"How do i record the ruslts in excel."

Excel is pretty easy to use and to setup databases. To record everything, you either have to type it all in by hand, or if you are lucky you can rip it off a website that is in a table format (like the rugby league results at stats.rleague.

Building up databases is pretty time consuming, but once it's done it's alot easier to mess around trying to create systems (maybe you just want to handicap?).

As far as research to do, look for trends. Try to get 5 years of data, and at least 3. Find a trend (eg home underdogs) and see how they performed over 1 year. Then test this trend over 2 or more different seasons. If the results then meet a certain rarity then they could or should be profitable, depending on how good the results were.

You can use this formula to test how good a test/record is:

=BINOMDIST(W,T,P,FALSE)

Just copy and paste into an excel sheet.
W= number of wins
T= total number of bets
P= probally of winning if you are guessing (eg a linebet =0.5)

So if I was betting against the line and won 50 out of 75, it would go:

=BINOMDIST(50,75,0.5,FALSE)

1 in 100 (1%) is nothing.
1 in 1000 (0.1%) means the trend might be profitable
1 in 10000 (0.01%) means the trend is almost certainly profitable.

Datamining for trends is the easiest way to profit off sports (in terms of creating 'pick').

It is VERY important that when you find a trend, you DON'T use that data in the test. Eg if I search for a trend of season 2, I only test on other seasons (eg 1 and 3 if you have 3 seasons). If you include the search data in the test it will be biased. In short, when you test a hypothesis, you can't include the data that was used to form that hypothesis. Very important.

As far as trends to look for, things like home faves/dogs and away faves/dogs. You might also look at performances when a team has had to travel alot.