HN Explorer

Ask HN: Has anyone built a poker bot?

5 points by beginnercounter ↗ HN
I'm looking to build a poker bot and want advice from people that have actually attempted it. What was your general approach to building it? Did you see profits at a small scale / at low stakes? What prevented you from scaling it to many instances running on multiple sites?

8 comments

[ 2.9 ms ] story [ 24.6 ms ] thread
Nextgrid
If it was profitable, wouldn't the market equalize anyway as everyone would be doing it? Your bot thus must know some secret and be better than all other bots as you need to assume your opponents are bots too.
beginnercounter
If all the opponents were bots then I'd agree that it would be hard to make a profit. But have you tried playing micro stakes? There are some very bad human players making mistakes on a regular basis. Particularly on weekend nights. The bot just needs to break even against other bots and beat weak human players.
jstx1
Two factors that don't equalise the market

- there are still human players in the pool

- using a bot is against the rules so sites use bot detection, ID verification etc. to prevent it (some obviously do a better job than others)

jstx1
The obvious thing is that it's against the rules on every major site. You're basically saying "I'm looking to cheat, has anyone else done it too?".
beginnercounter
I'm looking for discussion around computer vision, AI etc. This comment isn't really adding much.
jstx1
I think that if you were really looking for that discussion, you wouldn't have written this:

> Did you see profits at a small scale / at low stakes? What prevented you from scaling it to many instances running on multiple sites?

Which implies deploying the bot. And doing that is cheating.

beginnercounter
The first thing I wrote was:

> want advice from people that have actually attempted it

It seems like you have no experience in this area. If your issue is the overall post, why don't you just report it instead of complaining to me in the comments?

burntoutfire
I had a profitable bot around 10 years ago. It made around $500-$1000 per account per month. It didn't have a future though, as it played a particular kind of Sit'n'Go tournament that was phased out shortly afterwards (I believe mostly due to collusion rings that were active in those kinds of tournaments).

There used to be a semi-thriving community of botters on pokerai.org forum (was even frequented by some of the celeb pros, like Patrick Antonius, who were also trying to do bots). The site is down now, but perhaps you can read through it on a wayback machine. One of the interesting conclusions from there was that all the sites except PokerStars didn't care about detecting bots and some were in fact offering special deals under the table to encourage botters to come to their site (because they generate a lot of revenue for the site). In my own (fairly short) botting career on two major sites my bots were never detected. What's more, I actually detected a ton of bugs in their software which I had to work around in my bot-to-casino interface code :)