PDA

View Full Version : Reliable source of game results?


Foosinho
28 Nov 2006, 10:17 AM
Do any of the leagues here in the US provide timely, easy to parse (mechanically) results? I mean, MLS provides timely results, but damn if I don't want to avoid writing a scraper to wrangle that HTML mess into XML, or even a flat-file format. The USL schedules page is just as bad - and in the past when I tried to compile results, they were often *very* *very* late.

So... anyone? If not, I could use some help creating scripts to pull all these results down into a database in a timely manner. I'd be willing to - at the very least - make a flat text copy (regularly updated, of course) of the database available for download, and for those of you who provide significant help, I might even consider allowing you access to the MySQL database.

I want to start back up the old rankings databases I maintained on the old CrewFan.com back in the day. That died because of the tediousness of manually tracking down and entering results wore me down over the years. Now that I write on a relatively major blog, I have even less time to do that kind of data-entry grunt work, and we can all benefit from some cooperation getting this set up.

scaryice
01 Dec 2006, 07:40 AM
There's only a few results per week, what's so hard about doing it manually? Actually I don't even know what you mean by "rankings database."

Foosinho
01 Dec 2006, 09:23 AM
There's only a few results per week, what's so hard about doing it manually?
I did it before - it sucks. You try compiling all the results from MLS, USL1, USL2, and PDL every week. That's like 100 teams. And - at least in the past - sometimes results couldn't be located for over a week after the games happened, which meant I had to keep checking over and over again, to ensure I didn't miss one. It sounds easy, but after doing it every week for 2+ years, it gets tedious.

Why not have a button that does it for me? And why reinvent the wheel - if somebody's already done that work, it seems reasonable not to re-do it.

Actually I don't even know what you mean by "rankings database."
That was a clumsy turn of phrase, wasn't it? I used to provide computerized rankings of pro soccer teams in the US back on CrewFan.com - not unlike Sagarin, or Pomeroy (basketball). The system I used then was based on FIFA's system, with weights assigned to each division (difficult to measure relative strengths, since the leagues are mostly not connected statistically). I'll probably reinstate that (off of the newer formula) as well as RPI and ELO rankings, provided for each league, at a minimum. I'm studying a few other (predictive) methods to see if they can be applied to MLS/USL.

But to do all that, I need match results, I need 'em in a timely manner, and hopefully I can aggregate them with little/no ongoing weekly work (I've got a job, family, and blog I have to maintain too).