

Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
The randomness of computer-dealt hands in comparison to hand shuffling in card games. It discusses the methods used by computers to deal cards randomly and the importance of ensuring randomness. The article also addresses common complaints about computer-dealt hands and provides statistics on hand distributions.
What you will learn
Typology: Lecture notes
1 / 3
This page cannot be seen from the preview
Don't miss anything!
For those of you curious about the randomness of hands from the Dealing Machine vs. hand shuffling it is very interesting.
Computer-Dealt Hands—Truly Random or Impishly Fixed? by Dave Willmott
Computer-generated hands have a bad reputation in some circles-undeservedly so. “These computer hands!” is a common cry, but how different are “computer hands” from people-dealt hands?
Let’s tackle the second question first. It is a simple matter for a computer to deal a set of 52 cards. To explain how this process works, using an oversimplified analogy, consider how a postman places letters in a set of sorting office boxes. Each piece of mail has an address, and the postman places each piece in the correct box. Similarly, each card of the deck is randomly assigned a value which determines which “box” (hand) it will be dealt to.
There is a catch, however, and it has to do with the word “randomly.” It is important that the computer performs this task in a random fashion. Otherwise, there is a risk that the same set of hands will be generated more than once. Many computers use their internal clocks as a “random” starting value to assign each of the cards an “address.” In practice, however, this method isn’t sufficiently random enough. The software that is used is a random number generator and uses a linear congruential algorithm that will repeat after 2 to the 47th power (140,737,488,355,328) deals have been generated.
Returning to the question of whether computer-dealt hands are different from people-dealt ones, remember the great emphasis placed on creating a random deal. To create a random deal manually, players must give the deck a thorough shuffle at least seven times to guarantee that the cards are mixed. It is common, however, to see players shuffle the deck only two or three times before dealing. This is not a random deal.
It has been suggested by various mathematical authorities that this inadequate shuffling may lead to “flatter” distributions. In other words, suits may break evenly more often than expected. Players who are accustomed to their trump suits dividing evenly all the time because of inadequate mixing seem genuinely offended when they encounter a 4–1 trump split in a computer-dealt game, even though that division will occur 28% of the time.
One riffle shuffle reorders the cards significantly. But the reordering is not random, by any means. The top card is still near the top. The bottom card is still near the bottom. Most pairs of cards are either still adjacent to each other, or not very far apart. And so, several shuffles are necessary. How many shuffles?
Mathematics shows that at least seven shuffles are needed to randomize the cards. More shuffles than that do not affect the randomness. Bridge players often complain about the poor hands that result from computer dealt cards. This would seem to indicate that these people are used to inadequate shuffling.
Another factor which causes some players to believe that computer-dealt hands are different from people-dealt ones is that hand records are usually available after a computer-dealt session. Players then have the opportunity to see all the cards of every deal, and some of these players will use the hand records as “proof” of the “strange distributions” supposedly caused by the computer. If these same players could see hand records from their local club games, they would find that “strange distributions” occur in hand-dealt games as well.
Sometimes players complain that the cards seem to run one direction (usually the direction of their opponents). “North-South had all the cards!” This does occur sometimes, but it happens in people-dealt games as well. George S. Kaufman, the famous New York dramatist and bridge expert, once quipped that clubs should post which way the cards are running as a courtesy to the players. He said this, however, during the 1930s. People have been complaining that their opponents hold all the cards since the days of whist. Finally, let’s not forget the advantages of computer dealing. Without computers, hand records would not be available at sessions. It would be too time-consuming to copy all of the information from people-dealt hands.
The choice for our club players is simple – computer generated truly random hands or hand shuffling of doubtful quality. Naturally, bridge players (especially newer players) are very suspicious of computer-dealt bridge hands. Their suspicion is somewhat akin to a superstitious belief that pre-dealt bridge hands must be set up to give them troubles.
As with duplicate games in which the boards are well shuffled, you will have occasions when the hands in play are more distributional than usual, or more flat than usual, or any number of other coincidences that are perfectly normal for such a tiny selection of truly random hands. In my personal experience, I don’t believe I’ve ever attended a club game where someone didn’t say to me, “The hands have been really strange tonight!” That’s a sign to me that the hands are perfectly normal.
If you look at the table on the last page it shows the normal distribution of hands over 20,000 both actual and expected deals. Add up the number of hands with a singleton or void and you find out that about 1 in three deals will have a singleton or void. Note that the 5-4-3-1 hand is the third most likely hand.
So adding up the statistics for a given deal: 64.3393% of the hands have at least 2 cards in each suit. 30.5542% of the hands have a stiff in at least 1 suit 5.11064% of the hands will have at least 1 void.
Note that this is an average and on a deal of only 36 boards you can expect a significant variation from the average. Using these statistics says that there should be one hand with a 10 card suit in 30,000 deals.
Contrary to rumors we do not select hands or throw out flat or easy hands. It is way too much work to do that. We have a choice of dealing software but most clubs use the one that uses the most number of bits and randomness available. It is called “Big Deal” and uses the real time clock and mouse position to