Poker RNG audit

puzzlefish

puzzlefish

🙉 🙈 🙊 student of the donk arts
Loyaler
Joined
Feb 18, 2018
Total posts
4,426
Awards
3
CA
Chips
275
Hi everyone,

I am a long time lurker on multiple poker forums and NLHE enthusiast for about 2 years now. I have read my fair share of threads about RNG rigging and bad beats. To that end, I want to propose to you that online poker RNGs are likely not rigged to favour certain players but rather that there are algorithms in play that will favour certain players to win against others depending on player performance relative to others at the table over a number of previous boards.


I would like to demonstrate to you what I mean, but it will require the collection of data from broadcasts that show all players' hole hands. I have done this so far for pokerstars's Sunday Million final table from February 11th. The trouble is that this takes hours to do manually and to format the data to show the required variables:


1. Player chip count before each flop
2. Player position at the table (Dealer, SB, BB, etc.)
3. Player hole cards
4. The board cards (if any)
5. The winning player
6. The winning hole cards (out of all players including those who folded)
7. The change in each player's chip count as a result of the transactions that took place with each deal.


To be clear: I am proposing that the anomalies observed in online poker are not due to card rigging but are based on player performance against other players at the same table (i.e. wins, losses of chips, frequency of bets, etc.)




Can anybody help me out with this? I have good results so far with my first data set but will need a lot more to establish statistical significance. Any programmers/statisticians that can help?
 
smallfrie

smallfrie

Lucky Ducky
Loyaler
Joined
Sep 9, 2015
Total posts
2,661
Awards
2
US
Chips
153
I am proposing that the anomalies observed in online poker
What anomalies? Are you talking about losing hands that happen in poker? What makes you think there are anomalies? A hunch?
 
WhereDidMyEVGo

WhereDidMyEVGo

Visionary
Silver Level
Joined
Jul 20, 2014
Total posts
547
Chips
0
What? So it's rigged or it's not rigged? It could be rigged against you in some cases and rigged for you in some cases?
 
puzzlefish

puzzlefish

🙉 🙈 🙊 student of the donk arts
Loyaler
Joined
Feb 18, 2018
Total posts
4,426
Awards
3
CA
Chips
275
smallfrie and WDMEVG:

The simplified answer is that, in my hypothesis, RNGs "rig" the game for everybody equally, but they do so according to a set of rules which have nothing to do with the cards that are actually dealt. The rules apply equally to every single player.


When I mention "anomalies", I don't mean just any good beat or bad beat. To me, anomalies are situations that are statistically unlikely and which happen to coincide with critical tournament or game events. For example, somebody at a table wins a big pot and with a full house and gets dealt AA the very next hand versus KK, 99, and AKs which ends up in a flush for the player who lost the previous board. That is an example of an anomaly because you have several inducing hands that warrant shoving and a statistically unlikely flush as the winner right after a full house.


I hope this clarifies what I am talking about a bit and shows you that I am quite serious about what I am talking about... Not just your usual player complaining about bad luck.
 
puzzlefish

puzzlefish

🙉 🙈 🙊 student of the donk arts
Loyaler
Joined
Feb 18, 2018
Total posts
4,426
Awards
3
CA
Chips
275
I would also like to add that the primary purpose of this thread is to seek help with data mining and analysis as described in my first post. If I can get a few folks on board that know how to do this efficiently, the statistical analysis of the data itself will define the anomalies for us.
 
Peppinotom

Peppinotom

Legend
Silver Level
Joined
Apr 7, 2017
Total posts
1,201
Chips
0
That would be easy to discover, if we just played a CC Freeroll with open cards, to see, what is actually going on. As long as you don't have ALL data, you will ALWAYS have a subjective result. And I am pretty sure, that IF THERE IS REALLY ANY RIGGING, it would be adjusted regarding the levels of buy in. So you will NEVER find out. Quit playing, or be more careful.
 
TeUnit

TeUnit

Legend
Loyaler
Joined
Jan 20, 2009
Total posts
4,734
Awards
14
Chips
77
Think this would require a very large sample size and it would be difficult to do without a large amount of data mining. But, then it would be very easy to analyze on pokertracker or holdem manager.
 
NoPlace4U

NoPlace4U

Rock Star
Silver Level
Joined
Sep 5, 2016
Total posts
468
Chips
0
I think the RNG is not rigged but it is flawed, a lot of statistical anomalies occur.
 
Mikeisanace777

Mikeisanace777

Rock Star
Silver Level
Joined
Jan 24, 2017
Total posts
150
Chips
0
It's a computer not realistic like live poker

Assuming an rng uses 52 numbers-18--3-1-1-1-1 it's going to have some odd variability like video poker. This represents the deal x9 -flop-burn-turn-burn-river



The Rng's probably doesn't use cards they use numbers like 1x as 2 of hearts,1c as 2 of spades etc.. This can make 208 variations floating in free space pre deal-18 for 190 numbers free-floating preflop and is probably not realistic or fair. In live poker, you can assume 9 outs to 3 outs total for the river and turn to make a flush, however, if you play with a deck those outs are sometimes gone or to low in the deck and shall never appear in that particular hand.


Now imagine Live if they dealt 9 players in then reshuffled the deck burned one reshuffled their deck to a flop/burn shuffle/turn a burn one more shuffle then a river it may be similar online.


The only way online poker can be fair is if the 208 variations pre-flop stopped as static in a sequential stack and simply went -18-1-3-1-1-1-1 the randomness occurred pre flop as a constant shuffle the after occurrence would be similar to cards falling not free-floating variable numbers which is unfair and way more unpredictable or unexploitable so this could be a motive of such a game.
 
fletchdad

fletchdad

Jammin................
Loyaler
Joined
Feb 3, 2010
Total posts
11,706
Awards
2
Chips
122
I dont remember the site, but there is one that checks the RNGs of many major online sites, (independent site, maybe someone knows the name, I was there a couple of years ago, but cant recall its name) and they have literally millions of hands from every site they check, and they check the randomness of hands in all respects.

Do you mean something like this? Cause it is already being done.
 
puzzlefish

puzzlefish

🙉 🙈 🙊 student of the donk arts
Loyaler
Joined
Feb 18, 2018
Total posts
4,426
Awards
3
CA
Chips
275
Sorry for the late response. I have been away. I don't think the current checks being performed on RNGs assess the specific property of the RNG that I am proposing in this thread.

Another way to think of it that may make it clearer to understand: Think of each unique board as a unique scenario with a predetermined outcome (no randomness at all). All such possible boards put together will appear statistically normal in terms of which cards are winning in the long term. However nobody is looking at the statistics of which players are in the winning position versus the losing position (and the patterns involved in their placement there) regardless of the cards being dealt.
 
Top