Sunday, June 04, 2017

Picking winners …

… What a 1930s horse racing guide by a conservative philosopher could teach sports scientists today. (Hat tip, Dave Lull.)

The more we know about data in sport, the more the Oakeshott position – confidence in good judgement rather than scientific “proof” – gains strength. Daryl Morey, the general manager of the Houston Rockets basketball team, is a leading exponent of data analytics. Yet he is in the vanguard of a
new, subtler understanding of the relationship between data and judgement. The quality of the mind interpreting the theory is as important as the theory, Morey argues. “You have to figure out what the model is good and bad at, and what humans are good and bad at.” No model can deliver results without a wise, shaping intellect: all theories need to be marshalled by judgement.

3 comments:

  1. Jeff Mauvais12:04 AM

    Sounds about right. In June 1977, I quit my job in LA and devoted the next two months to daily wagering at Del Mar Racetrack just north of San Diego. I did the analytical stuff over breakfast, using data from the Racing Form to pick my favorites, and then headed to the track where all of my focus was on behavior, both equine and human. Horses have moods -- focused and calm, skittish and distracted, or even something that looks like a bad hangover -- so it was important to get a look at my paper favorites as they paraded through the paddock before taking their places at the starting gate. If I didn't like my favorite's mood, I'd sit out the race rather than wager on a horse that I hadn't flagged in the Form. I'd also sit out a race if the odds on my favorite dropped below 3:1 -- anything lower that that meant that the rewards of winning were not worth the risk of losing. So, mostly, I spent a lot of time sitting on my hands. A bit of data analysis, a little horse sense, and the habit of playing against the betting behavior of others: I started with $5000 in July, and walked away with $11000 at the close of the meet in September.

    ReplyDelete
  2. Very interesting. Back in the early '80s I was the guy doing the racing charts for The Inquirer. I noted that if you bet the winningest apprentice jockey to win every race he was in you could earn a pretty handsome wage. Of course, that jockey isn't going to be an apprentice for much longer, but I was surprised at how consistently this observation bore fruit. I still go to the track from time to time, but only for the fun of watching the ponies. I bet, of course. I bring a certain amount of cash that I am prepared to lose. I have never come close to losing all of it.

    ReplyDelete
  3. Jeff Mauvais3:42 PM

    I never noticed that! But since you were the guy actually doing the charts, I'm sure that you came to see patterns that users like me missed. Dare I say that you were an expert?! I always started my review of the Form by looking for horses moving up or down in class - maiden, claiming, stakes, etc. - for the race. It was a piece of information that allowed me to begin constructing a narrative. I'd never bet on a horse for which I couldn't put together a little story. Imagine a horse with good racing fundamentals who'd been on a bit of a losing streak recently. To help jumpstart his performance, the owners have decided to drop him in class for the race at hand, and to put a really good jockey like Bill Shoemaker or Laffit Pincay in the saddle. I'd jump on that betting opportunity every time, while most of the punters at the track would ignore it, focusing on the dismal recent performance. The odds on such a horse were inevitably 10:1 or longer, and I'd bet to win, place, and show. More often or not, I'd walk away from the payout window with a pocketful of high-denomination bills.

    ReplyDelete