*NOTE* Seeing as this post is the one that has caused a great deal of traffic over the last few days, I have decided to “sticky” it. All normal and more recent posts follow immediately below this. Thanks. *NOTE*
As some of you know, I have been playing online poker since October. At one site, Poker Room, I managed to make a great deal of money, but ONLY from playing exceptionally tight. The reason for this was the fact that every time I played a reasonable hand, that was solid, but not near the nuts, I would get beat by a horrendous play from a fellow player. I won over $500 my first week at the game, but once I hit $500, it was a struggle to go higher, as the beats got worse, and worse, until I finally cashed out in late November for good, only doing Freerolls with my remaining player points.
Lately, my bad beats have been continuing in tournaments, even small play money ones. Beats that are so bad, that I believed they could not be attributed to the “everyone here is a chaser” theory, which states that a lot of people find the risk small enough to chase draws until the river. I spoke with other players who felt the same way, and decided to use mathematical analysis to try and determine what (if any) was the difference between playing a home ring game with average players, and online with average players. Before I start with my analysis, I’d like to say that my initial prediction was that I would find overwhelming evidence that Poker Room is indeed stacking the deck against certain players.
So… Is Poker Room rigged?
To determine if the outcomes at this online poker site are fair, we must first have a comparison from real-life games. Instead of using many complicated mathematical formulas, I decided that the most efficient way would be to watch a real live ring game. I decided to take stats on the winning hand, to see if their was a difference between “big wins” and “small wins”
In this real live ring game, I counted the number of times each type of hand took down a pot. I did NOT count pots won on folds on betting before or after the flop, after the turn, or after the river. Only hands that resulted in at least 2 players seeing the hand through were tabulated. This was for the simple reason that I did not want to tamper with the game, as well as the fact that I would not be able to see the same kind of hands online. Splits were counted as the hand that was split.
WINNING HAND - # OF HANDS OUT OF 250 WON - PERCENTAGE WON
2 Pair - 76 hands - 30.4%
Pair - 68 hands - 27.2%
3 of a Kind - 33 hands - 13.2%
Straight - 23 hands - 9.2%
Flush - 20 hands - 8%
Full House - 20 hands - 8%
High Card - 7 hands - 2.8%
4 of a Kind - 2 hands - .8%
Straight Flush - 1 hand - .4%
Royal Flush - 0 hands - 0%
These stats also seem to gel well with webmaster Chris’ math from This is How Things Should Be.
Now, I did a similar study on Poker Room at a $1-$2 NL table with 10 people. Over the five hours and fifteen minutes I watched the table, there were 10 active players for all but about 47 minutes, and even in the 47 minutes of sub-10 players, there was always at least 8 players present. Again, the only hands that were counted were hands that resulted in at least 2 players seeing the hand through. Splits were counted as the hand that was split.
WINNING HAND - # OF HANDS OUT OF 250 WON - PERCENTAGE WON
Pair - 67 hands - 26.8%
2 Pair - 62 hands - 24.8%
3 of a Kind - 33 hands - 13.2%
Flush - 29 hands - 11.6%
Straight - 26 hands - 10.4%
Full House - 26 hands - 10.4%
High Card - 5 hands - 2%
4 of a Kind - 2 hands - .8%
Straight Flush - 0 hands - 0%
Royal Flush - 0 hands - 0%
OK, so now we have some data. A brief glance shows us that the numbers do not look significantly dissimilar. They are much closer than what I thought we might see prior to conducting this test. But once we break these numbers down, we discover some startling facts.
We first tabulate “big hand wins.” For all intents and purposes, a “big hand” is either a flush, straight, or full house.
OFFLINE: 23+20+20 = 63 hands taken down by big hand = 25.2%
POKER ROOM: 29+26+26 = 81 hands taken down by big hand = 32.4%
That’s over a SEVEN PERCENT difference from a regular home ring game. If you play roughly 30 hands per hour to completion, this is somewhere between 2 and 3 extra “big hands” per hour.
What this also proves, is that since these “big hands” were counted in the final tally, their bets must have been called by another player, and the hands were played to completion. While I did not tabulate the losing hands (or “bad beats” in this case), this statistic implies that they also had, at the very least “solid hands” (Top pair with good kicker, 2 pair, 3 of a Kind) to warrant such a call to and on the river. This tells me that good hands at times are dealt to more than one player with the possible intention of garnering more action, therefore, a bigger rake for the site.
Another scary statistic I started taking note of over the last 50 hands was the amount of times a flop came out of three of the same suit. Mathematically speaking, this should occur once every 65-70 hands (1/4 * 1/4 * 1/4 = 1.56%; Note: not exact because not accounting for odds lost to cards being out on table). Therefore, I should have seen either zero, one, or an outside chance at two flops that contain all suited cards. This happened five times - over 6 times the mathematical projections. Granted, the sample size is not big enough to draw a final conclusion, but it would seem that this trend of flops of a single suit is meant to keep players with weak overall hands in the pot on draws (ie Jc-3s with Ac-6c-2c on the table). There were also 2 times in the 50 in which FOUR cards of a suit came out on the table, significantly higher than the math would predict. (In both of those cases, players with significantly inferior starting hands were rewarded for their persistent chase of a draw.)
So… Is Poker Room rigged?
I feel the numbers speak for themselves. While I’ll admit, they weren’t as lopsided as I originally thought, I feel as if they are skewed far enough for ME to probably cease playing for real money at Poker Room for the time being, until I look at some more hands. The stats can be interpreted if you’d like to show a favoritism for extremely loose players, in what seems to draw them into more hands for rakes, and draw them back to the site when their cash is low. Again, I haven’t taken enough stats to come anywhere close to confirming this, but I’m not a big fan of them. I cannot condemn Poker Room with these stats, but I feel that reviewing the numbers shows that it is not in my best interest to play there for a bit. These stats only show a trend, not an official study, so I merely offer these numbers as a thinkpiece, this is NOT meant to get everyone to stop playing at Poker Room. It is a free country after all, and instead of keeping these in my Excel, I thought it would be nice to post them for all to have a look at.
UPDATE: 12/30/04
OK, since I have been temporarily incapacitated of moving thanks to some minor surgery, I decided to study another 750 hands at Poker Room over the last 3 days… Here is what the new totals (including previous observations) look like. I did 3oo at a $1-2 NL, 300 at a $5-10 NL, and 150 at a $10-20 NL table.
WINNING HAND - # OF HANDS OUT OF 1000 WON - PERCENTAGE WON
Pair - 265 hands - 26.5%
2 Pair - 237 hands - 23.7%
3 of a Kind - 135 hands - 13.5%
Flush - 118 hands - 11.8%
Full House - 109 hands - 10.9%
Straight - 102 hands - 10.2%
High Card - 22 hands - 2.2%
4 of a Kind - 10 hands - 1%
Straight Flush - 2 hands - .2%
Royal Flush - 0 hands - 0%
Big hands won - 329/1000 - 32.9%
Statistical probability - 27%
As for flop statistics– Out of 800 hands:
Times flop has been suited cards - 46/800 - 5.75%
Mathematics predicts about - ~ 1.755%
Statistics continue to hold fairly true. I’m still amazed at the number of flops that continue to come out suited. THIS may be the reason why flushes are so prevalent on this site, is that a high card chases the flush when they know they only need one more of that suit to hit the nut flush. I have also noted that the board has paired NUMEROUS (I did not take stats) times on the flop has well which probably leads to the higher number of full houses as well.
More stats to come as I feel, but trend seems to hold true.
Recent Comments