Reverse Queen's Gambit

After 1. d4 d5 2. e3 c5 3. dxc5, this turns into a reversed Queen's Pawn.

Critique my play and let me know what you think... Thanks!

I feel that stuff like this, or say 1. e4 e5 2. d3 - just gives black equality straightaway. That's all that is wrong with them.

Surprisingly, 1. d4 d5 2. e3 c5 is being played at the highest levels - just search at http://www.chesslab.com/PositionSearch.html ! Earlier even Capablanca played it as white (but lost) - ive given that game also.

Here is a recent highly rated game which starts like "SomeGuy vs. Bowens" until

4. b4, where white wins:

[Event "XV Open"]
[Site "Albacete ESP"]
[Date "2007.??.??"]
[White "Cabello Rodriguez,S"]
[Black "Martinez Rodriguez,P1"]
[Result "1-0"]
[WhiteElo "2105"]
[BlackElo "2304"]
[ECO "D00"]
[Round "4"]

1. d4 d5 2. e3 c5 3. dxc5 e6 4. b4 a5 5. c3 b6 6. Bb5+ Bd7 7. Bxd7+ Qxd7

8. cxb6 axb4 9. cxb4 Bxb4+ 10. Bd2 Bc5 11. Qc2 Qc6 12. Nc3 Na6 13. Qa4 Ne7

14. Nf3 Nb4 15. b7 Rb8 16. Ke2 Rxb7 17. Rhc1 Na6 18. Qxc6+ Nxc6 19. Na4 

1-0

The game is so short it looks like black blundered somewhere - 

i am too tired for anything to register ...

 Here is Capablanca's one:

[Event "?"]
[Site "Philadelphia"]
[Date "1910.??.??"]
[White "Capablanca,Jose "]
[Black "Milnes,B "]
[Result "0-1"]
[ECO "D00"]
[Round "?"]

1. d4 d5 2. e3 c5 3. dxc5 e6 4. b4 a5 5. c3 Nf6 6. Nf3 Be7 7. a3 Ne4 

8. Bd3 f5 9. Nd4 O-O 10. f3 e5 11. fxe4 exd4 12. exd5 Bh4+ 13. g3 dxe3 

14. Qf3 Nd7 15. gxh4 Ne5 16. Qxe3 Qxd5 17. O-O Nxd3 18. Rd1 Rd8 

19. Qd4 Qxd4+ 20. cxd4 Nxc1 21. Rxc1 Rxd4 22. b5 Bd7 23. a4 Rg4+ 

24. Kf1 Rxh4 25. Nd2 Rxh2 26. Nf3 Rh1+ 27. Kf2 Rxc1 28. Rxc1 Kf8 

29. Kg3 Ke7 30. Kf4 Be6 31. Re1 Kf6   0-1

[contd] the chesslab database has only 8 games with 3. dxc5, the result being the one white win i posted; 2 draws and 5 white losses [including capablanca - i posted that game also - and frank marshall !] so barring theoretical breakthrough 3. dxc5 cannot be at present considered good for white..  i wish a strong plyer could comment more on this - give some insight ... meanwhile lets look at the stats..

again restricting oneself to chesslab; as far as  1. d4 d5 2. e3 c5 is concerned there are:

from 1991-present: 

95 games ; white won 35%, black won 37%, draw 28% - which is not good - but!

 compare with pre-1991:

57 games, white won 26%, black won 52%, draw 22%

Looks like theoretical improvements for white are being found.

I see that post-1990 3. Nf3 and 3. c3 are the most common moves. [i sampled only games won by white for this]

I try to play it often. I think the best response after 4.b4 here would be a5, breaking white's queenside apart...wouldn't it?

- 3. dxc5 is not so good.
- 4. b4 is probably just weak.

The correct way to play this seems to be 3. c4 and this will most likely end in the 4-knights variation of "improved" Tarrasch after 3. ... e6 4. Nf3 Nf6 5. Nc3 Nc6.

The other idea is 3. Nf3 followed either by 4. c4 with transition of moves, 4. c3 with Colle-attack or some b3 stuff (Colle-Zuckertort).

I was thinking Bxa3 was not so good, but it looks like black got such a good game that anything wins, but certainly white did not play the best moves.(11.Qb1 and 13.e4 and 14.Bxc5 in particular)