Knowledge 135 The 80 characters of this poem versus the 52 playing cards and the greater choice offered by 26 letters increases the problem geometrically. Taken together the probability of accidentally getting this poem is vastly less than a perfect hand of bridge, 1 in 10® against the perfect bridge hand of | in 10”. That’s the difference between the number of atoms in the known universe and the number of atoms in a jug of water! Numbers get big very quickly when we are looking at the permutation of information. And there is another problem with our bridge analogy. All the bridge players in the world are part of the machine finding the perfect hand. When a human sees a perfect bridge hand they are amazed. It is an event that usually hits the local newspapers and a couple of years ago one reached the national papers in Britain. Each bridge player looks at every hand, they play so there is a huge amount of processing going on during every bridge game. To replicate this for our poem, we would need millions of poetry classes spending hours each evening reading through computer printouts of gibberish. I should also add that sightings of perfect bridge hands are almost certainly hoaxes. The probability of it happening even once would require everyone on Earth to play bridge continuously for a thousand years. It is reported somewhere in the world about two or three times a year. If we are charitable, we might assume people failed to shuffle the deck properly but I suspect some mischief is going on! The numbers don't stack up... You might think the problem is one of improving the efficiency of the filter so humans would only have to examine a smaller number of possibilities. Surely I could improve things by writing a simple program to ban all non-English characters, words and poor grammar; things that don't pass the Microsoft Word grammar checker. This would generate a more manageable number of potential poems. Lewis Carroll shows this does not work; my idea to use a g