My baseball model has excelled for me for years. However, through the first month and a half of the season there has been a significant reverse in a trend that is the basis for my model. It's unwise for me to assume the books got sharper on this particular trend as they had not adjusted for many previous years and the trend has done a big reverse, not just a couple percentage points. I know its early but the sample size of games my model has applied to so far this year has been somewhat significant. I have adjusted so far by narrowing some of the inputs but i would hate to give up on the wider selection of games the model was so good on for so long. anyone have any comments? thanks
-
-
It would also be unwise for you to assume the books haven't got sharper on this particular trend.Comment -
In some sports (NFL comes to mind immediately) the sport changes so much due to rule changes and the fact that coaches/players must react to them that one see "trends" coming and going all the time.
This will not help you fix your trend problem, but perhaps to realize that relying on trends will most likely fail in the long run.Comment -
maybe my semantics were misleading. the current problem is a definitely "trend" and statistical outlier(somewhat significant sample however), the long term model has been solid for years, but i guess you answered my question and that i should just assume the trend will correct itself. I know thats the answer but I think the strategy is just to stick with the narrowed criteria thats still working for now until the corner has clearly been turned on this current trend. Less money being made for now but as buffet says rule #1 don't lose money, rule #2 don't lose money. very frustrating
without going into too much detail, the chance of this being the case is remote. being as that I live and die by statistics, I have to give it SOME chance but it is very unlikely. the unadjusting of the past years is only one piece of a few that has led me to believe im right on this countIt would also be unwise for you to assume the books haven't got sharper on this particular trend.Last edited by king9; 05-13-11, 02:54 PM.Comment -
mispriced subsets win... until they don'tComment -
Sports evolve and statistical trends tend to fade. If a syndicate modeler finds a significant trend, they will tend to pound on it with significant action and the market will adjust to the pounding.Comment -
stat errors
here's a great little bookComment
