|
JCapper Message Board JCapper 101
|
| Home |
Register
Log In |
| By | Newbie:Where do you stop??? | |
| rosenowsr 10/21/2011 2:57:26 PM | I am sure this has been asked at least a thousand times but here goes anyway. Example=I have a UDM that has a single profile figure 0f 250/1000 with a positive ROI. When do you stop adding positive and negative profiles to that UDM?
| |
| jeff 10/21/2011 9:32:01 PM | A few thoughts on UDM theory if I may... I'll begin with some background about handicapping factors in general: If you pop open up a DRF and start looking at traditional past performances - and use colored pens to circle data points you think might be relevant - and spend a couple hours per card doing a thorough job of traditional handicapping.... I would not be at all surprised to see some of the more experienced players reading here end up circling a close to 100 (or even 500) unique data points. Ours is a complex game. I would venture a guess that most of us can probably name several hundred (some of us 1000 or more) unique handicapping factors. But if you step back and look at it in a big picture kind of way... Every handicapping factor that you can name fits neatly within the context of 7-10 different handicapping factor categories:
Sometimes the lines are blurred. Class horses have the ability to run faster speed figs than those of lesser class. So in that way speed is class and class is speed, etc. But for the most part, every handicapping factor that you can name fits neatly into one of just 7-10 handicapping factors. Next, UDM foundations: I prefer to start each UDM off with a solid foundation - one that gets a good win rate and flat bet roi right off the bat. It doesn't have to be a positive roi right off the bat (but at the very least it should fetch substantially better results roi-wise than say post time favorites do.) I also want my UDM foundations to be based on simple logic. If somebody were to ask me why I used a particular UDM foundation - I had better be able to answer that question intelligently using no more than 2 sentences. IMHO, the worst possible answer to that question is: "Because I'm grabbing a bunch of factor combinations that appear to work well together." My experience says that such UDMs often fail to produce going forward. (The player may as well have said "I'm backfitting but it's ok because I'm hoping for the best.") A far better answer is one that touches on big picture cause and effect: "My UDM foundation emphasizes horses that have an edge in areas X and Y over their fields while also having a twist in area Z. Large sample R&D clearly indicates the betting public shies away from area Z." Ok. On to your question... When is enough enough?... When do you stop adding factor constraints? After starting a UDM off with a solid foundation, the next step is to try to give the UDM factor constraints that touch on each of the above factor categories. I don't go nuts in any one category. But I will make the effort to involve every one of the categories named above - but only when results from the Data Window tell me I should. I would say if your UDM has a solid foundation, and one or two factor constraints from each of the above named categories - you've (probably) done your job in creating the UDM. The next step would be to test the UDM. Confront it with fresh races that are not part of the development sample. -jp .
| |
| JustRalph 10/22/2011 7:15:59 PM | After 3 years with Jcapper I have learned that "Less is more" very very often
| |
| ryesteve 10/23/2011 8:37:52 AM | I agree with the above. All too often, factors which do a great job of filtering out losers and boosting your ROI in your development sample, end up filtering out winners in your fresh data. You always want to keep independent data available to test anything you're doing.
|
| Reply |
|
Copyright © 2018 JCapper Software back to the JCapper Message Board www.JCapper.com |