The USCF ratings committee is against the usage of rating floors. They were created only because of tournament (money) considerations. The timeout issue you speak of was fixed a long while back. If you have 100 games in progress, and time out in all of them, only the first 5 will negatively affect your rating. After that, the opponents will gain rating points but you will not lose any.
This has happened to me more than once. I am playing a slightly lower rated oponent and as the game goes on the guy has me on the ropes. I use my wit and check up on the opponents account to see if I can see some of his games he has played in the past to see if I can dig up some dirt (steer the game into something that seems to discomfort him). While on my dirty deed I am horified to discover that this player is a "Sleeping Dragon" A guy who has a peak rating 200 or even 300 points higher than his current rating and with this my illusion of winning has become but a faint shadow of its former self. My point is perhaps chess.com should implement an anti sand-bagging system to not allow this to occur, something along the lines of "If a player has held at any time a peak rating of 2000 then their rating cannot go more than 100 points bellow this. (P.S. with this I single out no one as sandbagging for cheating purposes [most of the "sleeping dragons" are a result of timeing out in a-lot of games all at onece])