- Joined
- Nov 23, 2008
- Messages
- 1,683
- Reaction score
- 4
- Points
- 38
The gambler's fallacy is that if there is a string of events of one result then the next event must be the reverse "to even things out". That is not the case if the events really are independent.
But you can still multiply those chances together to get the chance of them all occurring together assuming they are independent. For example, you can calculate the chance with a fair dice of getting sixteen consecutive 7's as 1 chance in 6^16, which is 1 in several trillion. Still even if there were fifteen consecutive rolls of sevens, if the dice is fair the chance of getting some number other than 7 is not increased on the next roll.
In reality though if you did observe that string of dice rolls that had trillions to one odds against it, you'd assume the dice was loaded. That assuming the dice is loaded means you are looking for a physical explanation.
Bob Clark
people explain the fallacy this way, but the real fallacy lies in the failure to see that all outcomes are equally unlikely (or all equally likely).
Any string of dice rolls has trillions to one odds against it. If I told someone they can win my game if they roll the exact sequence of 14 numbers that I write down, their odds of winning will be just as terrible no matter what sequence I write down. This is the idea behind the lotto type games. The people who run it rely on huge amounts more money being spent on tickets than is being payed out to winners.
I would only suspect a loaded dice if after hundreds of dice rolls(large data pile), the dice did not statistically resemble odds of outcomes, meaning there would be more sevens than twos. However at their core, the odds of getting numbers with dice is really based on each dice having a 1 in 6 chance of getting any certain number, the better odds for 7 only come about because of the way we count the pipets. If we made the different 7's that can be rolled unique, then they would all have the same probability as rolling 2. The rules of the game skew the way we perceive the surfaces of the cubes and also the way that people tend to think about probability. When two different outcomes are considered equal by the game, people forget that rolling 4 and 3 is different than rolling 3 and 4 and both are equal in probability to rolling 1 and 1 or 1 and 6.
When it comes down to it, we simply do not have enough data to know how many close calls and impacts are statistically normal or abnormal. All we know is that compared to our short lives and history, large impacts are what we would consider infrequent and tiny collisions (super tiny) happen on a second by second basis.