Looks like part 1 and part 2 are the same video.

Fixed.
Thank you.

Dave is right about Deep Learning/Neural Nets being prone to overfitting to the data at hand. In my experience, when we are dealing with tabular data (as opposed to hard to feature engineer data types like images or video input) even more primitive algorithms like Logistic Regressions can outperform overly complicated and brittle pure Deep Learning/Neural Net approaches.

AI Best Horse: Odds Sensitive Handicapping. The AI BH odds go <2/1, <3/1, <9/2, 3/1+, 4/1+, 8/1+, 17/1+. Curious why the <9/2 and then 4/1+. I know the B=6 and we're looking at 5/1+, so it doesn't factor in, but just curious.

B=6 is 5/1+ and B=11 is 10/1+. So, are the odds always minus 1 of B? B=8 would be 7/1+.

Looking at the race example MVR05. #11 Scarlet is showing AI Odds 14.63. The 17/1+ column is circled red. Should that have been under the 8/1+ column. It's still a 1 & 1. Again, just following along and trying to understand the screens.

The choices were due to research done by a simple engine I created a long time ago that crawled the data trying different odds ranges.

These were the significant breakpoints that produced changes in the factor makeup.

Thus, 4/1+ horses contain a better factor makeup than 7/2+ - more CONSISTENCY IN LIKENESS.

IOW, they were built not to fit the need for human conformity to a strategy but rather where the data took me.

Have totally enjoyed these videos. Knowing the Balo in a race would really help your win percentages.

Have totally enjoyed these videos. Knowing the Balo in a race would really help your win percentages. — Ed Malz

By itself it is enough to win.

Dave,

It sounds like your 3,600 systems using TDA overlap races within your "1 segment of data." Is that correct? In simpler terms, you trained 3,600 clusters of overlapping race samples based on a rule-based approach that I assume groups similar races. Am I in the ballpark?

No.
One giant sample and let the AI do what it will.
IOW, you might have a 4f race for young horses that the AI chose to handicap like a mile on the dirt for older horses.

I simply have no idea how they are segmented.

I only know that it is powerfully consistent.

But know that I did zero segmentation.

Ah... you're thinking about the Topological Data Analysis!
Actually a good way to think.

But what I did was let the AI segment THE OUTPUT instead of the input.

the output meaning the outcome of the race and the associated factors of the winners?

the output meaning the outcome of the race and the associated factors of the winners? — Dennis

The output of the race... but no idea what the factors are.

"I think it deserves a Nobel Peace Prize, but I know nothing about dynamite."

Not so sure about that. deTerminator looks like it has the potential to blow up
the way horse race handicapping is done!! The Dave Schwartz version of TNT.
Old Alfred may have been bested.

Definitely worth a Nobel Prize for Science, though.

"I think it deserves a Nobel Peace Prize, but I know nothing about dynamite." — ponyplayer

Can't believe I missed this reply for NINE DAYS!
If you look up Nobel peace prize, you'll find that Alfred Nobel produced weapons and explosives.
WIki Article: Nobel Prize

You are correct. That's how he made all the money he is giving away.

There are many categories available for awards -
Nobel Prize in Physics – Awarded by the Royal Swedish Academy of Sciences
Nobel Prize in Chemistry – Awarded by the Royal Swedish Academy of Sciences
Nobel Prize in Physiology or Medicine – Awarded by the Karolinska Institute
Nobel Prize in Literature – Awarded by the Swedish Academy
Nobel Prize in Peace – Awarded by the Norwegian Nobel Committee
Nobel Prize in Economics

deTerminator is about blowing things up in the handicapping community.
Definitely not for "Peace".

I believe the proper category for your great advances in the discovery of new
munitions (deTerminator) would be the Nobel Prize in Physics. You definitely
deserve a prize; you just had the wrong category (was my point).

Hope you make as much on this project as old Alfred did on his!

You definitely deserve a prize; you just had the wrong category (was my point).

LOL - Perfect.


Wait until you see how much more there is... some is already in DET and more that I have to share.
Just the tip of the iceberg. I'll try to find time for some screenshots soon.

Meanwhile back to my single bug and the rest.