Poker-AI.org

Re: PokerStars Bot or reinventing the wheel

2023-03-31T06:36:31+00:00

There was quite a lot of discussion about Stars detection capabilities on this forum years back, probably on the private subforums which weren't archived. It was fairly clear that they could detect anything on the machine connected to the casino and suspicious mouse movements. It is not all that hard to detect bots and the only reason other casinos don't do it is because they are fighting for market share and are prepared to tolerate bots to get more players.

Statistics: Posted by spears — Fri Mar 31, 2023 6:36 am

Re: PokerStars Bot or reinventing the wheel

2023-03-30T08:53:58+00:00

How could you know that pokerstars could detect it? Do you know if someone tested this and verified?

Statistics: Posted by nbiresev — Thu Mar 30, 2023 8:53 am

Re: PokerStars Bot or reinventing the wheel

2023-03-28T11:51:08+00:00

Quote:

I have done human mouse play back it looks quite realistic.

Make sure the machine generating the mouse actions is under realistic load.

nbiresev wrote:

Do you know if poker site could detect that the mouse move and click come from software API and not from hardware by moving mouse?
Same for keyboard.

If the automated mouse and keyboard actions are realistic and come in from another machine via usb (or serial) they can't tell. Otherwise they can in principle. Different sites have different abilities. Stars can and will detect everything.

Statistics: Posted by spears — Tue Mar 28, 2023 11:51 am

Re: PokerStars Bot or reinventing the wheel

2023-03-24T18:23:01+00:00

Thank you for the answer. I have done human mouse play back it looks quite realistic.
Do you know if poker site could detect that the mouse move and click come from software API and not from hardware by moving mouse?
Same for keyboard.

Statistics: Posted by nbiresev — Fri Mar 24, 2023 6:23 pm

Re: PokerStars Bot or reinventing the wheel

2023-03-21T11:13:26+00:00

Ideally the rotating scaling etc should be kept small because the mouse trajectory is statistically different for different directions and distances.

I'm not at liberty to tell you why you should test it, but believe me, when you do you are very likely to find something you have not anticipated. Plot out the generated mouse position every few hundred milliseconds and visually compare the result with recorded real ones.

Statistics: Posted by spears — Tue Mar 21, 2023 11:13 am

Re: PokerStars Bot or reinventing the wheel

2020-04-02T18:25:06+00:00

spears wrote:

baddy wrote:

Guys,

Thanks for useful replies. However, I have last, but not least problem. Is it possible to emulate human input? Scaling and automation aren't possible without that.

It's possible to move the cursor in a human like way. Be sure to test this properly because there are pitfalls.

What kind of pitfalls you mean?
Would it be enough to record human mouse movements, and then if you want to move mouse between A and B, pick random path, and rotate, translate, scale points to fit between A and B?

Statistics: Posted by nbiresev — Thu Apr 02, 2020 6:25 pm

Re: PokerStars Bot or reinventing the wheel

2020-04-02T06:16:15+00:00

baddy wrote:

Guys,

Thanks for useful replies. However, I have last, but not least problem. Is it possible to emulate human input? Scaling and automation aren't possible without that.

It's possible to move the cursor in a human like way. Be sure to test this properly because there are pitfalls.

Statistics: Posted by spears — Thu Apr 02, 2020 6:16 am

Re: PokerStars Bot or reinventing the wheel

2020-04-01T16:46:03+00:00

baddy wrote:

Guys,

Thanks for useful replies. However, I have last, but not least problem. Is it possible to emulate human input? Scaling and automation aren't possible without that.

This is the whole point of creating a poker strategy.

Statistics: Posted by chadthelad — Wed Apr 01, 2020 4:46 pm

Re: PokerStars Bot or reinventing the wheel

2020-04-01T16:35:03+00:00

Guys,

Thanks for useful replies. However, I have last, but not least problem. Is it possible to emulate human input? Scaling and automation aren't possible without that.

Statistics: Posted by baddy — Wed Apr 01, 2020 4:35 pm

Re: PokerStars Bot or reinventing the wheel

2020-03-31T07:52:09+00:00

Poker stars is the holy grail. Maybe start to develop what you mentioned above, but don't exclude trying it on other sites.

If you do get things going for you I'd be happy to be involved and maybe share thoughts.

Statistics: Posted by chadthelad — Tue Mar 31, 2020 7:52 am

Re: PokerStars Bot or reinventing the wheel

2020-03-31T05:57:27+00:00

- Don't start with PokerStars. They have the best bot detection.
- Specialise in the AI or the Casino Interface, at least to start. Hopefully somebody will rent you an AI or sell you a Casino Interface eg
viewtopic.php?f=33&t=3299
- I started with a program like https://www.poker-genius.com/ If you want to use that check the API actually works. There is also something called pokertestbed.
- Spend a few days reading this forum. There is some good content but it is very spread out.

Statistics: Posted by spears — Tue Mar 31, 2020 5:57 am

PokerStars Bot or reinventing the wheel

2020-03-30T21:16:24+00:00

Hi everyone!

First, if you have any thoughts, ideas and suggestions, I'm open for any team work.

Several years ago I had the idea to implement PokerStars Bot. I have quite good experience in software development. Now it's clear to me that I'm ready. I'd like to present my vision of Bot and highlight the most narrow places where PokerStars (and maybe others) can catch it.

Let me start from business goals.
1. Undetectable by Room.
2. Fully automated. Support of scheduling work (MTT or playing cash, I mean human like).
3. Simple setup, not for sale, but for scaling.

My ideas for implementation.
1. Human like behavior.
1.1. Undetectable reading information from screen (VM is not an option, maybe KVM, maybe video driver, it's problem for investigation).
1.2. Recognize situation on tables including table and event name, bet sizes and positions, player names (as far as I researched it's obvious nowadays).
1.3. Making decision service (I'm planning to implement like separate "online" service. It will be mix of machine learning and charts).
1.4. Low level mouse driver implementation with human emulation (another yet problem, the idea I have is to save all my track moves and randomly use them).
2. Full automation(I suppose it's obvious, in case all implemented above).
2.1. Run client by schedule. Should have significant events schedule in DB.
2.2. Open scheduled event table or cash tables. Using previous recognition and inputs, of course.
2.3. Monitoring balance. Bot should stop work by balance limitations.
2.4. Close in human like manner.
3. Simple setup. I mean that it should be thick client that is on the same PC with Poker Room client. Decision service and maybe recognition module are elsewhere.

Maybe, I didn't take into account some security issues like VPN or geo location, let me know, please.

Thanks in advance.

Statistics: Posted by baddy — Mon Mar 30, 2020 9:16 pm