I think 2000+ people may have some knowledge of a broader array of middlegame positions, while 1800 level may be more specialized and play significantly worse when out of their comfort zone.
What makes a Class A - Expert Player?

I think the most prominent feature of anyone above 2000 is consistency. Even if they aren't playing fantastic moves, they rarely completely crack with a blunder that can be refuted in two moves. In other words, they are telling you that in order to beat them, you have to find a creative way to do it. I think they would be more able to finish off positions -- they can often convert extra pawns in the endgame, and will insist on not making a huge tactical mistake to let their opponent back into the game. In other words, they are starting to have some decent technique, which is almost nonexistent in anyone under 2000 (including myself).
They will, of course, still end up doing some of those silly things, but much less often than weaker players. Anyone under 2000 can easily crack, even if they have played well for the first 40 moves! Experts are much less likely to do that, at least without being put under strong pressure. Less unforced errors, if you will.
All this said, Estragon's post #20 is more to the point, and based less on interpretation. Make less mistakes, play more accurate chess; what skills they are particularly great at can vary, because chess players have different styles, but they usually have to be pretty good at every part of the game, else they would have glaring weaknesses.

I think the most prominent feature of anyone above 2000 is consistency. Even if they aren't playing fantastic moves, they rarely completely crack with a blunder that can be refuted in two moves.
last year a strong highly rated junior did exactly that against me.
I think that was psychological, as 1) I'd been defending all game and he'd had all the chances, and 2) The rook had a whole rank to move on so you don't expect it to get trapped. But that's the only time a 2000+ player has fallen for a "simple" tactic against me, in general they only lose by positional errors.

Each player up to super GM has weaknesses in his game.
The trick is to understand where you are weak and how to improve.

piphilologist -- that's a heartbreaker for him, but must have felt fantastic for you, to ironically be the one to outlast him, even though that was your opponent's very strategy!

piphilologist -- that's a heartbreaker for him, but must have felt fantastic for you, to ironically be the one to outlast him, even though that was your opponent's very strategy!
I felt very bad when later his mother told me he hadn't realized the game was rated though (neither of our opponents had turned up and we were in different sections so got paired together as a "friendly")

I'd like to reach 1900 or so. I once reached 2349 on the tactical tool, and about the same on Chess mentor, but then dropped back to 1700 or so. I could not maintain the high level.
Experts can play some bad games. I once beat one, playing 6th board for our chess team, at a time when I was only rated around 1500. I did not play well that day; he just played worse. I've played better games and lost.

While I can say that I have not played in any official chess federations, I can tell you how I managed to improve my overall style of play and my abilities overall to the point where I have had solid wins and draws over both Class A and Expert players and held a solid Spanish Game as Black for around 30 moves against a FIDE Master that comes to the coffee shop I play at. Namely it has been due to:
1. Openings: As a overall rule I tend to play lots of closed games and so normally play 1. d4 as White almost entirely now. Rather than soak up vast amounts of theory and lines, I tend to look at the ideas behind the openings. I know that if I play cxd5 I might help Black with his c8 Bishop and steer away from those lines. Aside from offbeat lines or Benko games, I do fairly well. As Black, I play e5 to e4 and d5 to d4 since barring a Gambit opening, I can set the variation.
2. Middle game: I tend to study tactics a lot and GM Alburt's book is one of my favorites. The tactical themes are in no set order just like the game play is not. I also do not try to go for cheap tricks and instead just try to play good solid chess.
3. Endgames: I regularly go through Dvorestky's Endgame Manual because while all games have an end, not all have an endgame. If I get someone into a decent ending, I will win or draw more often than lose now. e thing I noticed is that many players, including 1800-2199 players and some NM/FM Players are lacking in endgame play here in the US. Likely because they have not studied them much and since most books in the US show you how to win with openings and tactics.
4. Patience: I have 30 minutes on the clock so why play like it is 3? Now I am able to look positions over clearer and while not always finding the best move I can find good ones s lot of the time now.
5. Type of games played: I almost never play Blitz anymore because I tend to play sloppy in them and I don't see the board as clearly. Now I play OTB and CC games and I am getting better. My blitz here is mid 1600s and Online at 1600 from 1400 and 1500 respectively. My slow game is better so I focus on this now.
5. Learning from losses: this has helped immensely.
6. Playing superior players. I tend to play my best against higher rates because I am forced to. Against lower rated I take too many chances.
7. Confidence: I may not beat everyone all the time but I know at anytime anyone can be beat so I not only enjoy playing but I play as if I can win instead of thinking I will or have to win. In bad or lost positions I will make an opponent prove their win unless. I am busted.
To illustrate: currently in the 2 tournaments against members of my group. I have first place in both events and need one more win in one of them to seal it. Ratings are from 1000-1750. I have not played this well though I have been told I would be a solid Class B if not lower Class A USCF based on how I have improved in the last three years.
Ultimately I would assume the difference in an Expert in a Class A falls into the same way as well.

Having played against both I will say that Experts can have flawed strategy though frequently play solid moves. There comes a point when they just hand you equality or even a clear advantage if things are sufficiently sharp. The worst thing you can do is assume a move is sound just because the guy who played it is over 2000. The board contains the same information whether your opponent is 1000 or 2000 and has the same laws regardless of rating. If you for example play good moves in the opening not even Kasparov can refute your choice because the Catalan, Paulsen, Archangelsk Ruy Lopez, etc. are correct whereas if you play the Englund or Latvian against a 1000 rated player he can potentially win by force as the potential is there for him to find the best moves.
Basically experts aren't A players on steroids although they do assess, evaluate, and calculate better, but the main thing that seperates them is thinking process, time management, and strategic understanding.
What I really want to know is what seperates experts from NMs. Heisman's Improving Chess Thinker sheds some light and there's a contrast between expert and FM but I can't seem to put my finger on the essential difference I sense is there. Protocol M-3 gives me hope that I can make FM someday since I too saw the ...Kg7 and ...h5 plan with the idea of activating the rook.
De Groot B was the hardest position in that book for me though it was still unclear. The two bishops restrict pawn chains nicely, and two bishops are worth more than a rook especially if they're very active and pointing at the king, but seven pawns against four is never fun especially when the pawns are solid and hard to attack.

What I really want to know is what seperates experts from NMs.
Seems to me a big distinction between class players is what it takes or how often they get confused with what's going on in the position. e.g. which features are important and which don't matter. Of course a low class player never knows :p
To me, you have to make the position sharp / imbalanced to confuse the master. Otherwise they're really not going to blunder or get confused where their position starts deteriorate for no real reason.
Of course that's just me. I'm sure a GM would say masters are clueless and make big mistakes all day long.

Having a rating 1800+ & 2000+. Hope this helps.
lol
And so the Dredging Up A Dumb Topic trophy goes to: OogieBoogie.

What I really want to know is what seperates experts from NMs.
Seems to me a big distinction between class players is what it takes or how often they get confused with what's going on in the position. e.g. which features are important and which don't matter. Of course a low class player never knows :p
To me, you have to make the position sharp / imbalanced to confuse the master. Otherwise they're really not going to blunder or get confused where their position starts deteriorate for no real reason.
Of course that's just me. I'm sure a GM would say masters are clueless and make big mistakes all day long.
Sounds reasonable at first, but higher playing ability by definition means they'll be less likely to make a mistake and our mistakes are more obvious to them than theirs are to us. Whenever I go through an under IM game from the Big Database 2013 I analyse and assess positions then run it through a computer to see how close I was to the mark. If there's a mistake I try jotting down what principles such a mistake violates. Maybe it's a forced sequence that at best gives the opponent the critical e4 square for his king in a certain endgame or recapturing the wrong way whereas the other pawn capture gives him serious chances as it's a candidate passer?

lol

Having a rating 1800+ & 2000+. Hope this helps.
lol
And so the Dredging Up A Dumb Topic trophy goes to: OogieBoogie.
I have an idea of what seperates them but found this topic because I'm looking for a definitive article on it. Maybe Heisman or Dvoretsky have an article lying around that explains the most essential differences and the most critical areas of study.

According to Yermolinsky, the main difference between experts and NMs is that NMs 'handle the transition between opening and middlegame better', which I interpret as NMs know better how to play the various pawn structures stemming of their opening repertoire (might be a stretch, but I think that's the idea anyway).
I notice that above 2000 openings become far more important and you can't get away with a bad opening against an expert.