By |
Stagerediness and Additional UserFactors |
freddie 3/30/2009 1:46:55 PM | Jeff the update is a handicappers dream....a couple of questions......1 What is factor stagerediness? 2 Is it possible to add and additional 7 user factors? UF6 thru UF12..... I've run out of them....I have far greater success with UF'S than UDM'S......Thank you for a sensational program Regards Freddie
|
jeff 3/31/2009 12:11:27 AM | Stage Readiness is still in the concept stage... something that I've been working on.
Basically horses fall into one of three separate categories or "stages" of readiness:
0 not measurably ready
1 readiness obtained from effort(s) in prior race(s)
2 readiness obtained from work(s) since prior race
Right now it's a "grabbable" factor in the IVTable Wizard only. I purposely put it there because I've been getting some interesting results after adding it to my own UPR. If you use the "Weight like Jeff" option (I've recorded part of the video already) the UPR Tools Interface assigns it a factor weight based on the win rate and roi from the sample used to generate the current Key Factors Report (although it doesn't yet appear on the visible part of the report itself in the Data Window) that you see in the UPR Tools Interface. The current sample win rate and roi are then compared to historical benchmarks for win rate and roi as determined using some pretty large (2 year all tracks everywhere)data samples and an estimated optimal factor weight is determined based on the ratio between the two.
I should point out that I'm still working on the weighting algorithm itself... trying to get it to closely mimic how I weight factors myself when creating an initial UPR.
At some point in the future, once I complete further R&D to the point to where I'm ready to say I've nailed down hard definitions for Stage Readiness that allow me to conclusively classify horses into the above 3 categories based on their records (just like I do now for CLEAN/DIRTY, etc.) then you'll likely see Stage Readiness added as a grabbable factor in the UDM Wizard and you'll be able to query it in the Data Window.
Sorry for being so long winded. But the question deserved a thorough answer.
-jp
.
|
jeff 3/31/2009 12:33:37 AM | To answer your second question about additional UserFactors...
Right now I have a prototype version of JCapper that handles 8 UserFactors. but those additional three slots do slow the program down somewhat during database builds and Calc Races routines - which is the reason why I haven't made them part of a program update yet.
At some point I plan to revisit program flow and try to address database build and calc races speed and possibly the number of factors handled by the program.
Every new factor that I add requires additional fields in playlist and daily export files - which slows the program down incrementally. But the cumulative effect of adding more and more factors over the years is that the current program takes longer to build a database and run a Calc Races than the older program versions that handled a smaller number of factors.
One thing I've considered is a program version where the user can define the number of factors to be handled by the program. For example, imagine a box on the System Defs screen where the number of factors handled is set = 50 and a second interface where the user has chosen his or her 50 factors from a list:
Factor 1 = CPace Factor 2 = PMI Factor 3 = Form Rating Factor 4 = JPR Factor 5 = Alchemy... and so on.
Such a program version would be capable of producing lightning fast Calc Races and Build Database Routines compared to the current program version.
I can envision having the initial download package come with a default factor mix. I can also envision some users going with as few as 30 factors and others (me for example) going with 180 or more.
Just a few random thoughts...
-jp
.
|
freddie 3/31/2009 8:11:19 AM | Jeff thank you for the in-depth explanation...what a program and what support!!!!!! Regards Freddie
|
dvlander 3/31/2009 10:35:07 AM | Jeff, I'll pipe in that I think your idea of the user limiting the number of factors considered to speed up the "Build" and "Calc" functions is a great idea. Actually, even with crowded Saturday cards, the "Calc Races" is never too lengthy of an endeavor. However, the "Build Database" is a major deterrent especially when you are trying to fine-tune UPRs.
I continue to be amazed at your resourcefulness and your ongoing support.
Thanks, Dale
|
Tom C 4/2/2009 1:38:52 AM | Jeff,
I have wanted to ask you if there could be a way to limit the number of factors so as to speed up the program but thought you would not be interested in pursuing that. Looks like you are one step ahead as usual.
Would limiting the preset filters also have the same effect?
Tom C
|
jeff 4/2/2009 12:12:36 PM | Limiting or doing away with Preset Filters wouldn't materially affect Calc Races or DB build time. It would however speed up Data Window queries just a little bit. Try running a query where the drop down is set to +Apply Special Filters. Then run another against the same playlist file where the drop down is set to -No Special Filters and note the difference in query speed between the two.
-jp
.
|