To prove that you really have knowledge of NL cashgame bots.
I seriously doubt you could show more than a break even N5/NL10 SS bot.

As Ongame was mentioned, I checked my NL50 Ongame database. I have 120 players with more than 10k hands and found only 2x2 deep stacked guys with identical stats. The "identical players" have never played against each other, probably the same bonushunter played on 2 skins. Maybe Ongame NL50 isnt the target of bot users. By the way the guy who wrote the working FR SSNL bots for the disaster like new Ongame software can be a super talented software developer, Ongame should hire him ASAP

If you play on smaller sites, colluders are far bigger problem than bots. When I played on Ongame the weak NL50 SS guys never caused any remarkable problem but I several times had to leave tables where 2-3 guys from the same country played as a team.

Yes, your post was a massive prove of the opposite. Thank you. So many people on 2+2 are accusing me of botting - now I can just refer them to your post.

I ain't gonna discuss my play online. If you want to verify my knowledge in anything, ask some technical questions. I wouldn't guarantee you a reply, but I would give it a thought.

Here's what I think it comes down to.

Poker is a negative sum game. The rake creates a head wind that anyone playing poker must overcome to win. Since over the long term everyone gets the same cards, making fewer mistakes than your opponents is the only way to win.

A well programmed bot by definition will make fewer mistakes than a human. Therefore, a bot programmed by a winning player at X level will do better than the winning player at X level, who will make mistakes because they are human. Two bots of equal skill playing HU will lose money because neither can sustain an advantage over the other, thus just losing rake. As more people buy Barry Greenstein or Doyle Brunson in a can, the edge one program would have over another would disappear completely.

Therefore, the only situation where a bot is profitable is where it is not up against other bots. Someone advocating free use of bots in games is either naive or disingenuous. I think it is clear which side indianaV8 is on.

Why? A bot is only as good as it is programmed to be. If it's badly programmed it will make mistakes. You say "well" programed but what that means? Do you mean a "winning bot" here, or otherwise don't understand your point with that sentence.

Normally, if bot and human play the same core strategy, human will have massive advantage over the bot, in the way he can adapt and exploit other players, spot patterns and so on - in a short term session. Bots might improve in that in the future, but they are not there now.
Normally, bots will compensate their disadvantage to humans, by much more in-depth analysis of the game. Anyway, again I don't know if you have particular point here, so let's move to your conclusion.

True.

There are many people that will build bot for the challenge of it, beating the other bots for money. If there is healthy ecosystem of bots you can make enough money against people tryint to build a better bot. You are basically trying to tell that all bots are "pros" and they, as any other player, need fish to be profitable. Which assumption is not true.

Finally, I still don't get it why would you opt to get personal. I am adovcating bots, being marked as such, on special tables, e.g. on PokerStars. People would be able to choose if they want to play against them. If they don't choose to play - they will not. If noone choose to play agaisnt them, we would be in the situation that we are now, so nothing would be changed. So what is your problem such setup? Why would you oppose it with arguments, if, in case you are right and this never flies off, we would be back to the current situatuion we are in?

Back in April/May when I got suspicious I got some data mined hands and went about an in depth investigation. To give you an idea of why I had suspicions:



Also note
- These guys had played against each other roughly 20 hands each over this sample
- They seemed to play in pairs, you would never see all 6 at the lobby, only two at a time.
- Two of them were from Harrow (England I assume) and two from London, a fifth was from "Arrow" and the last I can't remember - its somewhere in my notes.
- They had a high F3Bet so I tried out 3-betting them every single time (no joke) they did not adjust.
- Every other stat you could find that had a meaningful sample followed the same pattern.

The "case for the defence" would be that looking at their postflop hands I could see some randomness in their play, they were very very very aggressive and not incapable of three barrelling air. Also if I recall correctly I hadn't found a pattern or anything out of ordinary in their session times.

As I was doing the investigation all five vanished overnight (I had informed no-one of my investigation), I cant remember the exact date but it was before the ongame P5 update.

So because I'm not 100% sure and they seem to have been dealt with (no one with similar stats has appeared) I figured it wasn't worth pursuing. Felt it was relevant and interesting enough to show here in context.

SS that sounds like 90% of the shorties I've seen.

Forgot to state, these guys played 100bb deep 100% of the time.

I don't believe this is a valid conclusion. It depends on the skill of the programmer as a programmer, and their skill as a poker player. Any outcome is possible.

I'm pretty sure that in the context of this "debate" we can assume that a well programmed bot is going to be one that has been programmed with a winning style of poker.
It's pretty obvious that if my mom was a master programmer that doesn't mean that she will be a master Poker programmer. you have to be able to beat checkers before you can beat poker.

I love how some people try to use semantics to derail the argument from the main issue.

Indidana, I didnt speak about SNGs
Do you know undetected winning FR NL50+ bots playing on PS or FTP?
IPOKER is really full of SS NL bots?

Lol, xxsevendxx and xxeyeeyexx are in my NL50 database too, they are one of the pairs of identical players I found. They played 19/16.5 3Bet ~5.2 at NL50 and I have never thought a 19/17 player can be bot. xxcrashxx (18.5/16) also played NL50, other guys are unknown for me. Yep, this is quite strange. These 3 guys have the highest PFR% among the 120 regs in my database

If you were to build a bot, it would be best to have it play at a euro site. I honestly don't think they have the resources to look for bots like PS and FT do. It would be an interesting exercise though and I think it would massively help your game to think about it on that level.

The poster used that argument to justify his next statement; that a bot can only profit by playing human players. Since the original argument was not valid, it's not just semantics. (Although I have to say his second conclusion doesn't logically follow from the first anyway)

You can't conclude much about the ability of a bot by referring to the poker playing ability of its creator. It could be significantly better, or it could be significantly worse.

Indiana's idea (marked bots) is OK in principle. In practice, it means people get to develop bots on the "legal" tables without having to worry about being banned. Once they're developed, they can then start operating them on tables they should not or, worse still, start selling them for that purpose (not explicitly, of course, but everyone would know that's the intention)

+1. The problem with stealth bots is they will kill the golden goose. If some sites want to allow bots and others not, that is fine, but botting against the ToS is wrong. The sites have made bots against the ToS because they deem bots a negative factor for the long term health of online poker.

See the pokercast with Mason Malmuth (I think it was him) where he talked about computers essentially solving backgammon killing the game (from a wagering standpoint).

I assume everyone here, including most of the "I would bot" folks, would like to see the poker ecology prosper going forward rather than die.

Exactly. More to the point: should MLB allow the use of said bat? Of course not!

Sure, technology will advance but that does not mean we have to let people ride motorcycles in the 200m foot race event.

The question is not "will technology advance" but "for sites that want to prohibit botting, will they be able to detect bots at a high enough rate to keep botting from destroying the game and their business".

Interesting find. I wouldn't be entirely convinced from just pre-flop stats that this wasn't just a team of people using the same auto-folder software, which I assume could have been written to auto-3-bet as well.

Late to the party on this thread, my answer was no - but a few years ago I would have answered yes. I actually thought about this for a while (more about the programing challenge than the idea of unleashing my bot to go collect me $$$) and the logic involved in a FR bot just became too advanced for my (<skilled, >novice) programming ability. Very preliminary thought on trying to figure out some PF logic on position+cards+villian(s)+action started to melt my mind. I am sure it is possible, but only by smarter people than me. Making it fairly undetectable seems like it would be easier than making it profitable.

It looks like xxsevend85xx and xxcrashgardonxx played differently postflop. Maybe they are not the same bot/person or at least the bots worked with dissimilar postflop strategy. Sounded Simple can verify if they played differently at NL100/NL200.



By the way a preflop auto-raiser/folder program is a type of pokerbots. I think every software which makes preprogrammed decesions during the hand can be considered as pokerbot.

A pre-flop bot handing over to postflop does make the most sense I agree.

The main point being that they (or he/she) had a serious advantage in terms of being able to play more hands for longer with much less chance of misclick/tilt/timeout pre-flop.

This.

Perhaps everyone is better than me, but I can't pokerstove every person's range against every possible hand I play is the 3 seconds or so that I generally have to decide what my action should be pf. I can't then take that range, figure out what they should continue with, calculate what my equity is against it to 3 significant digits in the next 3 seconds using their FCbet stats among other things.

I can't in the CO merge to 3 significant digits the likelihood of seeing a call or raise from the BTN and blinds to decide whether to steal or not.

I can't have in my memory how strong each player's hand is on the river when they bet and exactly how often they are betting with air.

But a bot can. Bots always play their A game.

Sure a player can try to switch up their game to fool the bot. However, the theories of adjusting the gain on a controller has been around for a long time.

I agree that the success of a bot depends on the skill of the programmer and poker player. But they would have to suck pretty bad to build one that was exploitable to the extent that someone could make money off of them.

I'm actually going to bump this thread. I was asked this morning if I would work with a player, knowing full well that the information was going to into building a bot (I obviously declined). Got me thinking, and then I remembered this thread....

Rush Poker is a great environment to bot in.

I suspect that, especially at the micro micros, botting is more common than I would ever have guessed.