Broken RNG

    • VladimirN
      VladimirN
      Bronze
      Joined: 22.02.2008 Posts: 438
      I have played 32.000 SnGs on Stars and I've been running really bad from day one. How bad? Mathematician friend of mine roughly calculated that I'm running around 1700 BIs Under EV.

      Before I go into how impossible this is, I'll first elaborate how we got to this number by using 3 key points:


      1) Poker Tracking Software Graph

      HEM says I'm running 470 BIs under EV.


      2) EV Line Error When Someone Busts Out

      I didn't have too much time or patience to search for this too extensively but I searched the last 500 SnGs and found 20 of them I finished ITM without going all in a single time.

      Graph looks like this:



      I took only the ones where I went out on the second place because these SnGs perfectly illustrate what happens when someone busts out.
      Not only is EV line 24$ below my absolute minimum possible profit from those HUs, it seems that it includes my AI equity as well which is just ridiculous.

      I went AI 17 times with an average equity of 44% and an average chip stack of 1990. That should give me at least some more EV right? Apparently not.
      I could say that I should have gotten +5$ on each of those All Ins and I'd probably be roughly right so let it be so (all of this is just a rough representation of my claims).

      5$x15 + 24$ = 99$ per 15 SnGs or 6.6$ per 1

      If I include those 5 SnGs where I ended up 1st as well, we get up to 132$ for 500 SnGs that I have searched. Which means that my EV line should be 17.6 extra BIs above where it currently is (per 1000 SnGs).

      Or 563 BIs per 32.000 that I have played.

      Now this is just for those SnGs where I went into the money without an all in, but what happens in other SnGs where I get to the bubble without going all in? What happens when I get to the bubble with an all in? Correct me if I'm missing something here, but I suppose I'm ripped off every time someone busts out, but I'll leave this to professionals to calculate.

      EDIT: I'll search some more SnGs for the ones with which I made the above graph, for more accurate estimate. It's quite time consuming though because I have to search each SnG individually. That's why I searched only 500 for which I needed 4 hours.


      3) The Set-Ups

      Don't get me wrong here, I'm perfectly fine with set-ups and I know they happen often. I have seen enough of Poker After Dark episodes to get a rough idea.
      What I do mind though, is that those set-ups are much more often in their favor.

      I kept track of this through the last 4.500 SnGs which should be a fair representation of my whole poker carrier (and it certainly feel so) where I had:
      Set-Ups in my favor: 97
      Set-Ups against me: 330

      By "Set-Ups" I mean loosing a flush to a higher flush or FH, loosing set to a higher set etc. Either way the criteria was the same for both sides.

      EDIT: The most important thing about this is that I went all in post flop and getting it IN bad, when the odds are saying I should be good.

      Out of 427 cases that I counted, it should have gone 213:214 let's say... which means I should have doubled up to 3000 chips another 116 times which is A LOT.
      Number of chips is roughly accurate because I didn't count cases when I was really short on chips, and the range is between 750 to 4500, not counting odd set-ups in HU when it gets up to 9000.

      So 116 extra double ups per 4.500 SnGs, which is 825 total.
      Now, what I could have done with those chips is a bit tough and I should go through my database to see how often I cashed after the double up, but I could do a super conservative estimate:

      Let's say that out of those 825 cases where I busted instead of doubling up to a healthy stack, I'd end HU around 35% of the time (I think that is more then a fair estimate, it should probbably be much higher).
      That is 289 HUs.
      Now let's say in HU I got 50-50 chance for a first place, which is again very conservative considering my table selection skills:

      144.5 x 54.29$ + 144.5 x 29.23$ = 12069$

      Or 805 BIs.

      EDIT: These Set-Ups has been constantly happening since I started playing on Stars. It's not just a thing from the past 4500 SnGs.
      -----------------------------------------------------------------------------------------------

      This all Adds up to 1838 BIs under EV which seems almost spot on to what my friend calculated.

      Now, bare in mind that we were extremely conservative and didn't include some things which I won't go into detail here like:
      *Fact that I'm flopping pairs less often then I should
      *When I flop a something, the board is extremely wet and it gets increasingly wetter
      *My opponents seem to hit the board too often
      *Didn't count point 2) with the rest of the SnGs which are influenced as well

      etc.

      All of this will be taken into consideration when I hand over my HHs to professionals for a complete analysis.


      "Please Bear In Mind..."

      ...that I'm aware that these examples that I gave, might be off a bit, but I did them conservatively and the chances are that I'm more probably cutting myself short then the other way around.

      Detailed analysis will be made when the time is right.


      "What are the chances?"

      Here is the simulation with ROI Simulator that I did with Million simulations:



      I'm not quite sure what I can conclude ofother columns exactly, but the first column is quite clear. The lowest I can get is 1% ROI which is 1322 BIs under EV.

      That is the floor and I'm at least 1700 BIs under, which is impossible. And that excludes other points that I didn't cover in here that might add up to 2500+

      I'm planning to do one with 20 Million simulations to see if some -1% ROI creeps in which would officially make it possible (although I'd officially be the unluckiest person in the poker world; the guy who got hit by lightning 7 times would be jealous). Although I highly doubt that.


      "What Can Be Done About It?"

      Believe it or not, I'm still waking up each day saying to myself that all of this is a coincidence and that my luck will turn around any day now. I've been optimistic for four years now and I think I've reached my limit. I'm not really stressed that much but I'm hurt financially which is a pretty big deal to me.

      One of the best lawyers in the country is a close friend of my father and I'm thinking of going all the way with this. Lawsuit... the whole shebang.
      I don't really care whether their RNG is broken or it's like this intentionally. They are going to pay dearly for this.

      Still though, I'm thinking of playing up to 50.000-th SnG to either collect some more evidence or let PS correct their mistake. This lawsuit wouldn't really bring me any pleasure and I don't want to waste my time in the courtrooms. I'd still like this just to be corrected, to get my BIs back and continue playing with normal luck. That is all that I want honestly.

      Sadly though, considering the trend, I expect to go under another 500 BIs by the end of the year. In that case (or if I bust before that) I'll take action, and believe me when I say this: I'm very capable and intelligent person, my HHs and math are everything that I need as evidence and I wont stop until PokerStars is in ruins.


      It's getting late so I'll end here and go to sleep. If I missed something I'll correct it tomorrow. Be free to comment, correct me or give some advise, but please for the sake of humanity: don't troll.

      Have a nice day.
  • 184 replies
    • VladimirN
      VladimirN
      Bronze
      Joined: 22.02.2008 Posts: 438
      <Reserved>
    • metza
      metza
      Bronze
      Joined: 28.01.2012 Posts: 2,220
      Pretty big claims. Might be a good idea to have your HM2 graph which shows 400BI below EV as a start.

      Personally I think you can never rule out the possibility of rigging, just that it would be very stupid of a site to do so. But then again, it would be very stupid of a site not to have all the player funds too, aaaand...Who knows.

      Gl with lawsuit.
    • Spikelema
      Spikelema
      Bronze
      Joined: 03.07.2008 Posts: 55
      I would be looking at http://www.psimg.com/pdf/cigital-rng-labresults.pdf and if the RNG is indeed incorrect talking to cigital as they are the company who verified the accuracy of their numbers.

      Remember they tested this all up to the "decking" of card. (Meaning up to the moment before they are dealt.)

      The only method would be to manipulate that "virtual deck" after the cards have been played. While this is not impossible it is not a very likely practice as it would be the easiest to trace in a legal sense.

      My biggest point to consider would be the size of the pokerroom. This does mean people have more Data, bots are more prevalent and you are more likely to meet a group of people who are operating in Unison.

      The other consideration about Set-Ups is that it is never going to be an even split you are talking about holding the NUT or just a Monster. If I limp in with any two suited cards in all positions I am going to be beaten in a "Set-Up" well and truly more times than I would if I played only high suited connectors. I would consider your hand range with this statistic rather than just firing it off.

      A 44% on average equity would state you should only win 44% of those hands which would mean you would lose 6 times out of every 10. (Roughly). You cant calculate a "33% bonus" (+$5) In fact the low equity value should be a -EV as you are only getting value on 4 out of 10 hands.


      Edit: Appologies if I have made an incorrect assumption. I am not a professional Poker player and I did a fair bit of internet research to craft my reply, which may mean I missunderstood something.
    • VladimirN
      VladimirN
      Bronze
      Joined: 22.02.2008 Posts: 438
      @metza

      HM graph would be a good idea. Unfortunately I couldn't keep this 4 year old humongous database on my HD because it's simply huge and outdated.
      This thread wasn't meant to exist as some sort of proof in the first place, so believing me or not is your choice and I just can't fill in my DB with another 27k SnGs just so I can show this graph.

      Rigging is definitely a possibility and I could go into why and how (because I'm an IT programmer) but I choose not to tackle with that at this moment in time.


      @Spikelema

      This depends on how they tested it and at this moment I can't imagine how it's done.

      Let's imagine for a second that RNG is made this way intentionally and that it's rigged against players with certain stats that show that they good players. Let's say that it changes shuffles against me as soon as my VPIP or PFR or AFq reaches some range (reason for this would be to keep my profits lower and feed worse players, so they need longer time to bust).

      How can they test this exactly? You said they tested this "up to the moment before they are dealt". That doesn't even seem like a test to me.

      Regarding my 44% equity: When I am HU, I can win either 29$ or 54$. My equity cannot go into minus because I already got my BI back and we are basically playing for these extra 25$ (estimated +5$ that I have taken is only 20% of that).
      I'll try to explain it this way: Let's say we got the same stack of 4500 and I got TT vs his AA. We go all in and even though I'm dominated, I still have 19% equity which should give me:

      First place 19% of the times and Second place 81% of the times.

      0.19 x 54$ + 0.81 x 29$ = 10.26$ + 23.49$ = 33.75$

      Which is +4.75$ compared to the minimum 29$ even though I got only 19% equity.
      But if it can go into minus then this is a huge mistake by PT4/HEM and I should take this into consideration as well.
    • belthazorrrr
      belthazorrrr
      Bronze
      Joined: 12.02.2011 Posts: 1,334
      Did you ever changed your nickname in pokerstars?
    • janushr
      janushr
      Bronze
      Joined: 29.08.2008 Posts: 261
      hey,i added you as friend... :f_biggrin:

      whats your stars name??
    • VladimirN
      VladimirN
      Bronze
      Joined: 22.02.2008 Posts: 438
      My nick on Stars is the same as here and I didn't change it (can you do that really?)

      I did block it completely on SharkScope because I just couldn't look at it. It showed around -0.3% ROI I think. I will probably just reset it to this time and see how the things go from the moment of my OP.
    • Spikelema
      Spikelema
      Bronze
      Joined: 03.07.2008 Posts: 55
      There are three parts to the poker software

      1) The "visual" which is what they use to display cards, stats, players, chips etc.

      2) The "data" which is where they store your cash, hands, results etc

      3) The "calculation" this is where the RNG is, it generated a new "deck" of cards and using a psudo-random number generator shuffles the deck to create as close as computationally possible "Random Deck". This is then passed to the visual side which SHOULD pull cards from the top of the random deck only (just like how you deal a deck in real life).

      The company who tests this software will look at the RNG and watch it operate up until the moment it has generated a random "deck" it will run the test a large number of times comparing each deck to the previous looking for emerging patterns. The certificate says that the company upon testing this found no emerging patterns.

      Now there is always a chance the company then "cheats" using this deck that has been generated. (Not dealing the top card) but that is a very unlikely case because you would be able to check easily by examining the deck that is created by the RNG with the hand that is played out on the table (In a legal case), without even having the need to examine the source code between.


      Also when you considered your All-In equity during those times you came second you should actually be considering all the times where you came first and also went all in.

      It should be

      All-In Lost (came second) + All-In Won(came first) Then calculated against your EV.

      Also I know for HM2 the EV line actually seems to count hands you dont even go all in with. So possibly that is why the line is so inaccurate.
    • janushr
      janushr
      Bronze
      Joined: 29.08.2008 Posts: 261
      i just want to add,that is wroten on CIGITAL page,that NEVER nobody test it in LIVE enviroment,only test equipment half way across the world...

      there was some guy saying that it must be wroten in that way,cause its legal stuff,but still i really dont have trust completly when comes to that...

      i also find it very strange that there isnt any harsher controls or any test results published on net,it only show that poker room have possibillity to do what ever they want...even when you start to play,you prolly must agree to all what they say and simmilar...and there is lots of money involved not to be tested and proven even daily if neccesesary...
    • Vygantas82
      Vygantas82
      Bronze
      Joined: 02.04.2010 Posts: 361
      Some people probably remember that pokerstars RNG was "tested" not only by Cigatel but also by BMM. That is the letter from company about their "test"(that they sent in 2010):

      Thank you for your inquiry. BMM performed limited analysis of the RNG algorithm in 2004. The algorithm was found to be adequate for the intended use. That said, BMM did NOT evaluate the implementation of the RNG, the operating systems, the games, or any other functionality, which is always required in more regulated gaming jurisdictions, certainly in live casinos. In terrestrial based gaming any time game software is modified, it is reevaluated and certified to ensure continued compliance to the jurisdiction’s technical standards. BMM has not been engaged to conduct any other work for this company since the original engagement and, again, it was a very limited scope. BMM cannot attest to the site’s technical compliance OR non-compliance as the analysis performed by BMM may be performed on an on-going basis by another independent test lab or by the regulator themselves. In addition, BMM has asked PokerStars on numerous occasions to remove any reference to BMM from the web site, obviously to no avail. If you have questions about the site’s compliance to technical standards, please contact the MCK. If you are not comfortable playing the site and are interested in continuing to play online poker, I would encourage you research other operators until you find one with which you are comfortable. Kindest Regards, Richard Williamson


      As you can see pokerstars only wanted to get logo from Cygatel and BMM, not full and fair audit. But BMM did not want to give their logo in pokerstars web site for the meaningless test
    • janushr
      janushr
      Bronze
      Joined: 29.08.2008 Posts: 261
      why they not get FORCED to make some test and make it public???is there anybody possiblle to do that??as i also tried to contact someone,but no use,as all poker rooms are just sitting somewhere in forsaken place just to avoid taxes,nothing else??why they would cheat their customers???

      for MONEY($$$),perhapse :f_biggrin: ???
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,295
      I don't play SnG's, but I've always had a suspicion that something is not quite right the way tracking software calculates EV.

      I think it's possible certain styles of play could magnify this error and cause it to appear as if you are consistently running bad. Why do I think this? Because you are not the first recorded case I've seen of a SnG player apparently running so far below EV.

      What makes you so certain this an RNG error and not an error in the way HM2 calculates EV?

      Now, I'm 100% not qualified to tell you what a standard swing in SnGs is. Either way, running this bad is certainly not impossible, merely improbable.

      Another thing to mention, from a cash game perspective, allin-ev is pretty irrelevant. It's no measure of how good/bad you are actually running. To give 2 examples -

      1 . You get dealt AA 50 hands in a row. Your hand only holds up 70% of the time. Allin EV line says you are running bad. You are clearly running good.

      2. You get 98% of your stack in with the nuts. Villain hits on the river and shoves that last 2%. According to allin EV you are down. Everyone knows the EV of the situation was actually positive.
    • belthazorrrr
      belthazorrrr
      Bronze
      Joined: 12.02.2011 Posts: 1,334
      You cant compare sngs to cash for the reason you said. I had been running way bellow ev. I just think that pokerstars is rigged somehow and i could justify it with card distirbution. I am a winning reg in pokerstars platform but is some platform had near good software i would have swapped

      I quit whining and i am in a mentality that no matter what pokerstars will always steal my money but still i will make decent money. When i run 250-300 bis bellow ev in a 20k sample then something is wrong with my account

      Again you cant compare cash with sngs. Because sngs is an close to all in game while cash is totally different
    • Vygantas82
      Vygantas82
      Bronze
      Joined: 02.04.2010 Posts: 361
      I don't play SnG's, but I've always had a suspicion that something is not quite right the way tracking software calculates EV.

      Only HM2 has such problem( If they didn't fix it yet). But i think OP uses HM1 which calculates EV correctly
    • Gentari
      Gentari
      Global
      Joined: 08.05.2011 Posts: 344
      I think it is rigged when I don't win and I play perfect poker when I do :f_grin:
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,295
      Originally posted by belthazorrrr
      Again you cant compare cash with sngs. Because sngs is an close to all in game while cash is totally different
      Btw, not comparing SnGs to cash. I'm showing how useless the EV line is in cash games. Why would the EV line in SnGs suddenly be any better and representative of true EV? For a start it's surely more complex when you have both $ev and chip-ev you could potentially think about.

      Better to just look at your green line imo.

      What's your viewpoint btw, that OP is just running really bad?

      @OP - graph please :D
    • belthazorrrr
      belthazorrrr
      Bronze
      Joined: 12.02.2011 Posts: 1,334
      Because sngs are much more an all in game especially in later stages that is more crucial and more heavily icm depedant

      My point of view is that something is wrong with OP account. Yeah all can laugh but this is my opinion. When you have 200-300 bis downswing in turbo games over 10k sample it cant be just coincidence. I had same dowswing which i never revived but again the only reason i played is because i still have decent ROI and with rakeback of supernova i make good moniez
    • Spikelema
      Spikelema
      Bronze
      Joined: 03.07.2008 Posts: 55
      One final point about why it would be moronic for poker rooms to cheat.

      You argue that they want the fish to stay and give them the good hands and the better players the worse hands so the fish stick around.

      Why would you give the player who understands the statistics/percentages/odds/out and all those thing, hands that defy those calculation.

      Its like trying to lie to a chemist about knowing what all the ingredients in shampoo actually do. You may fool the guy who lives around the corner but the one who actually knows is going to point it out.

      They would be digging their own grave. (It would also be more than a few posts by a very very small minority about how they think its rigged.)
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,295
      Hopefully this doesn't need to turn into standard poker is rigged thread. :f_love: