Scott, when you consider how few players at chess.com play in official OTB tournaments, astronomer's figures may not be far off.
Once upon a time, it was impossible to get an initial rating below 2000 in FIDE. Simply being FIDE rated meant that you were already at that threshold.
I don't know the break down of FIDE players. In the USCF, a player who is +2000 is in the top 96.94th %ile. Again, that only includes players who have played in official USCF tournaments, which is a pretty small selection of players.
http://archive.uschess.org/ratings/ratedist.php
I honestly don't know whether astronomer is correct or even how to prove it.
The answer to the original question is clearly "no". The way the Bell curve of the rating system is defined, a rating of 2000 corresponds to the top few percent of players. Clearly the 50th percentile (a median average player) will always be distinct from this.
"top few percent"?
How about top few hundredths of a percent. You're looking at a group that is clearly far above the norm
2000+ is 66,611/154,786 FIDE players, so not even close to top few hundreths of a percent, not even top third.