How to conduct sports betting research

[Beyond the Bets]

For those that clicked the "contributors" thread and want to contribute articles or research but don't know how, I figured I'd put together a very simple "how-to" post so that you have a better idea of how to approach this.

All research begins with a question or a point of inquiry. The question usually comes from something you observe. As an example, let's say you notice that the Duke Blue Devils rarely cover at home against no-name non-conference schools.

"Hmm," you ask yourself. "I wonder if this is true?"

You might decide to search Duke's ATS history over the last few years against non-conference teams at home. But before you do, it's best to come up with a theory and a hypothesis. A likely one:

Theory: Most people will blindly bet Duke when up against a no-name school. Bookmakers, knowing this bias, overinflate the lines.
Hypothesis: Duke will have a poor ATS record against no-name schools.

Then, you search Duke's ATS history and arrive at the results. The interpretation of the results and determining their validity is often times more important than the actual results itself.

Most people rely on sample sizes that are incredibly small. Others merely "data mine" for trends that aren't explainable and have no predictive quality. The majority of trends you see, I'm sure, are completely random.

Knowing what is and isn't, and asking good questions — questions that might produce interesting answers — is absolutely critical. Also, you want to make sure that your results are generalizable, if possible.

How the Houston Cougars basketball team fared ATS from 1997 to 2004 probably isn't going to help us much in 2011. There is nothing special about their team and they are relatively obscure.

However, it's possible that knowing how Duke fared ATS from 1997 to 2004 IS helpful. Duke is a brand. They are also a very public team. They were a brand in the late 90s and a very public team in the late 90s, no different than what they are today.

Some of what you find, then, could possibly be applicable to today's Duke teams.

Any questions?

[RyanB]

(11-16-2011 11:40 PM)Beyond the Bets Wrote:  Any questions?

Where do you suggest going to find old lines? Do you have favorite places online that you go to for concise information? Do you have a master list like you do for #twitter?

[oddsok]

(11-17-2011 12:07 AM)RyanB Wrote:  

(11-16-2011 11:40 PM)Beyond the Bets Wrote:  Any questions?

Where do you suggest going to find old lines? Do you have favorite places online that you go to for concise information? Do you have a master list like you do for #twitter?

Old lines or old ATS statistics?

Covers is a good place to start though.

If you're wanting to know whether Vandy covered in the 02/03 season vs Tennessee you can go here for example.

[Beyond the Bets]

Yep, odds nailed it.

[spanishredeye]

http://sportsdatabase.com/ is pretty useful, but some may find it complicated to use.

http://sportsdatabase.com/nfl.py/query has some sample queries for NFL records.

[Beyond the Bets]

I just started a thread posting all the references or tools that I'm aware of on the Internet.