But you're using information from the future if you will.
Most prodigies (and there are lots of them in chess) don't keep progressing to IM, GM, super-GM. You can climb 500 points in a year or two and then level off for a while (or the rest of your career).
That said, when the kids are still young (like 13) it's usually safe (as their opponent) to add about 200 points to their rating in your mind, just so you're not caught off guard.
Does ELO always accurately predict strength, and is there another metric (that of rating trend) that could be used to assess the likelihood of a win/loss?
Consider Carlsen when he was 13, and his rating was 2450. It's clear his rating is going up fast, so it's likely he is in fact stronger than 2450, based on this trend.
Consider then someone rated 2450, who has been in decline for many years, perhaps an ageing low level GM who never got much above 2500. Surely in this scenario you'd put your money on Carlsen, but the ratings suggest it's more likely to be a draw.