Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
AI beats Human Poker players. Could this be a game-changer for automated trading?
Could this technology be a game-changer for automated trading? IMHO, it will be:
“In a landmark achievement for artificial intelligence, a poker bot developed by researchers in Canada and the Czech Republic has defeated several professional players in one-on-one games of no-limit Texas hold’em poker.
Perhaps most interestingly, the academics behind the work say their program overcame its human opponents by using an approximation approach that they compare to ‘gut feeling.’”
I am sure it is a game changer, for several reasons, not only the AI technology as such, but also the computer power that is ready available with a few GPUs in a PC that today all can be bought in a retail computer outlet.
I do not think this is anywhere near as significant as Deepmind's work in beating Go. Even the Go work was largely notable because it used general purpose methods. But, this is just my feeling not having looking into this area of research too deeply. But, it would be yet another step, in a series of recent steps, in demonstrating capability of computers to outperform humans.
I'm reasoning from the fact that all poker hands have known values. So, I don't think it would be too difficult (famous last words) to develop some heuristics or develop a custom program that could play quite competitively. However, if they have a general purpose solution that was self-learned then that would be a bigger deal.
Will it effect markets? Absolutely...it already has dominated markets. One need look no further than HFT's. Whether you are for or against them, they have fundamentally changed how markets function, and more importantly how much more risk now exists because of them.
In many respects we already have AI at work, albeit not quite at the level of the link above. But it is indeed in our markets, and it will eventually dominate them (or perhaps it does already?).
Comparing poker to the markets is a bit of a stretch.
Poker is a game with fixed rules and a relatively small number of possible outcomes based on a deck of cards.
The markets are dynamic ever changing and adapting.
The only thing poker and the markets have in common from a player/trader perspective is that the only way to win in the long run is to employ some sort of probabilistic strategy.
I would venture to say there are market algos around right now that make that poker playing bot look like child's play.
Ok - so I have a friend in Israel who is a renowned persona in the AI field.
I remember a few years back - they were using historical large data sets of poker games to create a "Decision Support" tool in the from of a "Bot".
They raised capital for this research and after many months of work - they could not find an"Edge", Mimicking human behavior is a tricky endeavor.
Also I remember that a few years back some hedge funds decided to fight HFT algo's by purposely creating a "delay loop" which lasted a few extra milliseconds that took these algoes out of balance.
This reminds me of a funny movie segment "Buster Buster technology" - I think this says it all:
HFT -> Edge is simple but requires proper infrastructure.
AI -> Edge is complex and more like discretionary trading and does not require infrastructure but may benefit from it.
I have hypothesized and expressed in the past that discretionary traders should have an advantage over most traditional systematic strategies in the area of robustness because complexity is generally required or associated with modeling complex phenomena. Most system traders do not try to capture that complexity because without advanced data science techniques and expert domain knowledge, extra complexity will result in fragility and over-fitting.
What AI allows for though is for complex strategies that can capture the type of robust, market cognition edges that traditional systems can't capture. If you can build working complex strategies then it makes it more difficult for the crowd to copy your techniques which also lends to robustness.