy sports are plainly driven by statistics -- generally scores attributed to an individual player. To excel at fantasy sports, you must begin to follow at-large the entire sport you are playing in -- that's why you generally hear people who play fantasy sports knowing almost every key player in the sport. You start figuring out who has the best stats, albeit small numbers. You start maybe looking up weekly projections. And maybe you add a podcast on fantasy sports to your weekly listens. All of these things are stats-driven, but maybe not many competitive fantasy sports players see it as "statistics".

There I am, playing my first season of fantasy football. I draft like an idiot, but so do the rest of us -- we all just started playing for the first time. Since we are all horrible (and, as the season progresses, half of us are actively playing and the other half stop playing after a week or two) I end up doing well in the fantasy league. But I wanted to optimize my performance. That's always been my personality -- I can do it better, I just need to know how. So, as many before me have done, I looked into what earns the most points in our league. But how do I maximize getting the best players on my team when everyone has the same chance of selecting players to draft? That's where statistics come in.

So I looked at projections, but realized most projections out there were based simply on one (maybe two) years of performance before. The projections were weakly correlated for each player. And the stats themselves only took into account what an athlete would do over the course of a season -- not week-to-week, which is how the fantasy match ups were played in our league. This seemed ripe for analytical conquest.

So I very carefully found datasets for all the athletes I wanted to draft and noticed their performances were very match-up biased. In other words, these players would slaughter mediocre or poor teams but wouldn't do very well against above-average teams. This seemed to be exploitable -- good players playing against good teams would perhaps equal less fantasy points than a mediocre player playing against poor teams. So I looked at the best performers each week who did not start the game (theoretically, the players who start the game are the best on the team). There were a good amount of available players to select without having to waste drafting them -- meaning my proposed strategy of slotting in a mediocre player for a good player may hold water.

It did. That season I won quite easily. Again, this was against people who weren't the most competitive -- they competed, but I competed differently. So I entered a more difficult league the next year -- more competitive (included a money buy-in), and sure enough I did well (I didn't win, but I made the playoffs and stuck to my strategy).

Three years of playing fantasy football led me to even more analytics. Injury projections. Defensive projections. Match up ratings for a quarterback playing against an elite strong side rush versus an elite weak side rush. Teams who perform poorly against swing plays. Exploitable goal line defenses. Lots of analytics that I don't even care to use today but can recall like my birthdate. I didn't know I was doing statistics -- in fact, at that time in college, I believe I had earned an A- and then a B+ in our two-semester-long statistics course. Outside of simple averages and standard deviations, the statistics I was computing were different from the statistics I learned. But I didn't call them statistics. I didn't call them anything, they were my secret weapon.

At this point, I had moved out of my parent's home for the second time and transferred to a school about two hours away. A friend invited me into their fantasy baseball league. I thought sure, why not. It's all numbers. But secretly inside, I hated baseball. The game was long and boring. There wasn't really a team aspect that I could see -- one guy hits the ball, one other guy catches it. Sometimes they throw it around. Whoop-di-doo.

But fantasy sports ignores how long a game takes. Fantasy sports are simply numbers versus numbers. I didn't have to know anything about baseball to do exactly what I did in fantasy football -- I figured out what was valued the most, what players did poorly against certain match ups, and what players were undervalued due to non-prime-time match ups.

Again, I hated baseball. But I also didn't know anything about baseball. I didn't know what a strikeout was. I didn't know how many steals there were usually in a game. I didn't know that 10K's were elite and 3K's were more likely. I didn't even know what a K was, let alone a BB, or even ERA. I didn't know anything about it.

But that didn't matter. These random letters equalled points. So I went with my analysis, zero "gut instinct" based on players names, and really good headphones so I could ignore the roars of laughter when I drafted two pitchers in my first two rounds.

No matter. These so-called really competitive fantasy players got trounced by my lack of knowledge and analytics. I didn't win the championship, but I came in third. And didn't know anything.

The next season, I investigated more into the advanced statistics of baseball. How wRC+ was more accurate than ERA, what FIP and xFIP truly meant -- and how all of these were related to fantasy points. Again, I started to have a vague idea of what a strikeout was, or a hit, or all of these things, but I didn't really know what it looked like. They were just words attached to numbers. Abstract in all meanings of the word.

I sometimes would pretend to know what something meant when talking advanced stats to someone. Tim Lincecum's WHIP is Cy Young material this year! Oh, but his BABIP isn't very consistent. ...You must be one of those nerd guys.

I was. I still am.

Two years of fantasy baseball led me to finally go to a baseball game -- the first one since I started playing fantasy baseball. It was the Giants versus someone. Lincecum was pitching and he throws 9 K's. I find out that day how hard a strikeout was to get. Pablo Sandoval hits a home run. I find out how different parks are and why wRC+ is so important to a team like the Giants as opposed to a team like the Red Sox. All of those acronyms and things that equated to fantasy points actually had meaning.

A silly question I asked: why do they show how many were left on base but not WHIP? Oh right, no one around me cares about WHIP -- LOB is a raw metric to show how many potential runs could have been generated (like moral victories or something).

But as the game progressed, I realized the stats were alive. And not just in baseball, but all around me. I started looking at things in terms of statistics and probabilities. What's the likelihood I will need to wait for a taxi on this block versus the next block? What's the cost/benefit of taking the train versus a bus? What would a heat map for places where people stand in a small elevator versus a large elevator look like over the course of a day?

I tried applying stats to my own life. I kept a dream journal to see how many times I recalled dreams. I measured commuting in terms of time and gas and money spent. I had an ongoing count of how many times I tried to stick a USB plug into the port versus how many times I had to rotate it before it went in.

I started assessing my life in more toxic ways. How do I optimize seeing friends so I don't seem like an asshole? What's the latest I can arrive to a party without being a jerk? How late can I make plans with certain people without them saying no?

Stats, for a long time, ruled my life. I've ruined many relationships with statistics -- either because of my obsession over stats or because I statistically computed them out (which, as I re-read this during my edits, I sound like a freaking horrible person...).

I've regressed back to a less-stats-driven-life person now. It's been two years since I've played fantasy sports of any kind -- aside from March Madness brackets -- and I stopped keeping a dream journal and stopped keeping a black book on when I saw my friends and for how long. I took a step back and realized statistics have got me a lot of things, but it can't get me everything. And that's okay -- it shouldn't.

Before, in my early days of basketball statistics, I really tried to compute "one number to rule them all". It included both offense and defense, individual performance and team performance, and was weighted for match ups. It was a horrible predictor for wins -- I think it predicted maybe 12% of games? -- and I quickly abandoned the idea of one number to rule them all. There isn't one number. Just like there isn't one thing in my life that can make sense of everything around me.

I found sports through statistics. Through sports, I found a lot about myself. I found out that even though I am married to statistical analysis, there is a lot more than statistics. Maybe my life at the time was spinning out of control. I didn't have any particular calling and everyone around me was on a path to success and independence. Statistics was something I could do on my own without help. I tried to stats everything. And then I just ruined myself. I went into a field, cognitive neuroscience, that has married both statistics and psychological behavior -- where it's difficult to rely on one or the other, and the best results usually come when both are compatible with each other.

Just like when I saw baseball speak to me for the first time, stats and behavior of the game made too much sense. I sought that high and didn't find it when I was trying to create it for myself. I've learned to let those moments come to me more naturally. Because of this, they are seemingly coming more frequently in terms of opportunities and benefits for pursuing statistics and human behaviors (sports, neuroscience, etc). It's all seemingly due to baseball. Baseball is a frustrating, annoying sport. There doesn't really seem to be a team mechanism when you're just casually viewing. But when you change your lens a little, everything starts to make a lot more sense -- for better or for worse.

One thought on “Math led me to sports and almost ruined my life”

  1. Dr Winters says:

    fantasy teams that drafted brandon lafell made the play-offs (p<0.001)

Leave a Reply