We all know about recurring villains. They’re fun, they save you chargen, they exist as reasons for the PCs to get better and as targets for them to go strike at when they might be otherwise unmotivated—in short, they’re blasted useful, so we use them. Good stuff. On the other hand, that means that every time they show up before the scheduled final showdown is a time in which this chain of recurrence can be broken. I’m not going to talk about keeping the PCs from killing your recurring villain. That’s a matter of mechanics and varies by system. Instead, I’m going to look at it from another angle. If you’ve made the recurring villain too powerful than the PCs, how do you make sure you aren’t breaking things on the PC end?
(Note that I am assuming right now a storyline in which a. plot matters and b. the players are also for the most part invested in the plot as it currently stands; strange though it may sound, sometimes players will decide that a plot sounds interesting and attempt to render it doable. If my GMs, particularly the inexperienced ones, give me a choice in the matter and ask for help, I’ll cheerfully facilitate their plots up to the point that the Idiot Ball or acting OOC gets involved, at which point there is negotiation. I know perfectly well that there are people who think a GM should just let the dice fall where they may, even if one does end up with a Catch-22 leading into a TPK. If you’re one of those people, then I have other articles for you; please don’t start an argument in this one.)
On the whole, that should be pretty easy; there are plenty of ways to justify a villain sparing PCs. The villain’s emotions—hubris, unexpected attachment, whatever—or overarching plan leading him to decide it’s better that they be left alive, either on their own or as captives. Winning but only thinking they’re dead (which, needless to say, works better in lethality-optional systems like FATE or Legends of the Wulin, since it’s remarkably easy to kill someone by accident in, say, Exalted and near impossible to disable someone without killing them in D&D—unless the character is doing it on purpose, which breaks suspension of disbelief). Having an objective which is completely independent of the heroes’ lives, completing it and leaving. Being brought down to a certain point and deciding that right now, it really isn’t worth the associated risks. Deus ex allied NPC (please use sparingly).
Likewise, on the PC end, there’s an answer to this: run like heck in the opposite direction. If you can do a proper job of telegraphing to the PCs that they are outmatched, thank you very much, and if running seems possible, they might do it—at least, presupposing there isn’t some equally important reason why they feel that they can’t.
The problem is making sure, going into a fight, that whichever out you’re planning to take will actually work with your other motivations. If the recurring villain wants something from one of the PCs, then leaving them for dead leaves a massive plothole; if there wasn’t a cliff, raging fire, or similar unsurvivable situation involved, why didn’t he follow up and take what he was looking for? Likewise, if you can’t justify anything but the PCs running away, and there’s a good reason why they consider running away not to be an option, you’re going to end up at either a massive impasse or a TPK. So if you really want to have an early recurring villain run-in, then make sure you know exactly what both sides are trying to get out of whatever situation you’ve thrown the villain into, and that whatever plans you have take into account both sets of motivations before you throw in Antagonist the PCs Can’t Handle Yet. Otherwise, a plothole may be the best possible result.