NL Bots on Full Tilt

[Nortonesque]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
someone explain to me how these stats would prove they are bots anyway. Wouldn't it just prove they all know the system equally well and follow it exactly?

[/ QUOTE ]

Yes, that's what it would prove.



nlnut's contention is:

At some point they sat down with a pen and paper and wrote down a system, then spent a week memorizing it and threw away their notes. Then 3 people, through the power of sitting next to each other, managed to execute this strategy PERFECTLY, with so little deviation, that their stats matched exactly over, what, half a million hands and half a year.

They also discussed tough decisions, but these decisions did not create any perturbation in the stats (on the river no less). These discussions didn't create any deviance from the system, not even a beneficial one. They also made changes to the strategy and communicated them to all players instantly and with 100% compliance.

Also, at one point one of the team members left, and a new team member was brought in. Without the use of ANY written or electronic aids this person was taught the system so that they instantly and without a learning curve could generate the same stats.



My contention is:

Does not compute.

[/ QUOTE ]

They are SUPER tight and playing basically a fit or fold strategy, I dont see why its so unrealistic to think that whatever system they have is simple enough that they are all capable of playing the same way. Its not like they are going to have anywhere close to the same number of decisions as your typical player.

PF is insanely easy, I dont get at all why people are even using VPIP as any sort of bot "proof". Of course they all have the same VPIP they are all following the exact same starting hand guidelines. Same way as if I never changed my starting hand requirements my VPIP would be the same across DBs, and if its not its because of variance, which would be the exact same variance a bot would face.

When you arent playing many hands post flop is easy to. They obviously follow something like "if you raised PF then bet the flop", if you called, unless you flop a set fold." If you get raised fold unless your hand is greater than 2pair. blah blah blah. Obviously we think this is stupid because it means you wouldnt be playing very good poker. But its been shown they dont play very good poker. They just play breakeven poker, playing a very stringent tight set of guidelines, stacking of weak unadjusting unnatentive fish and playing so tight they dont bleed away all their profits to the good players. which to me doesnt seem out of the realm of possibility that its simplistic enough that real people could follow them exactly. So again, what does any sort of statistical analysis proof? Other than proving they definitely arent bots (well doesnt even prove that, since you;ll just be able to say "well the humans took over for a while" or "well they changed some code")


[/ QUOTE ]
This might be the first sane post in this thread. Anyone could devise a system that would produce the same stats for people who adhered strictly to it (trivial example: fold every hand). I don't know why people think only a bot would be capable of playing by rote.

Throw PT/PAHUD stats into the decision matrix and I can see a system being very marginally profitable at FR 200nl.

As for the sweatshop angle, I don't think the economics work out. Once I know the system, why would I accept anything much below the system's hourly earn rate?

[KurtSF]

Everyone, read the [censored] thread before posting!

[james129]

Username Games Profit Stake ROI Total Profit
1forthethumb 209 -$3 $28 -5% -$685
full_tilting 233 $0 $20 -7% $91
mariojr 222 $4 $24 13% $786

[StephenVaughan]

We Have to vote with are feet on this, its the only way.
It seems the Full Tilt is putting the huge amounts of rake multiple bots playing unbelievably long hours produce above the concerns of its users.

Why do i need to play on a site that has no safeguards against "bot fraud"?


Answer, I don't!

[2OuterJitsu]

[ QUOTE ]
I don't know why people think only a bot would be capable of playing by rote.

[/ QUOTE ]

see my post in probability

[neverforgetlol]

[ QUOTE ]
[ QUOTE ]
neverforget,

Can you address this?

[ QUOTE ]
These statistics alone are not enough to "prove" that they are from the same player. You can't use a "simple" hypothesis testing procedure like a chi-squared test, and that won't indicate anything. A statistician examining the likelihood they are from the same player needs to have to other peoples stats and see how they behave. How much do these stat differ across 100k+ hands tight-nitty players?

There is a second issue with that there may be some correlation in some of the variables. For example, the "flop aggression", "bet flop", "raise flop" and "c/r flop" are connected to each other and a player with a high "flop aggression" will reasonably have high stats in the other 3 categories. An analysis of the likelihood they are the same player must have any correlation effect isolated. How these correlated variables behave in reality can be examined by studying the correlation of other 100k+ hands players.

Extreme care must be taken to conduct any probability tests and it is too easy to use a poorly-designed statistical test that does not consider the matters I outlined above. This is a formidable full-time task and should be undertaken by somebody with a solid postgraduate education in statistics (or having similar experience). I am not defending anybody here, but it is too easy to get carried away and it is easy to "prove" they are from the same player, but such test would lack mathematical rigority.

I'm not a statistician/mathematician but these are just my views.

[/ QUOTE ]

[/ QUOTE ]

I know you asked Neverforget, but this is sort of what I had been saying earlier (brings up a new thing since I was only talking about VPiP though). I've done a few things like this before.

The big thing is that these 'statistical tests' most people used in this thread assumed IID distributions. That is, Independent and Identically Distributed. With that, the correct formula for SD is sqrt (P * (1-P) /N). So if we're looking at how many ones we roll on a six sided die in 1000 rolls, we can use the formula above, since the die doesn't remember the previous roll or is affected by anything outside (the day, what someone else rolled, etc...)

Now, with these poker stats, in most of them both I's are violated. Taking VPiP first--one enters with different hands from different positions. Thus the picks are not from identical distributions...early position might be a pick from a distribution with a 7% chance of success, and on the button it might be 20%. Thus the above formula for SD is incorrect (but probably somewhat close--that's why I used it and adjusted my interpretation of the results)

The independence comes in on any post-flop stat, and combinations of pre-flop stats. With these many numbers, a lot of people (and a good test) would be to look at groups of numbers to test them to see if they all could be so close. However, post-flop stats are affected by the pre-flop decisions (and so aren't independent of them) and other stats are also correlated. For a simple example, PFR <= VPiP. So statistical tests that combine several stats like that (or just look at later tests) also will not have independence.

So what do people do in cases like that? Well, if they have a model, they'll simulate it thousands or millions of times and look at the distribution that results. The simulations (if they consider everything) will take into account all the correlations--in this case, if we said to raise with TT from middle position, we'd account for the correlation with 'raise after the turn if we've got an overpair or better' automatically. From the distributions that were generated, 5% and 1% (or whatever % you want) bounds can then be constructed (5% of the data lies further away from the mean than that point) and then you can test the data. What the person above is suggesting is similar; instead of simulate, just get a lot of people who play similarly (tight aggressive set miners, it looks like) and see what their numbers look like.

Hope this helps, and hope I haven't stepped on neverforget's toes (or reply).

Shane

[/ QUOTE ]

A nonparametric test is probably best since things like vpip aren't really normally distributed, and it would be hard to ascertain their succinct distribution.

[BigBiceps]

[ QUOTE ]
One of the things that has surprised me about this from the beginning: Regardless of whether they are bots or humans, why were they able to make money with such a predictable, exploitable strategy?

I would think it would work for a while and they would do okay against fish, but I would expect that the regulars would have been massacring them once they figured out their game plan. I've never played full ring (lol) 1/2, but I have played plenty of FTP 1/2 6-max and the games are filled with regular pros (admittedly, I play the dayshift [img]/images/graemlins/frown.gif[/img]). Isn't full ring the same way? If so, were those guys all asleep at the switch?

I would think that if guys who played this mechanically were in my game, I would be targetting them and stealing pots from them left and right (even though I probably would not have guessed that they were bots). They were apparently the most prolific opponents, so why weren't people trying to figure out how to beat them?

[/ QUOTE ]

I played a couple thousand hands against 3 of the 4 "bots". I did not know they were bots because I didn't pay attention because I was 8 tabling. I just knew that they had TAG stats and I classify each SLAG or TAG as either a) bad, b) solid or c) dangerous. I had classified all 3 as solid.

That being said I am up a nominal amount on 2/3 and down a nominal amount on 1/3. All 3 are down against me due to the rake. BigBiceps > bot. However, had I known they were bots it would be >>>>>>>>.

[jkkkk]

cliffnotes plz

[billmagnet]

[ QUOTE ]

This seems like a very reasonable explanation. The guys who are running the accounts could be just totally on auto-pilot most of the time and clicking the button which the software tells them to. This would help explain the long pauses for simple decisions (waiting to see what the computer says to do) and also how slow they are to react to obviously being exploited.

[/ QUOTE ]


You people are so ridiculous. Why the hell would they do this? Do you know how easy it is to have the mouse click a button in a random section, in a random time?

Even if FTP had some sort of software monitoring which could detect automatic clicks (which they don't, not even close) then they could run the program on another machine and have a VNC client do the clicking for them.

These guys are smart enough to create a winning NL holdem bot and you don't think they can go 1% further to get it to click for themselves? They just have to waste their life hours and hours every day to manually do the last 1%?

[seemorenuts]

This issue shows the weakness of a static game.

Think of how poker or its variants will be played 20 years from now.

Dynamic rules known to all, a good mix of skill and luck, no reason that any software could keep up with the changing game rules.

Simple, but no one has said it.

A few simple examples, different number of cards, streets, rules, structures even within a single game or tournament, more complex than HORSE of course, etc.