Previous Page

Table of Contents [ Show]


Thinking Like a Computer

Upon this realization it becomes a really simple algorithm:

  1. Start counting the deck from 0 to 51, and set a counter for each hand starting at 0.
  2. Select a random player from 0 to 4
  3. If the player's hand is full (ie. their counter has reached 13), then repeat the previous step.
  4. Place the current value of the deck counter into the position of the counter for the selected player.
  5. Increment the counter for that player and the deck.
  6. Repeat from step 2.

Efficiencies

If the hand is full in step 3, there is is nothing stopping it from trying the same hand again when it returns to step 2, but the process is so quick that it will eventually find a free hand much quicker than the time the shuffle and bubble sort took. I did a bunch of stats to find out how many wasted cycles this would use but I won't get into that here. The point is that the first hand doesn't fill up, on average, until around the mid-fourties. I have a gist that does an equivalent calculation. The average number of attempts to deal to an already-full hand is a bit less than 20 per deal. It would require many more cycles to first check to see if the hands are full than it does to just fail and retry. This does conflict with my desire to find creative and elegant solutions, however.

This method is already faster than just the process of sorting and deal, but this also means that it doesn't then have to sort the final hands either. Because the cards were delt in ascending order and the slots within the hand were also filled in asceding order, the resulting hands also end up in ascending order automatically.

I will note that this solution is perhaps only the most efficient in games where the whole deck is dealt. If only a partial deck is dealt then suddenly the previous method becomes more efficient. No matter what, you would need a random selection of cards that are to be dealt before distributing them. Running the previous shuffle for only the necessary number of loops will do exactly this. Then it is faster to deal to the hands in order and sort after the fact than it would be to sort and then deal. This is true because: A) There are no failed deal attempts. B) It is faster to sort multiple smaller piles than on large one. This is the principle behind the "merge sort", a much more efficient sorting algorithm.

The moral of the story is that people are smarter than computers in a lot of ways, but or intelligence can sometimes get in the way. To avoid cheating in a card game we go through a process that is completely counter-productive every time that the cards need to be dealt. Computers don't need to think about what abusing their position could mean. In this case it means that they are can be a benevolent dictactor. It is also the reason everyone is terrified of a generalized artificial intelligence with poorly defined goals. I'll blog about that when I have a solution...