|
JCapper Message Board JCapper 101
|
| Home |
Register
Log In |
| By | UPR Tools Intro | |||
| jeff 10/24/2008 7:17:04 PM |
UPR Tools Button A large part of this feature is still under construction. But I'll post a brief overview... Clicking the UPR Tools Button clears the IVTable Wizard and causes it to expand, revealing several new visual elements. The new elements are designed to facilitate (speed up) the process of keying in IVTable entries to create UPR and UserFactors. Before going any further I'd like to emphasize two things: 1. ALWAYS make a backup of the following file BEFORE launching the IVTable Wizard: c:\2004\JCapper.mdb I can't stress the importance of this strongly enough. At least once a month I get a call or an email from a user who through human error has accidentally modified or deleted entries from a UPR he or she was happy with. The best - no strike that - the ONLY way to recover from such an event is to revert back to the last known good backup. Be religious about backing up this file. Get into the following habit: Don't even open the IVTable Wizard without backing up the file first. 2. Avoid Backfitting. IMHO, the most important part of creating effective UPR and UserFactors resides in the area of laying a solid foundation. Selecting factors from different key areas (such as pace, class, form, ability from speed figures, connections, etc.) that encapsulate what is actually happening out there on the race track goes a long way towards laying the right foundation. Once you have a solid foundation, then some Data Window R&D where you experiment with different factor weights can help you get UPR and UserFactors closer to "right." Please don't get the idea that grabbing a bunch of factors from a recent multiple factor Data Window export (just because the factors grabbed were roi positive in a recent short term sample) is going to get you a long term killer roi. The most likely outcome from that kind of thinking is backfitting. Backfitted models seldom produce good results going forward. Be aware that the UPR Tools Interface lends itself to easy creation of entries for UPR and UserFactors. Resist the temptation to blindly grab a bunch of recent plus roi factors from the drop down and propagate them into your UPR. If you do that, don't say I didn't warn you. Experience has shown that backfitting doesn't work. I'll come back and post more as time allows... -jp . ~Edited by: jeff on: 10/24/2008 at: 7:17:04 PM~
| |||
| jeff 10/24/2008 7:02:56 PM | When developing my own UPR and UserFactors, I tend to use large chunks of data. I'll run the most recent 4 consecutive calendar quarters through the Data Window breaking the data out by factors that interest me. Such samples (for the most part) provide clean orderly impact values that seldom need to be modified when creating IVTable entries for UPR and UserFactors. Perhaps the most common "complaint" I hear from users attempting to create UserFactors and UPR is "My database is too small to represent a meaningful sample. My impact values are all over the place." Enter UPR Tools... UPR Tools Visual Interface The Template Drop Down The Template Drop Down is designed to create matched pairs of IVTable entries for you. You decide how deep into the field and how steeply they are scaled. After clicking UPR Tools and New, select a Template from the drop down. Most of the required fields for creating a set of IVTable entries will auto populate with values for behavior = 0 Rank. From there you supply additional field values such as GroupName, Factor, Weight, Track Code, Surface, Min and Max Distance, Class Descriptor, etc. After that, click the Save button. Bang! You've just created/added a new set of entries for a single factor within your GroupName. The advantage of using a Template instead of impact values from the Data Window is that the entry process is fast. The entries are also smooth and scaled... especially helpful if you haven't yet accumulated thousands of data files. Will come back and post more later as time allows... -jp . ~Edited by: jeff on: 10/24/2008 at: 7:02:56 PM~
| |||
| DeanT 10/26/2008 5:40:35 PM | I have not done this yet. Concentrating on tracks specifically and playing them has been my method of operation. So, for a beginner: Would you/can you make All weather UPR's and dirt UPR's and turf UPR's, or do we just make one UPR for everything, like a JPR? Do we make them for sprints and routes? Do we make them for specific, tighter distances? I have been meaning to do this for a long long time and I think it is about time I get cracking.
| |||
| jeff 10/27/2008 12:32:06 PM |
Yes to all of the above. When I came up with the idea for JPR I looked at large data samples at all tracks everywhere. At the time I'd never even heard of synthetic surfaces. There were just two surfaces: dirt and turf. Large data samples clearly showed that the dirt surfaces of the tracks I was following were similar enough in nature that one algorithm could be used to produce reliable results at virtually any dirt track in North America. Ditto that (almost) for turf. Enter poly track, tapeta, cushion track, and now pro ride... all of which do a really effective job of getting horses with early speed to spit the bit about the time they turn for home. People are going to hate me for this next comment. But I'll call it as I see it. After watching a few thousand races run on artificial surfaces, I'm convinced the overwhelming majority of races no longer go to the swift. On artificial surfaces they go to the wide and the slow. Please understand I'm not talking about the Zenyatta's of the world. I'm talking about the maidens, claimers, allowance, and minor stakes horses that make up the bulk of most cards. I'm not here to argue the pros and cons of the various surfaces. My job as a bettor/software developer is to come up with ways to beat the game no matter what they run on. Hell, if they raced by swimming them in jello I'm convinced that there'd be combinations of JCapper factors that could be used to get an edge over the crowd. Like it or not, synthetic surfaces change the nature of the game. One of the biggest advantages of UPR and UserFactors over "canned" ratings like CPace, PMI. CMI, Opt Pts, and JPR is that YOU THE USER have the ability to make them surface and distance specific. You may never be able to devote the kind of time to this that I do. Betting horses doesn't have to be the central focus of your life. You can take a "canned" rating like JPR or PRating or QRating or Alchemy and use it as a foundation for an artificial surface UserFactor or UPR by giving it say 80-85 pct of the total weight in your factor mix and then use the tools in the program to see what happens when you add factors like LateSlant, PAL, AFR, and Running Style to the mix giving them something like a combined 15-20 pct of the total weight in your factor mix. Holy run on sentence Batman... but I hope you get my point. Artificial surfaces ARE beatable. If you want to specialize, you HAVE the tools. It doesn't take all that much effort to come up with UPRs and UserFactors that are tailored to specific distances at specific surfaces. You don't have to reinvent the wheel. Start by taking a "canned" power rating. Go from there by adding factors based on how you think the races are actually being run on the surface distance category you are going after. You'd be amazed what can happen once you have a solid UPR that's different from the way the public bets if you wait for the overlays. I hope that rant makes some sense. -jp . ~Edited by: jeff on: 10/27/2008 at: 12:32:06 PM~
|
| Reply |
|
Copyright © 2018 JCapper Software back to the JCapper Message Board www.JCapper.com |