Variegate Improvement Project
 

The question... Is it possible to improve the predictability of Energy's variegate above the current published 66%?

I believe the weakness in Energy's variegate predictor is due to a few factors that modern technology & cheap data files (DRF) may be able to help improve.

Proposition #1... All who have experimented with Energy have discovered the variegate is highly sensitive to contender/paceline selection, which is hindered by a manual data entry system that discourages multiple pace-line use, entry and/or replacement. Improvement may be found if it were possible to linearly view the electronic pace-line data within another program, make selections based on both linear and graphic screens which could then transport the selected paceline directly to an Energy file.

Proposition #2... Could the variegate predictor (and the overall performance) be improved by employing linear track-to-track and variant adjustments to the selected paceline prior to transporting that line over to Energy?

The program to create adjusted Energy files exists. I propose putting together a small group of like-minded individuals to explore the possibilities for fun, education & profit.


Yours truly,

JoeLong

Variegate Predictor
 

JL, I would not how to proceed with improving the variegate predictor. As you stated above Doc's studies showed the Variegate predictor to be correct within 1 66% of the time. By within 1 I assume he means a Variegate predictor of E/P as opposed to an E, or an SP compared to an S predictor. An E compared to an S is not within 1.

So, as you say their is a wide range for a possible false variegate predictor. We have the means to change a false Early variegate predictor by using the (C) Change Variegate procedure. The problem lies when you have a false other than Early variegate predictor. For example a S/P variegate. They don't have a procedure for changing this type of variegate.

Recently their has been many postings about the variegate predictor being off. Especially in terms of a false other than early predictor is where most problems occured. If the Variegate predictor said sustained the tendency was just to look at PARAGON B horses as win contenders, with the results being at times bypassing the win horse. Richie P finally decided to use other readouts such as the UXR to determine his win contenders.

To quote from page 58 of Follow-Up #26 Doc states " to compensate for the problem of a false sustained variegate, we have asked that you model your MUV and UXR. These numbers indicate just how far back a horse can be from the 0.00 MUV horse and still win. Hence if the Variegate is Sustained and the top two Sustained horses register a too high of a MUV or UXR number based on your records, you can mentally alter the Variegate to comply with your records."

JL, I know this does not really answer your question, but other than Doc's suggestion I would not know how to proceed on how to improve the Variegate. Others may have better ideas in this area. Your suggestion #2 concerns me about employing linear track-to-track and variant adjustments to the selected paceline before entering them in Energy. In affect wouldn't we be making another adjustment when these lines are entered in Energy which may have the opposite desired affect?

This is only my opinion on this topic, as others may have more knowledge in this area.

improving the variegate predictor
 

I have an idea on this and it has nothing to do with adjusting anything.Instead of worrying about a single paceline and whether it needs to be adjusted use the horse's ENTIRE past performances and consider any line in the pp that feels right. FOCUS on the break and first call to find patterns that horses exhibit.

Everyone looks to the far right of a past performance line for answers including adjustments. Maybe if we take time to LEARN better answers and MONEY can be found looking to the far LEFT of a line. NOBODY is looking there man. This also can completely eliminate any worries and work as far as adjustments are concerned. Food for thought.

BEFORE entering anything into software:

1) Select your pacelines for as many horses as you feel are needed.

2) Group all the "pure need to lead at FIRST call" horses together. match them against ONE ANOTHER and find the fastest at 1c and 2c. Take the fastest guy, write his line chice down, remove all others from WIN consideration. He is your "surviving early".

3) Group all the " early pressers" together and once again match them only against ONE ANOTHER. Find the one that PRESSES the 2f the best and has best overall speed. Write his line choice downand once again remove all others in THIS category for the WIN. he is your " surviving e/p"

4) repeat these steps for
a) presser
b) sustained.

at the end you should have MAXIMUM 4 horses left. the FASTEST in each of the individual running styles. After all we DO want to bet on FAST horses right?

NOW fire up the computer and ONLY enter these remaing horses into Energy! ( or Validator, Speculator, Synergism, Kgen, Homebrew etc etc) and see what shakes out. Results might really be something that helps everyone in getting a better "read" on the race ok?

************* CAVEAT ***************
ALL the above mentioned "styles" are based 100% on POSITIONAL/VISUAL and NOT computer generated Esp. If you choose to rely on computer stuff for this
Caveat Emptor


Hope this helps
Rich

Jim Bradshaw
 

EXACT quote from Jim's " 5 step approach to the Matchup"

Agreed, this is a total complete linear approach which Jim has mastered without the aid of a computer and is the crux of many handicapping programs which mostly errs on the side of the favorite.

Energy on the other hand is a departure from a pure linear model and takes the match-up concept into the realm of non-linear handicapping, which unfortunately appears to provide the user access to both brilliance and mediocrity in regards to picks when using on the variegate.

My point in all of this was to develop an approach to aid the program in improving its predictor model and hopefully produce more longshot winners, which I realize is a a mighty big task.

Hi Joe,
I am not sure anything Jim does can be considered linear! lol. he is kind of in a very different dimension.

If Energy takes the matchup into non linear areas( I really dont know what that is) then what better food to "feed it" than the TRUE matchup?

Fastest early
fastest early presser
fastest presser
fastest sustained

Kinda makes sense that this procedure is the true matchup.

PLUS NOBODY is doing this. always an added benefit just like " looking left" when the "crowd" is "looking right"

anyway that's all I can offer here as feedback.

great thread Joe!

Richie:)

Variegate Match-Up Manual
 

In the Match-Up manual that Bill has uploaded to the BBS Michael Pizzolla briefly gets into the Variegate, and I think he goes into the concept of how the Variegate is determined. His explanation is somewhat different than others presented in the Follow-Ups.

JL, I do not know if this will help you in your research, but starting on page 174 thru page 179 of the Match-Up manual you may find some of this information useful. Consider what he says in the first paragraph on page 178 of the Match-Up manual.

It seems to me the more we understand how the Variegate is determined the better chance we have of knowing when it may not be accurate.

Pizzolla Variegate
 

From what I can figure out what Michael Pizzolla is saying on page 178 of the Match-Up manual is something I have been looking at with the 0.00 Muv Horse and the Muv reading of the horses following it. What he calls lack of disparity in the second call I assume he means lack of seperation between the top MUV horses.

I made up some pacelines and beaten lengths to determine if I am on the right track here.

My horses, por and beaten lengths are as follows.

AAA 22.2 45.2 110.4 0-0.50-1.0-4.0
BBB 22.3 45.1 110.3 0-1.0-2.0-4.0
CCC 22.1 45.2 110.4 0-1.0-2.0-3.5
DDD 22.1 45.3 110.3 9.0-8.0-7.0-6.0

This is an extreme example, but this is just for illustrative purposes.

Variegate Screen 1
 

The Energy Generator

Variegate Screen 2
 

This is the second screen

Variegate Screen 3
 

The third Screen

Variegate Screen 4
 

This is what I think what Michael Pizzolla means about the lack of disparity in the second call. AAA, BBB, and CCC are not far apart in MUV.

Variegate Screen 5
 

The final screen

The Variegate Challenge
 

I think you may encounter many races like this where the 0.00 Muv and the horses following it are close. I have to believe the Variegate Predictor will always be other than Early in these cases. In this example I could see where some people might want to use DD6 as one of your win contenders since afterall he is number one PARAGON B. Not to say DD6 never wins a match-up like this, but his other numbers such as his low raw Total Energy and having the highest UXR should give one pause about this horse.

The only way you would really know if DD6 can win with a MUV of 3.01 at 6 furlongs is by modeling this number from your own handicapping. As I said this was an extreme example and some races are not as obvious as this.

As of now I really don't have any ideas on how to improve the Variegate predictor other than this.

Great stuff, things I had not thought about, thanks. My original thinking when starting this thread centered around the idea that paceline & contender selection coupled with surface & distance switches might be adversely effecting the variegate predictor, and with the proper paceline adjustments the variegate predictor could be improved.

My reasoning was that Energy was first developed on the West Coast under optimal track conditions where they can paint the variant on the wall because it rarely changes. Again, my thinking was that tracks outside of the So-Cal circuit were more challenging due to daily variant adjustments and if the paceline was adjusted and normalized prior to entering it into Energy maybe, just maybe there would be rewards for the effort.

I really appreciate all the deep thinking and sharing that goes on here.

All the best.

JoeLong

Houndog:

I was looking at your screen 4 and I do not think those MUV numbers are close at all. The numbers actually look fairly normal for an average race. DD6 has almost no chance of winning any race because an MUV over 2 (or so) is a definite disadvantage. I find that horses with an MUV over 2.00 rarely win - PROVIDED - and this is a BIG PROVIDED - the 0.00 MUV horse is a true contender with a valid line. Most of the time when I go back and look at a race where a horse with a MUV over 2.00 wins I usually see that I used a bad line (in retrospect) for the horse that earned the 0.00 MUV.

Gary

Yes Gary I gave an extreme example in this race. To you as an experienced user of Energy these numbers do not seem close. The program seems to think they are close enough to warrant a S/P variegate predictor, especially in the case of the top 2. As far as a horse like DD6 I used to play horses like this alot when I first started using Energy because of all the 1's in the sort.

I was just trying to illustrate areas where some users may need to look closer at other readouts. Obviously you know by experience and your records what the MUV cut-off point is.

Houndog:

Thank you for your comment. What I was trying to point out is that an MUV over 2.00 is always a bad sign. If the 0.00 MUV horse is a true contender then any horse in any race with MUV over 2.00 is in trouble. Do horses with MUV over 2.00 win? Yes, but not very often. It is a good way to determine to bet against an odds-on favorite.

Gary