Machine Learning: Pool simulations

justnum

justnum

Billiards Improvement Research Projects Associate
Silver Member

The only post on youtube I found which is closely related, feels more like trial and error than concise computer programming.

Developing a computer opponent for straight pool is my latest project.

If I get enough support, I will post a video of me riding the space elevator while playing pool. Teaser: You get to see the shift in gravity while I play pool on the space elevator.

Your support is needed in coming up with the best one liners or tips to play great 14.1.
 
  • Like
Reactions: Pin
Patrick Johnson said:
New to this planet?

pj
chgo

Virtual Pool 4 - Celeris Inc

Celeris award winning games & simulations developer. Virtual Pool 4 Home.
www.celeris.com
Click to expand...
I couldn't find any videos of that platform with the computer playing 14.1

1626806830314.png

A while back I did purchase a game with 14.1 and a 14.1 opponent. This game did not have a good computer opponent for the 14.1 matches.

Thanks for the recommendation.
 
justnum said:
The only post on youtube I found which is closely related, feels more like trial and error than concise computer programming.
Click to expand...
I mean conventional machine learning is trial and error, really, isn't it?

But I think it would make sense to come up with parts of the program yourself: a way to quantify shot difficulty, a semi-randomized ability/missing control, probability threshold for playing safe.

So perhaps the machine learning element would be what safe to play, and shot sequence. If you wanted the program to feel 'natural', even shot sequence might be better designed to mimic real players' patterns rather than a probability pattern.

And if you wanted to incorporate it into a video game, would an algorithm developed by machine learning be small enough? I've no idea.

And if it was just for a video game, would the player really care enough for the whole thing to be worthwhile? I mean they're essentially just waiting until it's their turn to play again. And they'd probably play 8 or 9 ball, not 14.1...
 
Thinking about it, if you can code an algorithm to pick a shot sequence in 8-ball, you'd expect a lot of overlap with the algorithm for shot sequence in 14.1. So maybe the two should be dealt with together.
 
Pin said:
I mean conventional machine learning is trial and error, really, isn't it?

But I think it would make sense to come up with parts of the program yourself: a way to quantify shot difficulty, a semi-randomized ability/missing control, probability threshold for playing safe.

So perhaps the machine learning element would be what safe to play, and shot sequence. If you wanted the program to feel 'natural', even shot sequence might be better designed to mimic real players' patterns rather than a probability pattern.

And if you wanted to incorporate it into a video game, would an algorithm developed by machine learning be small enough? I've no idea.

And if it was just for a video game, would the player really care enough for the whole thing to be worthwhile? I mean they're essentially just waiting until it's their turn to play again. And they'd probably play 8 or 9 ball, not 14.1...
Click to expand...

Its a programming challenge that is well known but few have attempted to work on it.

Chess got huge investment and exposure with Deep Blue.

The online market can grow because of advanced computer AIs. Imagine practicing and getting coaching at the same time from the computer.

It will be a pool player dashboard. The next upgrade is to enhance it with phone support to calculate best shot options on table.
 
I was helping some girls at the local bar last night that were worse than the AI program !! really !
 
I actually thought the video was interesting. A lot different than virtual pool, which I'm assuming had many, many, pre-programmed parameters to find "good shots", and many of the shots are still ridiculous. The AI in the video I think was much more "free thinking" (and resulted in even more ridiculous shots than VP).

During the first few seconds of watching the video, I wasn't sure if the OP or someone else made it. Once it got to the pool table graphic, I was sure it was made by someone else;)
 
Building a billiard community for programmers can help recruit and rebuild the college campus pool vibe.

Imagine a bunch of student engineers trying to contribute to a great billiards open source project.

There is not enough interest for college pool tournaments, how about college pool tournaments for AIs to compete against each other.

The best learning AI.

They have AI competitions in other esports.
 
What does a computer need to learn about pool? It isn't that hard, balls in holes.

As long as the aim is right, the computer is going to make the ball. There is nothing to learn.

Execution matters in the real pool game. Perils abound. Not safe for all. That's what makes it thrilling.
 
justnum said:
Building a billiard community for programmers can help recruit and rebuild the college campus pool vibe.

Imagine a bunch of student engineers trying to contribute to a great billiards open source project.
Click to expand...
I see, you are an idea man! You don’t want to actually do the work. Throw shizz at the wall and hope something sticks; someone else will follow through on your ideas. You want to inspire others through your passion for this beautiful game. Help them see it’s more than a mere way to pass the time at the bar or gamble with lesser skilled players.

Hey Bub, nobody is inspired by someone who isn’t willing to work for it themself. You lead, you inspire by example.

Someday you’ll actually follow through on one of your ideas. In the words of John Fogerty, “Someday never comes.” I hope that’s not the case. I believe in you. Grow up and get after it!
 
VarmintKong said:
I see, you are an idea man! You don’t want to actually do the work. Throw shizz at the wall and hope something sticks; someone else will follow through on your ideas. You want to inspire others through your passion for this beautiful game. Help them see it’s more than a mere way to pass the time at the bar or gamble with lesser skilled players.

Hey Bub, nobody is inspired by someone who isn’t willing to work for it themself. You lead, you inspire by example.

Someday you’ll actually follow through on one of your ideas. In the words of John Fogerty, “Someday never comes.” I hope that’s not the case. I believe in you. Grow up and get after it!
Click to expand...
Let’s just get him on that elevator to space. Dude is a troll, arguably the best az has seen.
 
You must log in or register to reply here.
Back
Top