Monday, November 16, 2015

New T-Rank features

In case you haven't noticed, I've made some major improvements in the T-Rank for this season:

1) A specific page for every team, with predictions, results, and team stats (more to come).

2) A page for each conference with similar stuff, and more to come.

3) Player stats, found at the bottom of each team page. Eventually I'll put up some overall and conference stat pages as well.

Basically, I'm now able to do a lot of the stuff that Ken Pomeroy does at his site. T-Rank is sort of becoming Kenpom Lite. But if you're reading this, you should definitely pay the $20 and get yourself a Kenpom subscription.

I've got a lot of other ideas for improvements that I think I can implement, given the time and motivation. Now that I've broken away from Excel and figured out how to get player boxscore data in a usable form, there's not much of theoretical limit on the kinds of analysis I can run.  (There are a lot of practical limits, though, of course.)

If you've got any ideas for something you'd like to see, let me know.


  1. Hey Bart - Nice site! Cool to see your twist on kenpom.

    May I ask what language ans source you used to retrieve your basketball data?


    1. I use Python and scrape (***shhhh!***) ESPN for the data.

    2. Really cool, thanks for the info. Would love to learn how to do this.

      How do you determine home vs. away vs neutral designation?

    3. I actually have my own master schedule file that I created before the season (and revise & update as necessary) which is cross-referenced with the scraped data. My master schedule determines the home/away/neutral. However, the espn box score pages do have data on arena and location, and I now have enough data that I could do it that way as well. But since I'm trying to project future games I need to know where future games are supposed to happen anyhow, so I need that master sked file in advance either way.

      I actually just taught myself enough Python to do this right before the season this year -- starting from scratch with no programming background (except for some proto-programming I did in 8th and 9th grade -- almost 30 years ago!) You can do it! I'd recommend the web page "Learning Python the Hard Way" as a good way to learn, that's how I got into it.