In theory, an unknown will "know" how to play pre-flop. This means that if he plays 20% of his hands, he'll play the top 20% only.
But one does not have to look very long at that statement to know that it is wrong, very, very wrong. For a couple of reasons:
a) It doesn't take positiuon into consideration. A player in a ring game can have a very high preflop raise % on the button and be nitty as hell UTG. This means that his hand range needs to be adjusted to position. While JTs is a perfectly fine hand to raise with in the CO in 6-max when everybody has folded, I don't know of many players that will do that UTG+1 for example. And if they do, take note and broaden their range.
b) Most poker tracker stats against a player are from a way too small sample to accurately put a player on a range of hands. You can gain a lot of very useful info (meta-game concerns such as overall blind play, if they limp often, if they check or bet postflop most of the time, etc.). This means that you have to take notes about certain behaviours and not only rely on the numbers. If you see a player open with JTs UTG+1 as was stated above, take note and adjust your play accordingly.
c) Pre-flop agression might be targeted more towards certain players than others. Let's say the guy is in position 5 and has two very lucrative fishes to his right. Both flat preflop. If he wants to get involved, most of the time, he'll isolate (unless he was "bought-in by the pot odds
) to avoid hemorraging money postlfop if he doesn't hit and still wants to clean the fish if the sharks to his left get in the dance.
So what I'm saying is basically: reads, reads, reads. Stats are good for overall behaviour, but when it comes to specific hand ranges or specific hands, take notes after you observe how they play (will slowplay AA, will try to isolate with a reraise with AKs, etc). Then, once you are on the flop and beyond, you weight all the options compared to previous actions and make the most sensible decision mathwise (usually).