/poker/

On an average day how many tables do you play at once and for how long?

There is no upper limit really, just the games get harder and harder, I play 10nl (the blinds are 5c and 10c) about 4000 hands a day and make 300-500 a month. You can concievably play 8-10k hands a day and average about 650 a month maybe.. when you get to 25nl plus you're talking at least a k a month for any decent volume 5bb/100+. Where I live though 300-500 is plenty of money.

4000-5000 hands usually takes me about 7 hours. Play 4 zoom/fast tables during peak times and table select about 8-10 tables outside (its best to tile cash games imo). Table selecting can be a bit of a pain tho, way fewer blatant fish at 10nl than there used to be.

thinking of making a bot, when I'm finished with my trading algorithm.
Any good books for advanced poker tactics?

what if the bot loses all your money

how can you even make a good enough bot

hey scumbag, hope I sit in some of your games when your bot is finished :)
bots get caught all the time, most bots are in hu, and if you are competent are sometimes easier to beat than the worst fish cos they're so predictable. bots can't really win, any consistent use of GTO lines will flag them up, any deviation that is pre-programmed should be exploitable.
Bots do concern me a little bit but not too much.

well the easiest way is just using the monte carlo method, simulating a few thousand games to determine what's the right move. Then I would list the moves of the other players, see if they are aggressive or passive. Maybe apply some machine learning before I incorporate them into the monte carlo simulation.

I will ofcourse have failsafe's for when my bot will make errors, and stop-loss.

If my bot doesn't win I will team up with a poker expert.

A lot of time will probably go in recognizing the cards/game, and making the bot play like a human. This is my specialty though.

You won't make a profitable poker bot, I can just tell.

Then you got nothing to worry about.

But yea you are right, my field of expertise is currently in computer-vision, so I will just focus on making a bot that doesn't get caught.

Making a good algorithm will come later, I also dl pokerstars and saw you can just sit in on tables (?). This way I can make whole databases on specific player behavior. I'm getting pretty hyped desu, good thread OP!

how can i learn to do thesame thing as you

This seems better than playing poker manually

in about a year or so OCR/ hooking will basically be impossible on most poker sites, and yeah you can sit tables but you can't see mucked cards so it's bordering on useless. When you get caught - which you will, if you make any real money, you will get fined to shit, and good luck signing up without passing KYC measures :)

you need to have programming skills, experience in computer vision, know some basic statistics, and like poker. It also helps if you're interested in AI.

I'm a mechatronics student, and I did a lot of projects using vision for recognizing real products.

>OCR will basically be impossible
if the characters don't look like the old Veeky Forums captcha I think I will be fine.

>KYC
I'm not gonna play stupid you know, just 1 'person', multple tables, normal hours, what is normal for a human. Then if I make money I will just sell my bot, or just keep it as a 1 'person'.

> you can't see mucked cards
but I can see folds, raises, hands when they win. There's a lot of information in that.

>you need to have programming skills, experience in computer vision, know some basic statistics, and like poker. It also helps if you're interested in AI.

yes but how can i learn

I'm a NEET with no skills. i've dabbled in programming ocassionally but always gotten bored becus i never had anything meaningful to work on. I've always wanted to make something that either earns money or has some sort of value to other people. I want to know how i can get started making something useful

So you'll make a bot thatll require you to be at your computer (like I say OCR will be impossible, trust me on this) for a length of time matching what a human could put in, not able to play higher than 100nl without insta detection, and then you'll sell to someone who likely works for a site and then lose everything. Ok excellent plan.

I also lost motivation sometimes in the beginning. You just got to push through it, once you know the basics of one big language (python, c/c++, java) you know them all.

Basic experience is necessary because otherwise you won't know where to begin with big tasks where you can make money from. (i.e. poker bot)
Just read a book, follow a tutorial, do some tests. (make a calculator, sudoku puzzle solver, make a database)
I would recommend Python.

> require you to be at your computer
nope, it will pretend to be human (it plays at the times a normal human plays)

> OCR will be impossible
ok, I'll bite, why?

> higher than 100nl
if this is true then I will be content with what I get, but again I think you're overestimating the capability of their AI recognition system.

> you'll sell to someone who likely works for a site and then lose everything
I do not personally handle the distribution of my software

Not him, but any recommended books for learning python?

I'll save you some time. Don't do montecarlo, google cfr+. To save even more time look through the poker-ai.org forum. If you want to save even more time I can tell you straight away, invest your time into something else, the good days of poker are over.

Some general advice. Dont' have a big ego. Move up in stakes aggressively. The rake playing NL10 is eating a ton of your profits. What what costs you the most money is not having the bankroll to play good games on higher stakes. Game select hard. It's better to play one table with a fish than 3 tables with regs only. Use solvers to find near gto strategies but also study exploitive play and always exploit any leaks your opponents have. Also look for population leaks on these lower stakes and exploit them vs unknown players (as opposed to playing gto)

I quit poker about a year ago after playing full time for 5 years.

There are tons of anti-ocr measures being developed. Countering poker bots will always develop methods at a faster rate than the creation of poker bots. Anyway I'll leave you be.

To actual poker players, live or online, what are you playing? How's it going?

The bot detection is surprisingly good on pokerstars (and shit on all other sites).

I fucked with C and then went into C++ and then tried C#.

The issue is that I could NEVER make the change from "Making stupid bullshit" to "Making something actually useful to someone and that's productive".

I don't know how to do it. How the fuck do you hop from "useless shit" TO "becoming a professional"?????

>enter room
>go all in
>every body folds
>collect entry fee money or whatever it's called
>leave room
>enter another
>repeat

>walk into KK+
>cry

Terrible idea. If you want to gamble with just a slight negative EV use holdem resources to look up optimal short stack push ranges and play the push or fold games on 888. But probably better of playing roulette or BJ

Learning python, learn python the hard way, python crash course.
Python is also very usefull for finance, if you're interrested you can follow up to date tutorials on youtube. A guy named sentdex has really good ones.

Well then why won't you try to make something, like a poker bot. Just start slowly, make a plan, divide your program in pieces, do research, draw flowcharts, start coding.
Start with a poker helper, just a script you can manually input all the known cards on the table. Then it will calculate your chances of winning. If that's done you can see what algorithms are available on the internet, and try to implement them in you're code.
Even if you fail, or your program isn't 'professional'. You will learn a lot and get a step closer to becoming a pro.

I don't consider myself a professional computer software developer, I never made an application that would be considered as professional software. They are functional, but I couldn't just sell them. I'm far away from being a full stack developer.

Point is, that there are 100 lvls between useless to pro, I'm maybe 5 or 10, and idc. The only way you're gonna reach lvl 100 is just grinding through the tasks that are ahead. The harder the challenge the more xp you get, even if you fail you still get a part of the xp and you can try again.

thanks bro ill do it

Did you ever think that you were not to first one to thought about writing a bot, right? There is already tons of them and tons of them gets banned daily and lose the bankroll in that account.

>fish sits on table
>sitting with not full buyin, random 88bb stack
>pays bb from EP
>Goes allin fucking first hand he gets.
>assuming noone gets the fucking obvious signals and won't adjust to call looser
>still runs into QQ+ AK %2.6 times
>loses 0.83bb every time he makes that move
>makes a post on forum "online poker is rigged"

Thanks to you ecosystem exists

Best to leave him be, he obviously thinks he knows best, either way his money will end up in our pockets.

do you guys know any of these measures being taken against bots? Like what their methods are to determine if a player is a bot.

Errorless play
Closely follow gto lines or deviate from gto lines at set times
Stats similar to other bot players
Following specific patterns against certain players
Check for hooking or ocr acting on the client
Check for non human mouse movement
See what software is running

Bots can't really win. Whilst it's true stars are the only ones who really are vigilant with bots, more sites are picking up and also receiving help from 3rd party analysts.

Please don't be so naive . I am grinding 100-200nl on stars and 200-400 on party,888 and can say with confidence that there are bots on all the above rooms . by far more are in party then on 888 then on stars . many players from Russia/Belarus/Ukraine are bots , I could even bet that >50% of the winners on stakes over 100nl are bots. There are certain stats that make it very clear that a player is a bot and even stars doesn't make any thing about it ,a specific player had to be reported more than 20 times from different people just to be checked if he is a not. Also don't forget the Russian bot ring that was caught but had made over 600k on 100-200nl zoom in 2013-2014

Yeah I'll bet it's under 5% though I don't play above 25nl, can't imagine a bot beating higher non-hu.

Really shouldn't have started this thread with >bots desu..

I make $500 in a short session at the local casino,

Post a graph or some stats/hands to get the conversation on rails

Can you share example bot stats?