Shuffling in Poker

Over the past few years one of the most controversial areas of debate in Internet poker has been the issue of online shuffling. Clearly the purpose of any fair shuffle is to create a random deck such that every possible sequence of cards is possible, while at the same time making it impossible for anyone to predict the position of any card in the deck. In principle, unless they are crooked or contain bugs, online shuffles should be closer to random than can possibly be achieved in a live-action game, where the cards are often just collected, riffled a couple of times and dealt, with the result that clumps of cards can sometimes stay together. As an example, the Paradise Poker website states that: ‘No deck of cards in any brick and mortar cardroom is ever shuffled as well and as thoroughly as we shuffle our cards. Each game, the deck is shuffled 10 times with each shuffle moving each card between one and 51 times throughout the deck. There is no bias to any card, any card patterns or seats at the table.’

However, the key rider here is unless they are crooked or contain bugs. In the early days of online poker, the shuffling algorithm for the ASF Software Inc. Hold’em games, used in at least three online cardrooms (including Planet Poker), was far from flawless. In September 1999 the Software Security Group at Reliable Software Technologies uncovered a means of calculating the precise deck being used for each hand, knowledge of which would have enabled unscrupulous cheats to know in advance the exact hands of every player, together with the future cards that would be dealt in that hand. Unfortunately, the ‘seed’ (or particular starting point) used for the ASF Software random number generator at that time was the number of milliseconds since midnight according to the system clock, which thus made it easily predictable once the RST program was synchronised with the system clock; and RST also identified other flaws in the shuffling algorithm. Of course, these problems were quickly addressed by the online cardrooms affected, but the fact that a shuffling algorithm had been cracked was very damaging for the credibility of the online poker industry as a whole at that time.

Nowadays every online cardroom is acutely aware of the need to reassure their clients that their shuffle is fair, random and unpredictable. A visit to the website of any of these companies will reveal the different ways in which they generate their shuffles. Clearly, the understanding and implementation of these complicated processes has moved on considerably since 1999, in particular with regard to the criteria used in selecting the seed used in random number generation, and shuffling has become more secure. However, that does not necessarily mean that it is absolutely foolproof at every single site.

Many newcomers to online poker, on experiencing an initial run of bad results, rush to blame their losses on unfair shuffling (or on collusion by other players) rather than questioning their own play. Undoubtedly it is possible to write software that gives the cardroom an edge in some way, perhaps by juicing the deck to produce more ‘action’ hands and thereby maximising the rake. The key question is whether it would be worth any major online site becoming involved in such a practice – their existing rake income is generally quite sufficient to keep their businesses running smoothly, so why would they risk such a practice in case they were found out (perhaps through being outed by a disaffected ex-employee) and lost their entire business?
Although it is true that online cardrooms survive by extracting money from their clients, it is in their long-term interests to do so honestly via the rake rather than dishonestly by cheating. Whether you are a long-term winner or a long-term loser, you will contribute more or less the same hourly rake over time, and companies with a satisfied strong regular player base (many of whom will then recommend the site to their friends) can expect large revenue returns for many years to come. Of course, it is not impossible that decks are being ‘juiced’, but there is no hard evidence to suggest that such practices are in operation.

One common complaint on poker forums and newsgroups is that players are drawn out on more online, with the online shuffle somehow generating more river cards that defeat made hands. Regardless of whether this complaint is actually justified (it may just seem like you are being drawn out on more because so many more hands are dealt per hour online than in a live-action cardroom, and so many hands are dealt with a full ten players seated at the table), being drawn out on is the natural occupational hazard of the winning player. One of the key advantages of most winning players over their fellow competitors is that they consistently enter the pot with better average starting holdings, so they will need to be drawn out on more often if they are to lose. In the meantime, they are being paid off handsomely on numerous other occasions by weaker players who are staying in the pot with poor odds, hoping to hit longshot draws.

Another common newsgroup and forum online poker myth is the so called ‘cash-out curse’. Many players seem to believe that when they cash-out from a site this will somehow lead to them being flagged by the cardroom and dealt a higher than average number of losing hands. This myth can be debunked in many ways, including the following:
•    At any point in time you are on either a good run or a bad run, with wins following losses and losses following wins. Inevitably, when you are on a good run you are more likely to cash-out and will subsequently appear to experience the ‘cash-out’ curse when the law of averages comes into play and your inevitable bad run arrives. When these players perceive that they are being afflicted with the cash-out curse, they are merely experiencing a natural regression towards the mean. Furthermore, those who are fortunate enough to maintain their good run after a cash-out will never have cause to report it; it is only those who lose after a cash-out who ever publish their experiences, and in so doing, perpetuate the myth .
•    When you experience a good run and cash-out, you may start to play in over-confident fashion, overplaying your hands and generally failing to pay as much attention to the game as you did before. Inevitably, this then manifests itself in a losing run.
•    Often when players cash-out they leave themselves with an inadequate bankroll, which is then vulnerable to the natural short-term swings of the limit at which they are playing.
•    There is no real financial reason why a site should dislike players who cash-out, but nevertheless continue to play at the site regularly. Whether individual players are winners or losers is basically irrelevant to the cardroom, so long as they keep coming back and thereby maintain their contribution to the rake.

To date no evidence has ever been produced to suggest that the shuffle at any online poker site is any way rigged. At Paradise Poker the shuffle is reviewed on a quarterly basis by Price Waterhouse Coopers. Using the log files provided, PWC have performed a series of statistical tests and affirmed that in their opinion every card has an equal chance of being selected. Furthermore, many news group posters have also independently analysed the hand histories from Paradise Poker and other sites, using samples of 60,000 hands or more, and drawn the same conclusions.