Any Stk/Stk Ultra users?

[upthecreek]

I was wondering if anyone was still using the software I found a download link for STK 2002,it seems limited and the support forum is closed Any info appreciated. Thanks!

[PaceAdvantage]

I think I remember that program. It was nothing more than backfitting at its best if I remember...I still think of it fondly once in a blue moon because of all the hope the test runs on past races would offer you... :lol:

[upthecreek]

I think I remember that program. It was nothing more than backfitting at its best if I remember...I still think of it fondly once in a blue moon because of all the hope the test runs on past races would offer you... :lol:

Im not sure what is meant by "backfittting" Youre suppose to add results(which I havent figured out yet) and then adjust the weights accordingly to which factors are producing the most winners for the track youre playing and then there is something about the odds line which I havent figured out yet either. Dartman51 is looking for his info/notes he has on the program to help me outIm going to post todays Sar picks over in the selections forum.

[headhawg]

Essentially, backfitting in this context is creating the weightings based on races that already have run in order to optimize the Win% or ROI. The major problem is that there is little chance that the weights have any kind of predictability for future races which is why PA gave it the :lol:. In other words, the program is giving a user the illusion that if he/she set these optimal weights they would have a 35% win / 1.35 ROI (or whatever) on races yet to be run.

I'm no stats genius, but if you could split the past races into two samples, run the "optimal weight algorithm", and then test it on the second sample to see if the Win%/ROI remained the same then you could be (slightly) more confident in the weightings.

Handicapping is all about backfitting -- we look at past performances to predict the outcome of future events. The trick is to find factors that have predictive ability and that few other people have. The vast majority of handicapping programs don't have that capability. I think the open source stuff that Delta Lover and others are talking about in the Software section is attempting to build a better mousetrap.

[upthecreek]

Essentially, backfitting in this context is creating the weightings based on races that already have run in order to optimize the Win% or ROI. The major problem is that there is little chance that the weights have any kind of predictability for future races which is why PA gave it the :lol:. In other words, the program is giving a user the illusion that if he/she set these optimal weights they would have a 35% win / 1.35 ROI (or whatever) on races yet to be run.

I'm no stats genius, but if you could split the past races into two samples, run the "optimal weight algorithm", and then test it on the second sample to see if the Win%/ROI remained the same then you could be (slightly) more confident in the weightings.

Handicapping is all about backfitting -- we look at past performances to predict the outcome of future events. The trick is to find factors that have predictive ability and that few other people have. The vast majority of handicapping programs don't have that capability. I think the open source stuff that Delta Lover and others are talking about in the Software section is attempting to build a better mousetrap.

I dont know if that Id call that backfitting , maybe my idea of it is different
Doesnt Handifast, MT Handi and Dave Mahiliks Matrix use weights based on how predictive the factor is?

[headhawg]

Well, that's the way that I understand backfitting and I think why -- under most circumstances -- the programs that use it don't make the handicapper successful. Backfitting is really only "predictive" of races that have already been run, not future races. You can't bet on past races. Kind of makes it useless to a certain extent, no?

I don't know anything about Matrix, but yeah, Handifast has user-controllable weights. As designed, it takes a tremendous amount of work to test weights to come up with some combination that's "optimal" as there is currently no way to batch-run files. There are a few people who have come up with some weight combinations that work for them, but I don't think that anyone is naive enough to think that their settings are going to take them down the road to profitability. And if they did, well then, that is just great for them.

[upthecreek]

Well, that's the way that I understand backfitting and I think why -- under most circumstances -- the programs that use it don't make the handicapper successful. Backfitting is really only "predictive" of races that have already been run, not future races. You can't bet on past races. Kind of makes it useless to a certain extent, no?

I don't know anything about Matrix, but yeah, Handifast has user-controllable weights. As designed, it takes a tremendous amount of work to test weights to come up with some combination that's "optimal" as there is currently no way to batch-run files. There are a few people who have come up with some weight combinations that work for them, but I don't think that anyone is naive enough to think that their settings are going to take them down the road to profitability. And if they did, well then, that is just great for them.

Well the way I have the weights set on Handi multi track produced 8 winners out of 10 races today at Saratoga with 6 top winners Of course its not typical but does seem to contain the winner in the top 4 70-80% of the time.

[headhawg]

Fantastic, UTC. I love to read that people are still using Handifast and having success. It almost motivates me to finish Version 3. Almost. :)

Best of luck.

[upthecreek]

Fantastic, UTC. I love to read that people are still using Handifast and having success. It almost motivates me to finish Version 3. Almost. :)

Best of luck.

Here's hoping that you do,looking forward to it. Did I read in a thread about Handi MT someone was working on a print function--thats the only draw back.
Thnx!

[Partsnut]

As I recall the original STK software was pretty damned good. Truthfully,
I used it many years back and made some nice money at Belmont.
The software evolved into STK ultra which in my opinion, was never as good as the original STK.
It was indeed a back fitting program but set up correctly and used with a progressive betting method success was achieved.
I started with a mere $147.00 bankroll and at the end of 3 weeks my bankroll grew to $3596.00. At that point I stopped at the request of my late wife. She had little confidence in gambling for a living. I am a peace loving man so I honored her request. :)
As far as I know, the original site and users group dissolved and to my best knowledge does not exist anymore. If it does then please correct me and let me know that it does.
The software was developed by an engineer whose first name was Vern. I lost contact with him many years ago. He was indeed a fine gentleman.
STK was one of the better offerings I have seen and of course the price was right. It was free.

[Vinnie]

Parts:

It is great to see you posting. I do believe that you are completely correct. STK was from I believe Vern Brown. He is a first class guy and he wrote an excellent program. I could never quite get with the STK Ultra program when I was using it, but, STK in its original form was really quite good.

Take Good Care Parts and nothing but the Best to you Always. :)

[upthecreek]

http://www.stkusers.com

[headhawg]

Here's hoping that you do,looking forward to it. Did I read in a thread about Handi MT someone was working on a print function--thats the only draw back.
Thnx!The MT version was primarily Handiman's work and AFAIK he wasn't working on a print function. (I hope that he is ok; I haven't seen him post in a long while.) In the latest version of Handifast HHX (2.3?) you can print as that feature was added by mmdorn. Unfortunately, you can only load one track at a time. In HHXv3 you will be able to load up to 12 tracks and print. :) I just have to get back to coding it.

[upthecreek]

The MT version was primarily Handiman's work and AFAIK he wasn't working on a print function. (I hope that he is ok; I haven't seen him post in a long while.) In the latest version of Handifast HHX (2.3?) you can print as that feature was added by mmdorn. Unfortunately, you can only load one track at a time. In HHXv3 you will be able to load up to 12 tracks and print. :) I just have to get back to coding it.

I think it was another member who had set it up to print for himself and he was going to release to other members if it was okw/Handi I'll search the threads

[Partsnut]

Vinnie

Parts:

It is great to see you posting. I do believe that you are completely correct. STK was from I believe Vern Brown. He is a first class guy and he wrote an excellent program. I could never quite get with the STK Ultra program when I was using it, but, STK in its original form was really quite good.

Take Good Care Parts and nothing but the Best to you Always.
-------------------------------------------------------------
Hello Vinnie,

Thanks for your kind words.
It's good to get back into horse racing after a long layoff.

[headhawg]

I think it was another member who had set it up to print for himself and he was going to release to other members if it was okw/Handi I'll search the threadsmmdorn added printing and the ability to change weights to the C-C, P-S, and PWR fields to HHXv3. More information can be found in the Handifast section in the Download Here.... thread. We should probably move further conversations about Handifast to that area as this thread was about STK.

[upthecreek]

Fantastic, UTC. I love to read that people are still using Handifast and having success. It almost motivates me to finish Version 3. Almost. :)

Best of luck.

Just one last note; yesterday(0808) at the SPA HandiMT had 9 winners in the top 4 out of the 10 races 5 in the top 2 and 7 in the top 3
Also boxing the 1st 3 selections in the feature resulted in a $300 tri
Ill move future comments to HF software thread and may post HandiMT selections for SAR today/tomorrow