Note - the reason for switching being the best method is essentially this: you have a probability of 2/3 of being wrong on your first guess. Whenever you are wrong, switching will make you win. So, if you always switch, you will win 2/3 of the time. Much more information with various methods of explanation is available on Wikipedia.
Some interesting statistics:
| Total amount of people who have participated: | 8625 |
| Percent who got the car | Percent who got a donkey |
| Those who switched | 66% | 33% |
| Those who stuck | 33% | 66% |
| Percent of users who did this |
| Switching | 56% |
| Sticking | 43% |
Some less interesting statistics:
| Door | Percent who picked on first choice | Percent who picked on final | Car was here |
| 1 | 34% | 32% | 33% |
| 2 | 41% | 34% | 33% |
| 3 | 23% | 33% | 33% |
How the experiment is set up
The car is assigned to a random door as you load the first page. Everything is randomised as much as possible. If two doors could be revealed a coin is tossed to decide which one to reveal, the buttons to stick and switch are placed in a random order so users do not choose them because of their placement.
Contact me
You can email me if you spot a bug or anything else. You can also visit my blog.
Update: it is now on Digg.