The theoretical probability is based on all possible outcomes. There are six sides with equal likelihood so the probability of each number is 1/6 (or 0.16666..)
Experimental probability is based on an actual experiment. So you might roll the die 100 times and then count the outcomes for each number.
So if you rolled 20 ones, 14 twos, 17 threes, 15 fours, 16 fives and 18 sixes the experimental probabilities would be:
One = 20/100 = 1/5
Two = 14/100 = 7/50
Three = 17/100
Four = 15/100 = 3/20
Five = 16/100 = 4/25
Six = 18/100 = 9/50
As you can see, the experimental probabilities don't usually line up exactly with the theoretical probabilities.
As for computing an experimental probability from one roll, that's not going to be very helpful.
If you roll a four, you'll say that its experimental probability is 1/1 (100%) and all the other numbers have probabilities of 0/1 (0%).
Basing probabilities off of one trial isn't a good idea.
Answers & Comments
Verified answer
The theoretical probability is based on all possible outcomes. There are six sides with equal likelihood so the probability of each number is 1/6 (or 0.16666..)
Experimental probability is based on an actual experiment. So you might roll the die 100 times and then count the outcomes for each number.
So if you rolled 20 ones, 14 twos, 17 threes, 15 fours, 16 fives and 18 sixes the experimental probabilities would be:
One = 20/100 = 1/5
Two = 14/100 = 7/50
Three = 17/100
Four = 15/100 = 3/20
Five = 16/100 = 4/25
Six = 18/100 = 9/50
As you can see, the experimental probabilities don't usually line up exactly with the theoretical probabilities.
As for computing an experimental probability from one roll, that's not going to be very helpful.
If you roll a four, you'll say that its experimental probability is 1/1 (100%) and all the other numbers have probabilities of 0/1 (0%).
Basing probabilities off of one trial isn't a good idea.
1 in 6
any number will come up 1/6 th of the time.