Research methods

This is something I also posted on the SBC forum, that I want to share with everyone.

I would like to discuss methods by which many of us do our research. I have to be systematic in the way I do my research otherwise it just won't get done, I even have a set procedure and I thought I'd share the basics with you all.

1. Decide on a topic, any one whatsoever as long as it's broad, such as "goalscorer", and put it at the centre of a spidergram on a plain piece of paper.
2. You have your main topic, "goalscorer", now do some research on niche markets. A good place to do this is on Oddschecker in "view all markets". Niche markets could include:
- Anytime goalscorer
- First goalscorer
- Last goalscorer
- Etc.
Add these niche markets to your spidergram and link them to the main topic at the centre.
3. Take a good look at your spidergram, decide on a niche market to research, try to choose one that's offered by at least 3 or 4 bookmakers.
4. Now you need a separate piece of paper where you're going to put the niche market you chose in the centre of the page and do another spidergram.
5. Come up with ideas that you think would directly or indirectly affect the niche market that you've chosen, for "first goalscorer" this could be:
- Directly:
* Opposition defence
* Opposition attackers scoring first
- Indirectly:
* Relationship with manager
* Not a fan's favourite, confidence low

Just some ideas but I think you get the message. Add as many factors as you can to your new spidergram.
6. Research as many of these factors as you can, there are lots of free sites out there where you can find the data.
7. Bring the data together, I find using Excel is the best method as you can put each factor in a separate column.
8. Analyse the data producing your own probabilities and afterwards your own odds, which you can then compare with bookmaker/exchange odds. It's not easy, but the more you practice the better you'll get, I'm sure of it!

Backfitting is when you force the system to work by over-filtering the data. My way is to use a thoery such as "Does Wayne Rooney score first more against lower league teams or not?", it's specific but if I can't prove or disprove the theory I move onto another one, such as "Does Wayne Rooney score first more against top half teams or not?"...and on and on until a theory comes up with a "YES". Then I work out probabilities, translate them into odds and compare them with the odds available from the bookies. If bookies odds are much better than my own, there's value, if they're a lot worse then I could think about laying them on the exchanges.

That's my way, would like to hear from others.

Koda