... But my question is, how, even theoretically, we can use AI for the games of chance? I mean, it can tell you what part of your bankroll to stake at once to, theoretically, last longer, but in reality your very first bet can be that with the highest multiplier possible, and no AI can tell you for sure that it won't happen.
It won't be very useful for games with mathematically determined odds, where the house has an edge. But (when it's able to collect data from external sources) it could be good for things like looking for positive EV opportunities on numerous different gambling sites (i.e. with high enough Jackpot and low number of players, the mathematical edge can sway on your side).
Yes, it could be, but in reality, if using an AI for that you would likely to end up on a scam site with very favorable odds, but no intention to pay you out ever.
Another problem with positive EV games is that you need to make many millions of bets to really take advantage of that.
But it'd probably most useful for gambling with "judgmental" odds, i.e. sports betting, where it could be looking for arbitrage opportunities (but I think we already have that for quite a while) or it could run an advanced analysis on teams performance, past results, patterns etc and could help you find good value bets (i.e. with pay-out rates higher than they should be). Then it could help you decide on the size of the wager for best long-term effects.
The problem is, if such tools will be available to players, bookies will also be using them in order to not get screwed.
The problem is, if such tools will be available to players, bookies will also be using them in order to not get screwed.
Yeah, bookies and other gamblers too. And this will change the odds in such a way that you will need to risk $100 in order to potentially win $5 with a 94% win chance. Sounds familiar? Yeah, those are the odds for a dice game.
