Let's say you're merely pretty good at darts, because you didn't waste your youth becoming great at it (you wasted it on other things). Then you run into trouble, because missing the triple 20 means you're just as likely to land on 5 or 1. That's why the board is designed that way -- it's risk versus reward.
So is that when you aim for the bull's-eye instead? Science says no. Forget about the goddamn bull's-eye, it's a red herring. Mathematicians ran some simulations of dart games where the player's skill level is assigned a value, with a lower number being better (it wouldn't be science if it wasn't a bit confusing). When they plugged in the numbers, the graph of the best place to aim according to your skill level looks like this:
They had similar conclusions in their urinal aim chart too.
If you're a great darts player (what they're labeling levels 1-16) and can very reasonably expect to hit triple 20, go for it -- the occasional miss won't cost you much. But if you're slightly below expert (anything above a "17" on their scale), the best place to aim switches to the lower left, starting at triple 19 and moving up toward the center from there. As your aim becomes more erratic, throwing at this line maximizes your chances of hitting something worthwhile.