Friday, May 11, 2007

The 'Not-Bot' Controversy --- Script Kiddies in a 1/2 NL 'Sweatshop' at Full Tilt?

Here's one of those stories that was just bound to occur. A couple of days ago, an experienced 2+2 poster went public with his accusations about a supposed 'bot' ring that was flooding the $1/2 no-limit games at Full Tilt earlier this year. The evidence --- on its surface clearly indicating that something unusual was going on --- was the detailed posting of hundreds of thousands of hands played by the four suspect accounts. The accounts showed highly similar, very tight betting patterns over the span of something like 200,000 combined hands of play.

The original poster/accuser had made a private complaint about the perceived offending accounts several weeks earlier, which at that time, as events would later show, resulted in the freezing of nearly three dozen Full Tilt accounts for as much as a month while that site's internal investigation unfolded. This included many players who had done peer-to-peer transfers with the offending core accounts, and who, it seemed, were quite innocent in the matter. Full Tilt's lengthy investigation apparently turned up no proof of botting activity. The suspect accounts were later restored to active status, hence the public airing of the original complaint on 2+2.

And what a firestorm that posting produced. In two days, the thread has already become on of the largest in recent 2+2 history, garnering nearly 2,000 posts. Included are the public outing of the group involved and an intense debate over whether the stats as presented provide unambiguous proof of poker-bot usage, or, as the group's leader maintains, it's merely a highly structured operation adhering to very tight playing patterns.

The three-person group doing the 'not-bot' playing was pinned to an eastern Pennsylvania locale early in the debate; four accounts were mentioned as participants because one of the players was replaced part way through the play period under examination. One of the 2+2 moderators was quickly on the scene as well, claiming personal knowledge of the players in question and preaching loud and long that the players were in fact real. The moderator --- not of the 'Internet Poker' forum where the thread appears --- would later admit to having considered joining or funding another facet of the operation, which led many posters to question this visiting moderator's legitimacy in serving as a 2+2 mod himself. (That's an apples-and-oranges matter, not that the moderator in question seems to be a high moral beacon in poker concerns... for reasons we'll get to in a bit.)

Toss in peripheral concerns, including the long-delayed responses of FullTilt's 2+2 representatives, FTPDoug and FTPSean, and this whole mess stormed to the front of the must-read lists among the 2+2 populace, and has started to pick up outside notice as well. Even from the bloggers --- Iggy was perhaps the first to mention the thread specifically, while I waited on this one to see just how the thing would unfold.

Well worth the wait, even if the thread quickly has become a porridge of repeated unproveable claims and personal attacks. So here's my read on the situation, having read every entire damn post in the thread (and its outliers) to date:

The players in question seem to be real players, not bots. However, the players in question have been less than forthright about the 'system' that developed to assist them in their multi-table grinding. In case you're wondering, these are only marginally profitable players in terms of the play itself; their system relies on extreme tightness to avoid the big mistakes. The bigger bucks they're making are likely coming from rakeback deals for the action they generate. Each of the three players plays up to 12 screens at once and has never, ever played a single hand with any of the other players. Since all three are purportedly playing side-by-side (literally), in a tight setup featuring a string of oversized monitors, and also playing from the same IP address, it's quite obviously part of the plan.

The accuser and his supporters focus on the extreme similarities of the players' PokerTracker stats as proof that a bot of some form had to be involved, though the strident claims by the accused players that it's all merely a result of adhering strictly to a very tight system can likely never be refuted beyond all doubt. Such is the nature of adopting a bot-like approach to play; it makes the action so automatic that the difference between what a human does and what a bot might do in the same spot becomes irrelevant.

What compliments the problem is this: The ringleader of the grinding operation was quickly identified as an active procurer of scripts which can be used to automate online play.

Cue the shades of gray, baby, because here's where it gets really cute.

The use of these scripts is a highly debatable tactic, though in their simplest form they're designed to allow players to make their own decisions easier across multiple screens. However, these same scripts are part of the foundation beneath bot programs themselves, which combine these scripts with strategy and action components to fully automate the poker-playing process.

At what point does the whole automation thing fit into a definition of a 'poker bot'? Where exactly is the line behind game-enhancing scripts and bot-infused, fully automated play?

There's a faction among the 2+2 populace that I'm going to anger right now. I'm going to call them the technological angle-shooters. To them, anything is fair game unless it is absolutely, specifically, redundantly prohibited by the site ToC in question, and even then there might be another way to do what they want done. In this instance, that hard line insists that a bot must both automate the play and devise the strategic decisions.

But what about a series of scripts that went halfway there? What if a series of scripts was devised that recommended an action to a player, based on general guidelines, but left it to the player to make the clicking of the choice himself. In some of these hardcore technowonks' eyes, this makes it no longer a bot.

The argument comes up here because the overwhelming uniformity of actions over the hundreds of thousands of hands played by this group suggests that some steering by an overriding script was involved in the process. Just from the top, it's laughable that three players could play hours and hours at up to 36 tables at the same limits and never once accidentally end up two at the same table. The only logical explanation for this is that the three computers are linked (probably to a fourth computer), with a script specifically warning them when such a situation is likely to occur. The players wouldn't want to be at the same table, anyway, because that would open them up to collusion charges of a different sort.

The problem, though, is the automation itself. It's highly ironic that the accuser identified the purported bot users through PAHUD, another live-time application that gives its users an unfair advantage. (Many sites specifically ban the program.) PAHUD usage is linked to another spurious practice, datamining, which means gathering results from tables the user is not at, simply for the purpose of gathering data on potential opponents.

I detest the practice. It's cheating. Its proponents rationalize datamining and PAHUD usage by noting that since other people are doing it, they have to, too, or else they'll be at a disadvantage as well. And a lot of sites have recognized the difficulty in policing these programs and have basically thrown up their hands at the whole mess.

A few of the technowonks are only now realizing that the line between swiping an illicit computer edge and veering toward bot-driven play isn't as clear as it might seem. In true self-serving fashion, a whole bunch of 2+2ers are up in arms over what they perceive as Full Tilt's failure to properly police its games and eliminate bots, even as these same posters are the foremost practioners of hunting down the latest techie tool to play a few more tables or get information on their opponents that they haven't earned by playing the hands themselves. I can justify PokerTracker's usage (though I don't myself), because it provides a rough equivalence in knowledge value compared to what would be lost due to online poker's speeded-up play. But the Pokerbilitys and the PAHUDs and the action scripts... damn them all. Collectively. For tilting the table balance, for moving online poker farther and farther away from the live version of the game, which it's really supposed to replicate.

It's all about hypocrisy and greed, really. Hypocrisy on the players' part, to imagine that their preferred technological angle-shoot is somehow more legit than someone else's. Poker bots are just the extreme end of that whole damn sliding scale. The greed, well that's both the tech-driven players and the sites. The tech-driven players are greedy for cheating the system to grab a larger slice of the pie than their poker skills would otherwise entitle them to, and to the sites, for enabling so many tables to be played by a single account that it virtually begs for these add-ons and enhancements to be used... just because there might be a bit more rake to be had that way. I'm not calling out any site in particular, either; it's an industry-wide issue.

And, a now-open can of worms. It's important, because it shines a cold spotlight on exactly what technology, in its worst incarnation, has brought to the world of online poker.


eblonk said...

Quit playing Texas Holdem is my remedy. I think that bots don´t operate outside this most overrated of poker games ;)

Anonymous said...

Hey Haley,

I think you may be giving PAHUD a bad rap. All it does is pull stats in your PokerTracker database (which you have said is fine) and displays them on the table. In theory one could go back and forth between PokerTracker and the table and get the same info.

I haven't used PAHUD in over a year so maybe it's picked up some new features that make it questionable but as far as I know it's simply taking your PT stats and displaying them over your opponents.

Bill Rini

Haley said...

Hiya, Bill,

No, I know exactly what PAHUD does and how it works. The problem with PAHUD is that it makes datamining a profitable real-time operation, and allows for the introduction of extra information [illicitly obtained, in my opinion] into the game. No way can someone 12-table and keep track of which opponents are solid or not without a PAHUD or its equivalent. That's a mechanical assist that's available for a price, like buying all the best cards in Magic: The Gathering, to give all the script kiddies a comparative marker with which they'd be familiar. Poker should be about who plays their cards better, not about who has fancier clothes or better toys.

Anonymous said...

I think the key is that with some sites, you can use PokerTracker (or other software?) to data mine hands that you're not even in. I know this is not possible on Stars, but this is where people can gain an advantage on certain sites. PokerTracker is a useful and I think valid tool for helping you have a really good memory of the hands you're in and the players you're playing against... but if you can import thousands or millions of hands you're not even a part of, and use that information to influence your decisions using PAHUD, then I think that's less ethical.

Biggestron said...

Nice post Haley.

I don't consider using any database or database display system (ie pokertracker or PAHUD) to be cheating unless a site specifically bans the use thereof.

The online sites should serve customers who don't want to face bots and set up tables that are specifically for play that cannot be software assisted. Full Tilt used to do this for all of their tables (they didn't provide hand histories).

Look at it this way: consider computers as a tool to extend our minds. If someone can create a tool that helps them perform better then they deserve to reap the benefits. That better tool could be any of larger monitors, faster computers, more reliable internet connection or even software to assist with game play decisions.

Similarly for live play - if someone finds a way to improve their mind (exercise, nutrition, drugs, whatever) that gives them better recall then their opponents then they deserve the benefits. Is it fair that people with ADD can take ritalin to calm their minds and focus?

So long as there is no collusion then I can't see how it is cheating. I'd go so far as to say that even sharing database information among the team members is ok (how many times playing live have you accepted information on an opponent?).

Meh, long comment - I should have probably posted at my own neglected blog about this!

StB said...

Excellent post. I love how you turned the tables.