It depends, I think. With a 2 hours / 40 moves game, the 2000 player with an extra Knight (or Bishop) would win, I am quite sure.
But with a 3/0 game, or something like that, Kasparov would win : it is very easy to make a lot of errors when you play quickly.
It is also very easy for a pro, to create a situation where there is no apparent way to make progress - making you think and lose time...
I suppose that 2/40 does give better chances for the 2000+ player in this situation - one can just sit down and calculate and figure things out with calm.
However, I wouldn't say that winning is a foregone conclusion, by a long ways.
I just crushed H3 Pro with a queen advantage, just to test this. Very easy win even for a 1000 Elo player like me lol.
With perfect play, you can win more pieces and gain an even bigger advantage. If not you can just trade off everything, although there might be a draw.
You might lose though if you fall into an early tactical blunder. Otherwise it is pretty hard to lose with solid play.
kasparov will got mating attack quickly
positon matters as well, not just piece advantage. If the 2000 ELO player has position advantage as well, he/she could win but without that, i doubt its possible.
Kasparov would still crush the 2000 player!
THIS IS THE GHOST OF ROBERT FISCHER, POSSESSING THE BODY OF THIS KID TO TELL YOU THAT I HAVE NOW SOLVED CHESS AND CAN BEAT LITERALLY ANYBODY DOWN A QUEEN AND THREE PAWNS.
Or not.
Yeah, if the 2000 guy will be the future number 1.
Righ question is, what rating player is equal as Kasparov without piece.
Depends on the time control, and which piece it is :-) The longer the time controls the more the piece advantage will show, but anyone thinking a piece advantage is enough for a weak player to beat a strong player should try playing a bunch of fast games as black against a strong engine and remove Nb1 or even Ra1.
Wild guess, Kasparov in blitz without Nb1 = 2200, in rapid 2100, in classical 2000.
Fischer claimed he would beat Gaprindashvili with piece odds, and some even believed him, but that was a very strong opponent that could score results like these in Fischer's days, and that without piece odds:
http://www.ajedrezdeataque.com/11%20Ajedrez%20Femenino/Torneos/Olssons_67.htm
I think that I could beat Kasparov if I had a Queen, two knights, two bishops, and two rooks odds. Other than that, though, I'd be up a creek without a paddle.
Kaufman will lose pretty easily against a 1800 player with knight odds, me too, of course. Talking about 1800 FIDE, not USCF, which is inflated by 100-120 points, at least.
My understanding was that the inflation was at the high end, up to about 100 points, but gradually tapering off so that there isn't much difference say below 2000. We can easily see the 'inflation' at the high end (last time I checked anyway) just by looking at the top 100 lists. But this is difficult to do at the "low" end because not many USCF players under 2000 have established (non provisional) FIDE ratings and many of these are improving kids. I hope I'm not mis-stating the following but I think Glickman's attitude was something like why should we try to align USCF with FIDE, they're two different systems. By the way, the USCF Regular and QUick ratings are similarly misaligned - you see very few players with an equal or higher Quick rating, most of the time nearly 100 points difference. This can be seen from the USCF online ratings, just pick any random section of the alphabet. The USCF is mum about this.
USCF uses the following formula to convert FIDE ratings to USCF.
USCF = (
720 + 0:625 FIDE if FIDE< 2000
-350 + 1:16 FIDE if FIDE 2000
By this logic, here are the USCF and FIDE values for 1500 to 2500 FIDE.
FIDE USCF
1500 1657.5
1600 1720
1700 1782.5
1800 1845
1900 1907.5
2000 1970
2100 2086
2200 2202
2300 2318
2400 2434
2500 2550
So 1800 FIDE is officially equal to 1845 USCF, which is not a 100 point inflation.
In the under 1700 range (and even in the under 2000 range), I would expect that there is little correlation as most US class players play in only a few FIDE rated tournaments per year at most.
Combine a selection of low data to go on and improvement in ability without games to raise the ratings, and many US players will have an extremely inaccurate FIDE rating.
In fact, since August of 2012 I have gained 409 rating points in USCF, while only gaining 4 rating points in FIDE.
The main problem is that USCF takes all US games into account, while FIDE only accepts games that are
a) in a FIDE rated event
b) against another FIDE rated player.
Even in FIDE rated events of 7 rounds, I may only have 3-4 games that are actually rated by FIDE.
Elubas, they speak here of FIDE 2000, not USCF 2000...
FIDE 2000 = USCF 1970, i.e. not much different but actually slightly higer FIDE than USCF at this end of the scale. This is reversed at the IM/GM end of the scale where the USCF ratings are higher than FIDE, but on average less than 100 points. The conversion formulas are given by Glickman based on all available data, but also see Shepi13's two insightful posts above.
Ok, those are the formulas, and I've seen such before, but can you give me any idea on why they would work that way? Why are ratings inflated on the higher end (e.g., someone with 2600 USCF may have 100 points less for their FIDE rating), but not at, say, 1900 level as you have just shown in your example? Does it somehow get easier to raise your USCF rating when you get to IM level or something?
I guess one problem with national rating systems such as those of USCF is that they, naturally, contain basically all American players and so forth for any other national system. But Americans on average are probably not as strong as Russians, so to get a certain USCF rating (do well against other Americans) may require less actual chess skill than it would to get the same rating under a (hypothetical, at least) Russian Chess Federation.
I can't say who would win but if i had to bet on it I would chose Kasparov.
I assume most ~2000 female players could have beaten fischer with decent time available.
Furthermore, exchange odds + pawn aka (wn for br + f7 pawn).