HANDIFAST VERSION X

[headhawg]

Agreed. And Light accomplished exactly what he wanted. A$$hat.

[Speed Figure]

It's going way down hill FAST!!!

[Tom]

Sorry Doug, good thread and discussions for the most part.

[REBel]

Since I was the one who provided Handi with the results he posted above, I thought I should explain the various figures. Here are the column headers:

Cd

RACE TOP 4 RESULTS PAYOUTS

R5 3-6-10-5 3-9-2-12 5.80

R6 7-10-2-8 10-1a-7-9 15.80

R7 12-2-1-8 12-2-10-7 8.40 ex 46.40

R8 2-7-6-3 9-3-8-1 out



Aqu

R6 1-8-6-7 1-8-3-7 6.90 ex 16.50

R7 8-3-1-6 1-8-6-3 19.40 ex 46.60 tri 554.00 supr 1,588.00

R8 1-4-6-10 4-3-6-12 24.00

The ones Handi had in RED were the winners and shows where they appeared the Top 4. However, the headers won't line up.

[Handiman]

ReBel....I didn't mention it was you because I didn't have your permission.

Handi

[REBel]

Handi, no sweat. Hope your health is coming on like the winner of the CLASSIC.

[douglasw32]

Handi, Headhawg, and anyone else who wants to try it.

I do not think Light is that far off center (other than his bluntness)

The way I (we) wrote/released the code was to be open and things could develop off of it.

I do not feel that manula paceline selction would be feasible and would take away from the time saving of the program , however.....

The 2 places that I average could easily be set up to look (cycle) down the running lines looking for some criteria. It would only be in the EP and LP areas and the Class and PCR would take into account the recent average and the overall record respectively.

Much like what Raybo is working on for his spreadsheet.

Maybe do one automatic then work on selecting the defaults for it like head put into the points.

I would even take light and anyone else who has ideas advice on what to use as a criteria for the auto, in place of the average.

I have been playing with the thought and had good succuss, with choosing the most recent same condition (fst, sft, sloppy) on same surface Dirt/turf, at same distance, if none stick with recent take close dist, if none go beyond recent, if none do not worry about surface condition... etc, still none use last race. if more than one use most recent if both are recent, closest finish.

or something like that for the EP and LP, before the FAST is figured out ????

Just a thought.

[lsosa54]

Doug: I think you defined exactly how you calculated the FM and PC columns in one of these FAST threads? You remember which one? Thanks. Lou

[douglasw32]

Gosh no, but this is the form points...

=========

Roughly, I think it has morphed into something slightly different...

Todays Weight plus or minus below:

ES points -4 if ends up below 0 use o

sprint today and 0 ES points -1

work since last add distance of most recent work.
either of last 2 works since race = 5f add 2
no work and >45 days off -2

Ran last 7 days add 2
within 15 to 30 add 4
Has not ran in 90 or more minus -3

any fin pos from last 3 is itm +5
if none -3

Won last not by 3 subtract -2
won last by 3 award 1
came 2nd add 2

same surf as last add 1
same dist add 1

won more than a race in lifetime add 1
more than 3 starts never won -1

lower odds last 3 each time add 1

new trainer add 1
1st time lasix add 1
blinker change add 1

Jockey one of top 3 add 1

3 improving speeeds last 3 add 1
3 declining -1

Good ES +
Sprint with 0 ES <>


=============

PCR is pretty much exactly like scott did it but I used the Winners Speed Figures as the class up or down and used the AVERAGE speed figure of the horse as the baseline.

So like a 90 horse and a race of 96 winner, take away 6

So it is like add all 1st calls, add all final calls, subtract all above class races, add all below class races...

Divide by number of horses in the race, or something like that, I no longer have Scotts' book... but I could ask for it back and look it up if you need me to.

[ranchwest]

Originally Posted by douglasw32

Handi, Headhawg, and anyone else who wants to try it.

I do not think Light is that far off center (other than his bluntness)

The way I (we) wrote/released the code was to be open and things could develop off of it.

I do not feel that manula paceline selction would be feasible and would take away from the time saving of the program , however.....

The 2 places that I average could easily be set up to look (cycle) down the running lines looking for some criteria. It would only be in the EP and LP areas and the Class and PCR would take into account the recent average and the overall record respectively.

Much like what Raybo is working on for his spreadsheet.

Maybe do one automatic then work on selecting the defaults for it like head put into the points.

I would even take light and anyone else who has ideas advice on what to use as a criteria for the auto, in place of the average.

I have been playing with the thought and had good succuss, with choosing the most recent same condition (fst, sft, sloppy) on same surface Dirt/turf, at same distance, if none stick with recent take close dist, if none go beyond recent, if none do not worry about surface condition... etc, still none use last race. if more than one use most recent if both are recent, closest finish.

or something like that for the EP and LP, before the FAST is figured out ????

Just a thought.

I weight factors to arrive at an automatic pace line. The highest rated race is the pace line race. I don't like the idea of having rigid rules for finding the pace line race.

[lsosa54]

Thanks. No need to get the book back - that gives me an idea. If someone still has the book and can put the PCR methodology up knowing that Doug has used what's available in the BRIS file, it would be appreciated.

The FM and PC ratings have been producing some excellent mutuels at OSA in the top 3 or 4. As an example of a couple, check the last 2 races at Santa Anita yesterday.

In the 9th, the 2nd FM rating won and paid $20+. The top FM rating combined for a very nice exacta.

In the 10th, the 2nd PC rating just got up to beat the heavy favorite and paid $30+.

By focusing on these 2 columns for plays, my hit rate is lower but the mutuels are solid. And yes, I don't play the ratings blindly and do some additional handicapping before finalizing my plays.

Originally Posted by douglasw32

Gosh no, but this is the form points...

=========

Roughly, I think it has morphed into something slightly different...

Todays Weight plus or minus below:

ES points -4 if ends up below 0 use o

sprint today and 0 ES points -1

work since last add distance of most recent work.
either of last 2 works since race = 5f add 2
no work and >45 days off -2

Ran last 7 days add 2
within 15 to 30 add 4
Has not ran in 90 or more minus -3

any fin pos from last 3 is itm +5
if none -3

Won last not by 3 subtract -2
won last by 3 award 1
came 2nd add 2

same surf as last add 1
same dist add 1

won more than a race in lifetime add 1
more than 3 starts never won -1

lower odds last 3 each time add 1

new trainer add 1
1st time lasix add 1
blinker change add 1

Jockey one of top 3 add 1

3 improving speeeds last 3 add 1
3 declining -1

Good ES +
Sprint with 0 ES <>


=============

PCR is pretty much exactly like scott did it but I used the Winners Speed Figures as the class up or down and used the AVERAGE speed figure of the horse as the baseline.

So like a 90 horse and a race of 96 winner, take away 6

So it is like add all 1st calls, add all final calls, subtract all above class races, add all below class races...

Divide by number of horses in the race, or something like that, I no longer have Scotts' book... but I could ask for it back and look it up if you need me to.

[hcap]

I have been doing a study of the various factors in HandiFast.

So far approx 500 races. Using current races at CRC, WOX, and OSA. And some old BEL files from 2006. Eliminating maidens provides 262 races. Using only the top ranked Fair Odds horse.

PLAYS
262
BET $524

WIN%
27.9%

ROI
0.2%

[douglasw32]

Cool, That is a ton of work...
Can you lose anything around two turns, 2 year olds and Turf.

Smaller sample size, but I am curious since that is what I subjectivly think works best.

no proof.

[hcap]

Using a few more "Ifs" in addition to the top ranked Fair Odds horse

1-Play only if The jockey is ranked among the top 5 in the race.
2-Play only if the morning line is at least 80% of the actual Fair odds of the top ranked Fair Odds horse.

190 plays
WIN%
30.0%

ROI
13.1%

.................................................. ............................

Add 1 more filter.
Play only if the top ranked Fair Odds horse is among the top 5 on the morning line.

179 plays.
WIN%
31.3%

ROI
15.2%

Not bad at all

PS: I am using headhawg's CSV output and Excel to keep track.

[headhawg]

Hcap has been doing a lot of nice research. He's just waiting on me for some things that I haven't quite gotten around to. Given the positive ROI shown, I will be good and refrain from saying something about pacelines here.