So am I wrong to think it is 3 because the theoretical chance is 1/3?
Again, I'd say the theoretical value and the number of times it must be repeated before, in theory, it would occur. Is this not the correct way to think statistically?
The final formula IS simple like that... but the reason is not simple.
To compute the average throws, you need to calculate the chance of it happening in exactly 1 time, multiply it by one... then sum it to the chance of it happening in 2 times multiplied by 2, summed to the chance of it happening in 3 times multiplied by 3... and so on to infinity.
So, basically, you have to find the solution to a series.
Such solution ends up being an elegant 1/p