Banner 468

Spotting the flaws in casino card-shuffling machines

(Image: Etienne Ansotte/Rex Features)

Ever feel like others at your poker table seem to have an unfair advantage? Like you were up against a team of expert card counters like the notoriously successful MIT Blackjack team?

Well, you might be, but at least you can take comfort in the fact that a team of statisticians from Stanford University, US, are now making things fairer. They recently discovered that cards dealt by an automated huffling machine  - designed to deal a random set of cards - were more than twice as easy to predict than a human-shuffled pack.

Automated card shufflers have been helping inept - and unlawful - dealers since the 19th century, though it's only in the last 50 years that they have been widely adopted by casinos. Now, though, the machines are sophisticated devices, rented to casinos for around $500 per month per machine - and with one on every card table, that makes for a profitable industry.

Statisticians help casino equipment manufacturers ensure that their shuffling machines spit out cards that are entirely randomly ordered to help foil cheats - and give the average punter like you or me a passing chance at winning a hand.

But, like poker hands, not all automated shufflers are alike, and, like most of my poker hands, some don't work as well as others. The latest trend in card-shuffling circles is the shelf shuffler: a machine which replicates the human riffle shuffle by randomly placing cards on one of a number of shelves, before re-assembling the deck by taking piles from the shelves, again in random order.

The team of Stanford statisticians was contacted by a manufacturer of casino equipment to test a new shelf shuffler, which was already designed and built. The basic tests to ensure randomness of cards coming out of the machine had, apparently, been carried out by the engineers, and the results seemed satisfactory. All the manufacturers wanted to do was double-check that a deck could be passed through the machine and then be used without concern in the casino.

They did not, however, bank on getting the hand they were dealt.

The high-rolling statisticians at Stanford took the task seriously. They set about testing the statistics on which the machine was based, and performed tests to measure how random the machine's results were. They found that a knowledgeable player could guess about 9.5 cards correctly in a single run through a 52-card deck from the machine, compared to 4.5 for a properly shuffled deck.

That boiled down to some cold, hard advice for the manufacturers: their machine's shuffling was not random enough to be used in a casino. In fact, the statisticians have reported that the the president of the company responded "We are not pleased with your conclusions, but we believe them and that's what we hired you for."

Whatever your game, stopping crooked dealers and card-counting players has to be a good thing. Just make sure that, next time you visit the casino, you sit at a table with Stanford statisticians - and steer clear of those MIT card counters.