UDMs - Framework and Style
UDMs - Framework and Style
A UDM or User Defined Model is an electronic definition, stored for frequent use, that describes a handicapping method in exact detail. In JCapper, users can create UDMs for specific tracks, distances, surfaces, and classes. UDMs can also be created to exploit the patterns of specific trainers, riders, and owners. Users can set factor constraints using rank, numeric value, and gap for any of the individual factors supported by JCapper. UDM definitions can be created using any of the factors in combination with each other. A UDM can be as as generic or as specific as the player desires.
A well thought out UDM is a thing of beauty and a true reflection of the player's preferred individual handicapping style.
My own live play is centered entirely around UDMs. If a horse isn't selected by one of my UDMs I refuse to play it. If a horse is selected by one of my UDMs I'll give that horse my focus - and make a play or pass decision as post time nears. That approach is a complete departure from what everybody else does. But it's an approach that has made me a winning player at the windows.
My own UDMs have a structure to them. I employ different types of UDMs to create a framework. That framework has a purpose to it. My framework reflects my style. Keep in mind that other successful players have a style all their own - and a UDM framework that fits their particular style.
Enough theory. Here are the basics of my own UDM framework:
Types of UDMs:
a. Tight Model UDMs
b. Loose Model UDMs
a. Rider UDMs
b. Trainer UDMs
c. Track Profile UDMs
That's my basic framework. Let's jump in.
At the heart of my framework is the speed pace form value business UDM. While I won't reveal my exact factor constraints, I will mention that ALL of my speed pace form value business UDMs bear a strong resemblance to the MY_Edge UDM presented in the Finding An Edge Help Doc.
I'll talk about two types of business UDMs. Tight model and loose model.
I'll cover loose models first. A loose model UDM is a UDM with very few factor constraints - one that is very loose (not overly selective) when it comes to picking horses. For example, a UDM having QRating min rank=1 max rank=1 as its only factor constraint would be an example of a loose model UDM.
As hard as it might be for most players to swallow, a complete newbie to the game using that exact UDM as the foundation or starting point for his entire game would actually do better than most players at the track or otb who use traditional past performances only. QRating rank=1 horses win about 31 percent of all races and return close to 90 cents for every $1.00 bet in the win pool. Most players who use traditional past performances and nothing else do far worse than that.
But such a UDM by itself won't make a newbie profitable because it picks too many horses. The UDM is simply too loose. It isn't "tight" enough to be profitable.
A tight model UDM has many factor constraints. Most of my business UDMs are true tight model UDMs. I've gone so far as to give some of them hundreds of factor constraints. I've purposely made them extremely "tight" (selective) about picking horses.
When one of my tight model UDMs does flag a horse, I know at a glance that the horse in front of me has a LOT of hidden positives and VERY FEW holes in its record. Many of my tight model UDMs produce single selections that can be played almost blindly to achieve a profit. And that's the entire point.
The My_Edge UDM presented in Finding An Edge would be an example of a good starting point for a true tight model UDM.
Let's talk about layering UDMs. Included in this category (for me) are UDMs based on riders, trainers, and track profile.
I use these types of UDMs to produce what I call a "Layering Effect."
I know of several players who do very well using these types of UDMs as primary sources of plays. I said earlier that one's framework should reflect
playing style. Keep that in mind as you read this. Your style may differ significantly from mine. That's the way it should be. No two players are
exactly alike. It doesn't make any sense that other players should see the game exactly the same way that I do. Pick a framework that fits
your style and go to it.
Here's what I'm trying to convey when I say the words layering effect:
Say for the sake of argument that one of my "business" speed and pace UDMs (the same type of UDM illustrated in Finding an Edge) is flagging a horse in a given race. As soon as I see the HTML Report I know at a glance that the horse being flagged has a lot of hidden positives in the areas of speed, pace, and form. But when I also have a rider based UDM pointing out the same horse, and maybe even a trainer UDM - and maybe even a track profile UDM - when all the stars line up - my understanding of that horse takes on a different dimension.
The converging UDMs tell me things about the horse at a glance that the public likely will have a very difficult time seeing - if at all. Not only does the horse have advantages in speed, pace, and form - but it also is being ridden today by somebody overlooked who's been riding lights out on the front end... maybe even for a trainer who's proven over the past several weeks that he's on his game too. And to top it off, the horse fits the track profile perfectly.
Such additional "Layers" of information have led me to some truly phenomenal scores over the years. Carefully kept records of my own live play continually tells me that horses flagged by Business UDMs that are also flagged by Layering UDMs are far stronger plays (better bets) at all odds ranges than horses flagged by Business UDMs alone.
I'll close with this final thought:
I want you to understand something. UDMs are the centerpiece of my game. I've spent thousands of hours at the Data Window over the years developing them. (I've been using computers and databases as a tool for playing horses since about 1985.) The UDMs I use for live play are the direct result of hard work. I want you to understand something else. Hard work has a way of paying off.
Best of luck to you,
Copyright © 2008 JCapper Software