Definition
Experimental probability
Experimental probability estimates chance from the results of repeated trials.
What it means
Run the chance process, count how many times the event happens, and divide by all the tries. The result is based on what was observed, so it may differ from a probability model, especially when there have been only a few trials.
Remember it
Count the hits, then divide by all the tries.
Example
A light flashes blue 17 times in 30 button presses. The experimental probability of blue is 17/30, which is about 0.57.
Watch for
Saying a fair six-sided die has probability 1/6 of showing 2 uses a theoretical model; it is not an experimental probability until trial results are counted.
Common mix-up
Experimental probability is not required to equal theoretical probability exactly. More trials often make the observed proportion steadier, not perfectly fixed.
Exact meaning
For repeated trials carried out under the same stated conditions, experimental probability is the number of times an event occurs divided by the total number of valid trials.