Theoretical probability: before running an experiment, figure out what the odds of something happening *should* be.
Experimental (simulated) probability: do an experiment a bunch, and figure out what the odds of something happening *were*.
Experimental probability should get closer and closer to theoretical probability the more you do it.
* If you roll a 6-sided die, what’s the chance of rolling a 1? Well, the theoretical probability is obviously 1/6. OK, so roll it six times. You don’t roll a 1! Gasp! Shock, horror! Was your theoretical probability wrong? Well, no, you just failed that 1/6 chance six times in a row – sometimes that just happens.
* Roll it six more times, and you get a single 1. Now your experimental probability of rolling a 1 is 1/12. Still not too close to the theoretical 1/6, but closer than the 0/6 it was a moment ago.
* By the time you’re up to 60 rolls, you’re probably getting closer to 10/60. By the time you’re up to 600 rolls, your experimental probability is probably really close to 100/600. By the time you’re up to 6000 rolls, you should be really close to 1000/6000, and if you’re not, then that’s a sign that there’s something wrong with your test (like a badly lopsided die).
Latest Answers