By “enough repetitions”, I meant that it makes more sense to use a straight expected value calculation the higher the ratio of the frequency to the inverse probability. So let’s suppose you’re playing Russian Roulette with 12 chambers. The inverse probability of dying would be 12 and lets suppose you’re going to be playing it 12 times (assuming you don’t die first). That is a reasonable ratio so it makes a decent amount of sense to use the expected value. Not that the ratio did not involve the number of times you died, but the number of times the game was played.
By “enough repetitions”, I meant that it makes more sense to use a straight expected value calculation the higher the ratio of the frequency to the inverse probability. So let’s suppose you’re playing Russian Roulette with 12 chambers. The inverse probability of dying would be 12 and lets suppose you’re going to be playing it 12 times (assuming you don’t die first). That is a reasonable ratio so it makes a decent amount of sense to use the expected value. Not that the ratio did not involve the number of times you died, but the number of times the game was played.