In 2015 I have played 3 dollar turbo DoN tournaments on 888 poker. In the long run it is fair to assume that you should at least run above EV in as many tourneys as you run below EV if you play a mathematical correct game. Usually you only face one opponent in an all-in but occasionally you face two. If you face two opponents then of course most of the times only one of these people runs above EV and two below EV for that particular pot. Since the blinds increase fast, in many tournaments you have to rely on your “luck” in a pre-flop all-in in the end of the tournament. Since the aim is to survive not to accumulate as many chips (money) as possible as in a cash game I count the tournaments. A tournament is the entity that matters economically, that is why it is counted in my survey. 50 % of the players win and 50 % lose in a DoN tournament so it is reasonable to assume that a player that plays close to mathematical correct should at least have 50 % of the tournaments above EV.
I checked this assumption with my 15 most common opponents. I made an estimate calculation, so one or two tourneys can be wrong, but doesn´t affect the conclusion. The results are as follows: Player 1: 12 above EV, 12 below EV, 7 no all-in pot, Player 2: 16 above EV, 11 below EV, 8 no all-in pot, Player 3: 13 above EV, 10 below EV, 6 no all-in pot, Player 4: 10 above EV, 5 below EV, 1 no all-in pot, Player 5: 7 above EV, 8 below EV, 2 no all-in pot, Player 6: 8 above EV, 4 below EV, 1 no all-in pot, Player 7: 7 above EV, 2 below EV, 1 no all-in pot, Player 8: 3 above EV, 3 below EV, 0 no all-in pot, Player 9: 6 above EV, 5 below EV, 1 no all-in pot, Player 10: 5 above EV, 3 below EV, 2 no all-in pot, Player 11: 3 above EV, 8 below EV, 1 no all-in pot, Player 12: 5 above EV, 6 below EV, 1 no all-in pot, Player 13: 5 above EV, 1 below EV, 3 no all-in pot, Player 14: 4 above EV, 3 below EV, 1 no all-in pot, Player 15: 4 above EV, 2 below EV, 2 no all-in pot. Totals for these players are: They have run 108 tournaments above EV, 83 tournaments below EV and 37 with no all-in pot. The ratio is 108/191 which is 56,54 % above EV and accordingly 43,46 % below EV. These players are in my opinion not better players than I am, for most of them, the opposite. This shows that if you play a mathematical correct game, in the long run, if the card dealing is fair, your above EV tournaments should at least be as many as your below EV tournaments.
This year I have played 297 of these tournaments on 888 poker, the goal was 300. At last, so to say, I had enough. I am unemployed and don´t like to donate my money to multi millionaires. On 25 of these I wasn´t involved in any all-ins so I will not count those. I have run below EV on 162 of these tournaments and above on 110. The ratio is 110/272 which is 40,44 % above EV and accordingly 59,56 % below EV.
Here is the graph. My profit is $51,4 and the fees I have paid are $59,4, which means the net result is minus $8. The blue line shows how many buy-ins I have “lost”/”gained” per tournament. If I run above EV the line increases with one buy-in and if I run below EV then it decreases with one buy-in. In the long run the line should oscillate around the x-axis. A notable event was that I cashed out some money when tournament 140 was played, and after that it was rather much downhill for the luck in all-ins. (Well it was downhill before that too.)
We assume the probability to run above EV if you play a rather mathematical correct game to at least 50 %, so we assign the proportion to 0,5.
H 0 : p = 0,5 (site is random)
H 1 : p ≠ 0,5 (site is not random)
p in sample = 0. 4044
t = (0,5 – 0,4044)/ √(0,5*0,5 / 272) = 0,0956/0,0303 = 3,155
The cumulative probability in a table shows approximately 0,99918
Null hypothesis is rejected. The probability of running this bad is 0,00082 which is 0,082 % or 0,82 per mille.
At the same time I have played the closest equivalent to this tourney on Pokerstars to have something to compare with. I have played 642 of these. The tournament is called fifty50. The tournament ends when there are 50 % players left as in a DoN tournament. In addition you get extra money for every 100 accumulated chips.
Here is the graph:
The green line is as above fees paid. The red one is winnings ex fees. And the blue one is the rig index line that should oscillate around the x-axis in the long run, which it seems to do in the beginning, then something “happens” as it did at 888 Poker too. Since you earn more money if you accumulate more chips I have included two net all-in lines too, the light blue and the violet. One is Hold´em Managers way of calculating and one is my own. Winnings are $12.81 and fees paid are $-109.0 the net result is $-96.19. “Stolen” in all-ins are $35.26 according to my method, which is the purple line and $47.3 if u use HEM´s method, the light blue line. According to me the Rig Index line more accurately reflects how much money is “stolen” since if you very often have bad luck in all-ins you have not the benefit of a large stack in the end of the tourney and misses the opportunity of terrorizing your opponents with all-in bets. It is also important to have the good and the hitting hands rather evenly distributed on the different tournaments. If you as a Poker Site want to cheat you can work in at least six ways. 1. Make sure the good player is more unlucky than lucky in all-ins. 2. Make sure the good player is unlucky in big pots and lucky in small pots 3. Deal the strong hands to the good player in a small selection of tournaments since the “excess winnings” in one tournament is worth less than a bunch of lost buy-ins in some other tournaments. 4. When a large portion of the strong hands to the better player are dealt, deal crap to the other players so that they fold. 5. Make sure the good player is “unlucky” when he is small stacked and make sure the small stacked fish have nine lives. 6. When the good player is small stacked give him KQ that is a no-brainer all-in hand and give opponent AK, or maybe AJ vs AK, then it is only “natural” that he loses. I only remember that I have run above EV on one poker site and that was one account at Cardgrinders in 2008 on the Entraction network. There I had no boosted ROI from freerolls since they didn´t offer any. Maybe that was the reason they dealt cards fairly randomly? Can it be that I am unlucky on all of these sites or can it be that they all use some algorithm to classify a player and deal cards accordingly?
A notable event was that I won $90 on a $20 free bet in the Sportsbook, when tournament 35 was played, thereby increasing my account´s ROI considerably and after that it was downhill for the all-in luck. I also won $26 on a free bet, when tournament 96 was played and then it was even more downhill for the bad luck. I cashed out most of my money when tournament 108 was played. The all-in luck line, a trend or just bad luck? I have always suspected that Poker sites use some algorithm that adjusts the odds according to the net winnings. Have I been less suspicious after this? At the two bottoms between 110 and 120 I use my last buy-in two times and both times I miraculously win with my last buy-in. The same phenomena I have noticed at other sites.
I have played 642 tournaments this year on Pokerstars, I wasn´t involved in an all-in in 108 of these tournaments. For 534 tournaments I have run above EV in 249 and below EV in 285. The ratio is 0,4663.
We test the proportions as above
H 0 : p = 0,5
H 1 : p ≠ 0,5
p in sample = 0,4663
t = (0,5 – 0,4663)/ √(0,5*0,5) / 534) = 0,0337/0,0216 = 1,56
Null hypothesis is not rejected.
We also test the mean for the all-in luck curve. The all-in luck mean in the long run should be 0.
H 0 : µ = 0 (site is random)
H 1 : µ ≠ 0 (site is not random)
mean in sample = – 0,0660
standard deviation in sample = 2,815
t = – 0,0660- 0)/ (2,815 /√ 534) = – 0,0660/(2,815 /23,11) = – 0,0660/0,11218 = -0,588
Null hypothesis is not rejected.
We observe that the all-in luck is constantly below 0. At the tables I play at there are some regulars. The one I have faced the most times is a “player” called Piskar88. He is a constant winner according to Sharkscope. If you check my name on Sharkscope I am a constant loser. I have summarily analyzed his playing and calculated all-in luck for all of the pots I have hand history of. He is a more loose player than I am so you would expect his Rig Index line to be below 0, but it is the other way around. I am a rather tight player and you would expect my line to be above 0 and it is way below 0. When I first played against him I soon discovered something peculiar about “him”. In the note section that you have for each player I wrote “Nostradamus” as a comment. He could move in as a dog and still win since he “predicted” what would come on the flop. Here below is his all-in luck lines and Rig Index line for the hand history I have.
Here we have three lines that should oscillate around zero in the long run (If you play a loose and stupid game the Rig Index line should be in negative territory but the other two should oscillate around 0. Does it look like any of these lines will oscillate around 0 in the long run? OK, anything can happen, but for me it sure looks like a trend not a coincidence. Is it any wonder this “guy” is a constant winner? On 115 tournaments with a buy-in of 3,3 dollars and 10 players he has excess winnings in all-ins of 1.08 dollars per tournament! I would have accepted this as a “black swan” if it had been 10 pots in a row or 5 tournaments in a row, but not on 70+ tournaments. When the pots are small he sometimes loses, when they are big he can move in with AQ vs AA and still win. When he moves in with a lower pair he hits so often it is rather obvious something isn´t quite alright. I don´t remember seeing him losing with a bigger pair against a smaller pair in a pre-flop all-in. So is this a human being or a liquidity providing robot skimming players money?
I do a check on Sharkscope and he has ITM of 53,4 % and a ROI of 3,3 %. I have ITM of 57,9 % and a ROI of -5,7 %. His Rig Index is 55 % and mine is 46,6 %, that fact combined with the fact of my bigger ITM number shows that I play tighter and that my Rig Index should be “better” than his. His Rig Index is 55 % and mine is 46,6 % and I play tighter! One conclusion that looks plausible for my high ITM and low ROI is that I play too tight and should move in earlier. That is not true. Nearly every time I have thought of moving in and changed my mind someone else has moved in and I had lost that pot and busted earlier than what was the actual case.
I have tried to chat with him and finally one day he answered and said he lived in a Russian town called T***. When I pointed out to him that he was very lucky in all-ins he said that I was just a bad player, a fish, that was the reason I was losing. Can a winning player like him, that is a good player, be unaware that he is lucky like that? No. Is he saying that just to bully and make me tilt, possibly. After he called me a fish he wrote something strange in the chat (strange if it actually is russian).
He wrote something like “teshe ne vse prosral” “o tupoj treved” I have asked a russian what this means. She didn´t understand what it meant. She called it “strange russian”. Tupoj is probably stupid. She thought it should mean that something is not the end of the world. If these lines was for me I don´t know. A russian that speaks “strange russian”.
One thing that enters a mind of a conspiracy theorist is an arrangement of many liquidity providing robots that are overseen by one single person. These robots work for someone (probably the poker site) and skims some per cent every tournament from ordinary players. These robots are overseen by some person that answers in the chat so that people doesn´t start to be suspicious since this winning player never answers the chat. Be what it may, his all-in luck is extraordinary. And it is also extraordinary that a constant winner doesn´t understand the difference between luck and skill and recognizes that he has been lucky. One thought is that since he is such a good customer generating so much rake, if he really is a legitimate player, the site wants to keep him “happy” with some “luck” so that he doesn´t leave the site.
We do some hypothesis testing here too.
We test the mean for the all-in luck curve. The all-in luck mean in the long run should be 0.
H 0 : µ = 0 (site is random)
H 1 : µ ≠ 0 (site is not random)
mean in sample = 1,2505
standard deviation in sample = 3,95104
n = 100
t = 1,2505- 0)/ (3,95104 /√100) = 1,2505/(3,95104 /10) = 3,165
Null hypothesis is rejected on the 99,9 % confidence level.
His playing is too loose to be mathematically correct so testing the Rig Index line is only interesting as information not as a conclusion. I would expect his Rig Index line should have been about – $20 ( a proportion of 45-47 % instead of 55 %) on these 110+ tournaments.
We test the proportions
H 0 : p in sample = 0,5
H 1 : p ≠ 0,5
p in sample = 0,55
t = (0,5 – 0,55)/ √(0,5*0,5) / 100) = -0,05/0,05 = -1
Null hypothesis is not rejected. As said above his Rig Index is 0,55, a fair assumption is that it should be approximately 0,45-0,47 in the long run.
Here are his “luck lines” compared to mine.
One is a constant winner and one is a “fish”, wonder why! The Rig Index line is positive for the looser player and strongly negative for the tighter player. The lines should be reversed. That suggests something is seriously wrong in the card dealing. If you look closely you see that the few times his all-in luck lines hit 0 they immediately bounces back to positive territory, for me it is the other way around. As soon as my all-in luck line bounces at 0, immediately it is back to negative territory.
Conclusion: I refrain from speaking my mind, if you read between the lines you will understand what my opinion is about their alleged random shuffling and dealing.
I am thinking of preparing graphs for my other most common opponents at Pokerstars.
Hand Histories can be obtained by contacting me.
Thomas Bengtsson
2016-09-07
daligvardegrund 