I'm posting this in hopes someone out there has had some experience in this endevor and would be willing to share some "lessons learned".
I had some experience in using this analysis method in graduate school, but not for sports. I'm considering a 90 -100 variable set and I anticipate having the bugs worked out by kickoff next year.
Thanks in advance for your thoughts.
0
To remove first post, remove entire topic.
I'm posting this in hopes someone out there has had some experience in this endevor and would be willing to share some "lessons learned".
I had some experience in using this analysis method in graduate school, but not for sports. I'm considering a 90 -100 variable set and I anticipate having the bugs worked out by kickoff next year.
You may use 90 to 100 variables but your weighting of each variable will vary greatly so as to render some of these considerations meaningless.
The only advantage to the regression part of the method is that you are acknowledging the fact that the results are headed to 0 reregression or a 50 / 50 point . (Such as over/under will eventually even out; theoretically) but the method is not accurate enough to choose a point at whch you have a definite advantage to make a prediction on a particular game.
You may be able to discover trends that we can then tweak and manipulate to our advantage; then let someone else post the games and results like Expert and gasman and srtzee have done.
The best results come from the Wikinomics aspects of this site; put out your well thought out and researched idea and then let everyone own it and use it to the benefit of all.
Other lesson learned: best results come from a mix of subjective and objective considerations, which is why I like handicapping. Test your methods exhaustively before making any claims.
All in all, thanks for contributing to our community. Anything of value is greatly appreciated.
0
JJ
You may use 90 to 100 variables but your weighting of each variable will vary greatly so as to render some of these considerations meaningless.
The only advantage to the regression part of the method is that you are acknowledging the fact that the results are headed to 0 reregression or a 50 / 50 point . (Such as over/under will eventually even out; theoretically) but the method is not accurate enough to choose a point at whch you have a definite advantage to make a prediction on a particular game.
You may be able to discover trends that we can then tweak and manipulate to our advantage; then let someone else post the games and results like Expert and gasman and srtzee have done.
The best results come from the Wikinomics aspects of this site; put out your well thought out and researched idea and then let everyone own it and use it to the benefit of all.
Other lesson learned: best results come from a mix of subjective and objective considerations, which is why I like handicapping. Test your methods exhaustively before making any claims.
All in all, thanks for contributing to our community. Anything of value is greatly appreciated.
Thanks for your post I have been wrestling with with weighting aspect of this approach (i.e. garbage in - garbage out). I may find after traveling this path that the variable set is too large and or contradicting itself as you have suggested above.
My thought is to historical data from the 2008 season to run trials and just see what may turn up if anything. I will also take great care in distributing any findings that may turn up. I have learned in a short time on this site that there are folks out there who seem to be very misguided in blindly following some of the picks and find themselves in a financial bind blaming the person who posted the pick. I want to try to avoid that if all possible.
Thanks again !
0
dabar
Thanks for your post I have been wrestling with with weighting aspect of this approach (i.e. garbage in - garbage out). I may find after traveling this path that the variable set is too large and or contradicting itself as you have suggested above.
My thought is to historical data from the 2008 season to run trials and just see what may turn up if anything. I will also take great care in distributing any findings that may turn up. I have learned in a short time on this site that there are folks out there who seem to be very misguided in blindly following some of the picks and find themselves in a financial bind blaming the person who posted the pick. I want to try to avoid that if all possible.
Hot Forum Topics
