College football just 4.8 months away!

Justin7 · 04-08-2008, 02:14 PM

I'm toying with undertaking a MONSTER project: handicapping all DIV I games this year in a similar manner to other sports: massive amounts of statistics.

If I'm going to do this, I need to get ahold of about 5 years of box scores that are nicely formatted, and I need data updated in a nice format every week.

Anyone here have good data scraping skills? If you do and want to work with me, I'll share the work product. I dinked around with this 4 years ago and found some monster overlays, but the model didn't get crazy strong until each team had played about 6 times.

Any volunteers?

Arnold · 04-08-2008, 02:39 PM

What exactly consists of the "work product"?

What data is required? Only box scores or individual player stats, etc?

I can do it, but I need to know first if it is worth my time.

Justin7 · 04-08-2008, 04:08 PM

I'm not worried about individual player stats. Mainly "team statistics". Rushes, rush yards, pass attempts, completions, pass yards, 3d and 4th down conversions, score by quarter, special teams yards (punting and kicking) and special teams scoring.

Also, each game needs an "ID" - this would include a home team, visiting team, week number, and date. It would be a plus to include a game spread, total, and halftime spread and total.

mofome · 04-08-2008, 04:39 PM

if you send me a pm to remind me (i dont read this forum a lot) i'll look for anything i can find. no need to give anything, its all about helping each other out. i'll see what i can find.

Arnold · 04-08-2008, 04:39 PM

This won't be a problem to get, if game stats for the last 5 years are publicly available somewhere on the internet. I haven't looked for it.

Are you building a model out of this?

Justin7 · 04-08-2008, 04:43 PM

Quote:

Originally Posted by Arnold

This won't be a problem to get, if game stats for the last 5 years are publicly available somewhere on the internet. I haven't looked for it.

Are you building a model out of this?

Yes. I have done lots of models - it's my main way of analyzing sports. I've maintained a 55% hit rate in NFL and arena, but NFL is mainly playing openers. I'd be closer to 51% if I had to play on game day.

Arnold · 04-08-2008, 04:59 PM

Well, if you're ready to trade your model for data, then I'm in. PM me if you're interested to discuss things in greater detail.

durito · 04-08-2008, 05:55 PM

If anyone has a link to a site with box scores that have data on sacks please let me know.

Thanks.