SilkBC Posted November 2, 2009 Share Posted November 2, 2009 I have recently been looking at the use of ELO ratings to assist me with my betting. I have been impressed with a1ehouse's work, in particular with his "LtD" thread as I primarily trade on BF doing just that. Since I am not knowledgeable, as yet, in databases, I have been able to crrate a spreadsheet that calculates the ELO rating for a team based on the work of Bon Runyon of the "World Football ELO Ratings", which I believe a1ehouse uses. I planning on doing each league as a seperate sheet within the wrokbook, with each workbook comprising a single season. (I have been entering the data needed for the calculations manually from stats sites, only for previous season as a proof of concept, but plan on using the CVS data from football-data.co.uk, which will allow me to add the odds for each match so I can do some back testing and also add the previous seasons relatively quickly, and for many different leagues) I have also seen the use of another calculation method used that I have seen referred to as "rateform" which involves a "kitty" to which the home team contributes 7% of their rating and the away team 5% of their rating. The winner of the match takes the whole kitty, which gets added to their rating. A couple of questions for those who are familiar with this: Are there any advantatges to using one over the other other, or does it basically come down to personal preference? (Related to #1 above) Has anyone compared the two methods and determined if one is more accurate than the other? If not, my template can be adjusted and modified to calculate the rateform points too with the same data, so it should be relatively easy to compare. It seems some reset each team's rating at the start of the season (to 1,000 or 1,500, or whatever points they decide to use), while others seem to carry over the points from season to season (making minor adjustments). Is it considered best to carry the points over, or reset them? I think those are my only questions for the moment. I may have others, or other questions may be created based on the answers to the above :-) Once I have enough data, I plan to use regression to determine for myself what the odds should be based on the Win Expectancy. That is the other reason I am doing each league as a seperate page in my workbook, because I suspect that each league is different (or at keast each country is different; I may combine al the English leagues into one, but I will have to see) TIA! :-) -SilkBC Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.