Here comes the AI

[ReplayRandall]

Originally Posted by Dave Schwartz

Thask,

Do they have restrictions on states to play from?

Recently I had both of my accounts closed because I live in Nevada.

The alternative required me to open a Nevada account and use an absolutely horrible mobile app to wager with.

Residents of the states of Hawaii, Montana, Nevada, and Washington are not eligible to withdraw funds from account balances. However, those players may participate in contests for the purposes of pursuing entry to on-site events through qualifying contests (such as the NHC and HPWS).

[romankoz]

Originally Posted by Dave Schwartz

It isn't past posting.
It's simply the lag time between when last minute bets are made and the 20 to 30 seconds that it usually takes for the data to come back to the track.

Roman Koz replies: Fair dinkum- 20 to 30 seconds!!! Gee, our time lag would be no more than the time someone sits somewhere and sees the gates open and bang the button is pressed to stop betting. I would be surprised if it was 2 (yes, TWO,) seconds and badluck to any whale here. Now I see what your problem is when you replied replied re my 9/2 to 7/1 posting. I would never ever bet in your situation: it reeks to be honest.

It is made worse by the fact that such a large part of the handle is coming from about a half dozen bettors, all of whom use similar statistical techniques. Logically, they will agree often.

RK replies: Yep, would agree: makes sense.

Worst of all is that they are the sharpest players on the planet.



I would agree with you.

If only AUS there was good digital information at an affordable price.

To get the kind of data an every day player in the US can easily get for $100-$150 per month would cost 20 times that in AUS.

Roman Koz replies:
. I would be very surprised at 20 times: depends on the depth of what you want/need.

[Dave Schwartz]

Originally Posted by romankoz

Roman Koz replies:
. I would be very surprised at 20 times: depends on the depth of what you want/need.

Last 10 races, with all the times, beaten lengths, jockeys, trainers, workouts, earnings box, equipment, all the class info etc.

The feed I was getting while working on a project was literally 20x the cost here.

If you can point me to a comma-delimited data source for all the tracks with all of that data for far less, I'd be very grateful.

Also, I need at least 4 years of back data.

[traveler]

used IBM Watson a few years back and it's predictive analytics and MC Sim features. Was interesting but horse racing is down the list of what I want to spend time on.

[upthecreek]

On At the Races discussing article in Financial Times about AI
Starts around 24 minute mark
https://stevebyk.com/broadcast/tuesd...-pat-cummings/

[headhawg]

The Beyer interview was ok. I thought the interview with Maury Wolff was a bit more informative but nothing earth-shattering. I was lucky enough to have a lengthy discussion with Dave Schwartz on Tuesday. (Truth be told I was mostly listening. lol) As a novice to machine learning he gave a great explanation of neural networks and the reasons that machine learning has moved beyond that.

If this kind of stuff interests you there is a video that Dave shared about how AI was used to redefine positions in basketball based on skill sets rather than traditional position labels. It helps visualize how data is shaped and the connections between the shapes. It's a good introduction.

New Positions of Basketball

[Half Smoke]

.


this is just anecdotal - not a thorough study at all - but anyway

twinspires posts their AI pick with the supposed % chance of winning

I've looked at that quite a few times -

much of the time it looked like a horrible, horrible pick

I wouldn't touch it with a 10 foot pole_________________(-:/




another anecdote


- I have noticed a few times that a horse got bet way down very late that looked bad on paper

of course, I didn't bet him

and the few times I noticed this he won and won easily a couple of times

it appears that some people knew something that I couldn't know just from the PPs


.

[ubercapper]

Originally Posted by ReplayRandall

Residents of the states of Hawaii, Montana, Nevada, and Washington are not eligible to withdraw funds from account balances. However, those players may participate in contests for the purposes of pursuing entry to on-site events through qualifying contests (such as the NHC and HPWS).

Washington State residents cannot play in any contest where a portion of the entry fee goes to prize money. Residents can only play in free contests and contests where the entire entry goes to bankroll.

[Dave Schwartz]

Technically, Machine Learning and AI are not the same thing.

Here's what ChatGPT said about it.

Artificial Intelligence (AI) is the broader field of creating intelligent machines that can perform human-like tasks. Machine Learning (ML) is a subset of AI that involves training algorithms to learn from data inputs and make predictions or decisions without being explicitly programmed. In other words, ML is a method used within AI to teach computers to learn and improve from experience.

This is accurate.

The nature of Machine Learning is to be more linear than an AI engine.

The whales use Machine Learning.


Google said this:

While artificial intelligence encompasses the idea of a machine that can mimic human intelligence, machine learning does not. Machine learning aims to teach a machine how to perform a specific task and provide accurate results by identifying patterns.

Artificial Intelligence (AI) vs. Machine Learning

Columbia University:

Artificial intelligence (AI) and machine learning are often used interchangeably, but machine learning is a subset of the broader category of AI.

[PIC6SIX]

So well said, Dave, so much so that I took a picture and will keep it on my phone for future interaction with numskulls.

[USMCGRUNT]

So, with all this AI talk which one is best that will make us win every race? Thanks in advance…

[Dave Schwartz]

Originally Posted by USMCGRUNT

So, with all this AI talk which one is best that will make us win every race? Thanks in advance…

None of them.

Working on it.

[USMCGRUNT]

Originally Posted by Dave Schwartz

None of them.

Working on it.

Perfect! I have about 30 good years left. I will be following, thank you!...

[andicap]

Originally Posted by thaskalos

I think you should be the one doing the reconsidering.

Talk to any established bookmaker and he’ll tell you that the identity of the bettor is more respected than the size of the wager. Mattress Mack can bet $2 million on a sporting event without any problem…but Alan Boston can’t even bet a tiny fraction of that amount.

That is one of the smartest things I have ever read on this board.
It explains why books will not move lines even when the preponderance of money is on Team A while well known sharps are on Team B.

In racing, knowing that stables are betting is irrelevant. What is critical is identifying the stables that win their bets.

[Robert Fischer]

Originally Posted by headhawg

The Beyer interview was ok. I thought the interview with Maury Wolff was a bit more informative but nothing earth-shattering. I was lucky enough to have a lengthy discussion with Dave Schwartz on Tuesday. (Truth be told I was mostly listening. lol) As a novice to machine learning he gave a great explanation of neural networks and the reasons that machine learning has moved beyond that.

If this kind of stuff interests you there is a video that Dave shared about how AI was used to redefine positions in basketball based on skill sets rather than traditional position labels. It helps visualize how data is shaped and the connections between the shapes. It's a good introduction.

New Positions of Basketball

I like that stuff. I have biases and jealousies regarding basketball stuff, but I like it.
If nothing else, it's better than the status quo. It also offers some new thoughts to build on.

The horseracing book that I started to write is vaguely similar to map and node style using various handicapping factors.
I can't see it being much of a success, and I wouldn't want to make the game worse if it was a big success.