Poker-AI.org

Ok, that day has come, you create winning bot. How much do you expect to earn in a month?

Same guy?

You may want to specify over what period of time.

Please, go read the first post.

With all the cold in Antarctica, I bet you could get a PC running very fast. Maybe that's how he does it?

so you don't earn 100k+ already ?

Should've added that "in a month" into poll title, it's very misleading otherwise

added

Wanna see that guy that does 20k+ avg. a month

Edit: For guys like me you should add negative numbers

This poll is about winning bot. I would even say that this is a hypothetical question.

You can create your own poll about lossing bot.

This depends on a lot of factors.
If we say mid/low-stakes, per account that multi-tables around 6 tables on average and plays roughly 8 hours a day I would say it's about 1k per month. That's a rough guesstimate.

If you want to back this up by math you can just calculate it:
Let's say you're winning at 1bb/100, NL100, that's $1 per 100 hands.
Roughly you'll get about 90 hands per hour I think in 9max but that obviously is higher in 6max and HU, for ease of calculation let's assume we get 100/hour, which means $1/hour.
Now we 6 table for 8 hours a day, 20 days a month, that's: 6*8*20*$1 = $960/month/account.

Looks like I was about right with my guess. Those are pretty reasonable numbers to achieve with a bot I think. Obviously if you play around with the numbers this scales pretty high pretty fast but you shouldn't kid yourself about your win-rate and the playing schedule. The higher you move up, the harder it's gonna be to 6 table consistently e.g.

_________________
Cheers.

You play about 60-70 hands per hour at 9-max.

Realistic estimate: calculate the estimate as honestly and precisely as you can, then divide that by 2.

Whoops. You are right, I haven't played online in too long.

_________________
Cheers.

1-3k

You can always run your bot on multiple sites, with multiple instances of your bot there should be a chance to make some decent money.

Achieving >1k/month off of one account on one site is really hard IMO

_________________
Cheers.

Why is it so hard? Shouldnt a good HU bot be able to make similar money to a human regular? Of course a rule based ringgame bot would be a whole different story.

I don't think you can build a bot that plays HU as profitably as a human would but I might be wrong. By profitably I mean adjusts well enough against bad opponents and still adjusts good enough not to expose leaks to good opponents.

I was mostly basing that statement on my experience with SNGs. What kind of math would you expect in HU? Limit, winrate, tables, hours/day etc.

_________________
Cheers.

I think you should be able to develop an FL bot, with a nash strategy that is on par with polaris NEQ strategy and you could add a simple exploit model for really bad opponents.

It should be able to run with at least 0BB/100 post rake.

You would play 8 hours a day, and should be able to play at least 1k hands a day on average, limirs would be 2/4-5/T, whi results in an average BB of about 6$, assuming you will get more opponents on lower limits.

FL has rather high rake, so you would get at least 10$ RB per 100 hands, giving you about 3k of RB per month.

Or 33$ SNG, 5 tabling on average, 10 mins per SNG -> 90$ rake per hour, lets say 50$ RB/hour

5 hours a day, 30 days makes about 7.5k, lets say you make only half of that , its still above 3k.

Seeing how bad I am at SNGs, a bot should be able to have a ROI of about 0%, but especially the FL bot looks realistic to me.

Are you familiar with the realities of fixed limit heads up games? I'm not, but a quick glance into poker stars lobby showed that, for those stakes, there's literally one game going on at the moment, while there are another 20-30 regs waiting for fishes (it's still an early hour at the moment, bot OTOH, it's Sunday, so fishes have time to play). Assuming the bot breaks even against the regs, making any reasonable profit of those super scarce amount of fishes seems extremely dubious.