JavaShuffle!

Sorry, you need a browser that supports Java.

Here is a set of hands, ready to be used for your next event. The hands were generated when you started the applet, no two pages will ever be the same.

Need more hands? Simply press the "Reload" Button and another set will be generated.

To print the hand records, press the "Print" Button". Unfortunately, on some Netscape 3.x versions, this results in a blank page. Netscape has promised to fix that in the next release.

Want a different layout? Feel free to copy the page. Then edit the applet parameters and load the page into Netscape. The applet parameters are explained in the comments of the HTML code for this page, the parameters allow you to adjust the number of hands, the number of columns and the size of each hand.

About JavaShuffle

People always complain about those Computer Hands so I did a bit of testing. On a sample of 100,000 hands (approximately 15 minutes of CPU time on a Pentium-90), I found that:

Each player held, on average, 10.000 HCP, where the number varied with less than than 0.005 HCP for the 4 seats.
Each side held, on average, 20.000 HCP.
Each player averaged 3.25 cards in each suit.
The number of times that each player held each of the 31 most frequent shapes (4432 up to 9310), was well within 1 unit of the statistical error expected from the sample size. (For example, north held 21536 hands with a 4432 shape, where one would expect 21551±68.)
The number of 9400 and hands with 10 cards or longer suits was within one sigma too.

There are more tests that one can do and I might even do them sometime, but at a first glance, the hands seem to be consistent with the expected values for perfectly shuffled hands.

The algorithm uses a 48 bit random generator and it can generate some 2.8x10¹⁴ different hands after each download. While this isn't anywhere near the 5.3x10²⁸ possible bridge hands, it will be a while before you see the same hands reappear.

The random seed is determined from the IP address and the time of the day. This makes it extremely unlikely that the applet is ever run with the same same seed (you'd have to run the applet on the same machine at exactly (in milliseconds) the same time) and excludes the possibility that two sets of hands will ever be the same.