|
JCapper Message Board JCapper 101
|
| Home |
Register
Log In |
| By | Jeff.......comments | |
| JustRalph 5/6/2010 2:55:32 AM | Jeff I have never been able to get the "require favorable comments" or the opposite, bad comments to work worth crap. I would love to be able to find actual words, like " weakened" "Faltered" etc........ can you point me to some documentation on how these two comments filters work .... I am sure it's covered somewhere, but I can't seem to find it. I am also interested in a UDM that would flag when a low percentage rider gets replaced by a much higher percentage rider? Any thoughts? ~Edited by: JustRalph on: 5/6/2010 at: 2:55:32 AM~
| |
| jeff 5/6/2010 10:41:34 AM | Hey Ralph, It's been a while since I last did any serious R&D into chart caller's comments. During my initial R&D I identified specific keywords found in the comments and then evaluated public tendencies/reactions when it came to underbetting or overbetting each specific keyword. From that I developed 2 general purpose preset filters allowing JCapper users to make the following comment types part of a UDM Defintion: COMMENT= require favorable chart comments COMMENTX avoid favorable chart comments I also created preset filters (beginning with the characters shown below) to enable "grabbing" specific keywords found in the comment line and making them part of a UDM Definition: BORE TROUBLE XSTART WIDE COMMENT more to come.... -jp .
| |
| jeff 5/6/2010 10:57:09 AM | I never found chartcaller comments to be a magic bullet either. That said, I DID find that applying specific keywords from the comments to a UDM Definition gave me the ability to improve UDMs that already had a solid foundation. Let's do an example... First, I created a simple 3 factor UDM and ran it against my calendar year 2009 HDW database:
The UDM Definition explained: The factor (from my factor setup) sitting in slot number F28 is JPRClass - making WHERE RANKF28 = 1 from the above sql expression equivalent to setting min amd max rank for JPRClass equal to 1 (for the benefit of those who are still running JCapper in the original playlist file mode.) The other two factor contraints from the above sql expression set an MLOR3 min val of 1.60 and a PScore (HDW) min val of 95. That's it. Those are the only 3 factor contraints used in the UDM. more to come... -jp .
| |
| jeff 5/6/2010 11:03:31 AM | The next step would be to validate the UDM going forward. I used the Data Window's Create Quick Index File Export function (Click MENU and select Exports) to create a quick index file containing data for all starters from the first 4 months of 2010. Here is the same UDM (this time in playlist file mode) run against that quick index file:
more to come... -jp .
| |
| jeff 5/6/2010 12:41:55 PM | Now, the question becomes: Can we use any of the preset filters (looking at specific keywords found in the comments) to improve the UDM? One of the preset filters based on chart caller comments produced the following results on the 2010 data for the UDM:
But which one? Discussion of specific keywords from the chartcaller comments... as well as general pointers for getting better results from selections no matter how they are arrived at to be continued in the private section of the board... -jp . ~Edited by: jeff on: 5/6/2010 at: 12:41:55 PM~
|
| Reply |
|
Copyright © 2018 JCapper Software back to the JCapper Message Board www.JCapper.com |