Hi,
i play Reti alot with white.

when someone answers c5 i always was pissed in the passed, so i started to answer b4 with the idea to sacrifice the a-pawn aswell if that gambit is accepted.

valid plan?

It doesn't look totally awful for White, but Stockfish 6 says that Black's best 3rd move is e6 instead of bxa3, with a slight advantage for Black, and though the engine can't always be relied upon, here I agree. The move follows GM Igor Smirnov's well-known principle 'to take is a mistake': by not wasting time on taking the a-pawn, Black develops his own dark-square bishop instead of White's.

In addition, after 3... e6, White has problems with the development of the queenside: the b4 pawn attacks the c3 square where the knight would normally land in the opening, and after axb4 Bxb4, the bishop is placed very well - White either gets his knight pinned on c3 or plays the pawn to c3 and hinders the knight's development.

The alternative way to deal with the b4 pawn - c3 bxc3 Nxc3 - is even worse because White is just a pawn down, with no development lead and no c-pawn left to support the center.

Therefore I like to play 1. Nf3 c5 2. c4 and 3. Nc3, transposing into the Symmetrical English. (This move order avoids the Reverse Sicilian 1. c4 e5.) If the very rare move 2... d5 is played, then White can respond by 3. d4 (transposing into the so-called Austrian Defense) or 3. g3 (the Reversed Grunfeld) with no problems. Explore those variations in the Chesstempo or whatever opening database if you wish.

I guess you get a wing gambit without the e-pawn committed to e4. Probably in a lot of lines you should transpose to that. I think those positions are fine for white but not a great fit for a reti player, what don't you like about 2. g3?

I think the e-pawn should end up on e4 anyway. The setup with the pawns on e3, d4 (and c3) makes White's dark-square bishop too bad for a position where White needs space and initiative to compensate for the sacrificed pawn.

The situation is different in the Nimzo-Larsen attack 1. Nf3 c5 2. b3 d5 3. e3, an offbeat but solid variation where White doesn't sac a pawn and his DSB can enjoy decent scope despite the pawn being on e3 initially because White can either put the pawn on d3 or exchange it for Black's c5 pawn (as the b3 one prevents c5-c4; White can recapture with the e-pawn if cxd4 is played).

Another possible idea is to put pressure on Black's center with Bb5 (preventing e7-e5) followed by c4, delaying the development of Queen's knight in order to give the b2 bishop control over the dark central squares.

E.g. 3... Nf6 4. Bb2 Nc6 5. Bb5 Bd7 6. Bxc6 Bxc6 7. Ne5 Rc8 8. d3.

5... e6 6. c4 Bd6 7. Bxf6 Qxf6 8. Nc3 d4 9. Ne4 Qe7 10. exd4 cxd4 11. Nxd4.

8... Qg6 9. cxd5 exd5 10. O-O Bh3 11. Ne1 leads to a bizarre position , Black has some compensation for the pawn.

Those lines are of course not set in stone, they've been pulled out of my head to show what ideas (c4, Bxc6, Bxf6) White has in general.

It's not that White is much better here than in the Wing gambit, probably by 0.2-0.3 of a pawn, but I'm more comfortable playing an attack with equal material and decent positional ideas, even a hypermodern one, than a gambit.

In addition, White can often reinforce the e5 knight by f4 at some point, transposing to Bird's Opening (using a better move order that avoids From's Gambit 1. f4 e5).

Still, the key idea behind the delay of c4 (when Black does play Nc6 early) is Bb5, the reversed Nimzo-Indian, with the difference being the queenside fianchetto. It's worth noting that Queen's knight is often developed to d2 (after d2-d3), where it doesn't blunt the bishop and from where it can potentially come to f3 or the c4 outpost (if Black plays bxc6 at some point and so can't kick that knight by b5).

If Black refuses to play Nc6 early, then White plays c4 and (if applicable) cxd5, as in T.V. Petrosian vs Hort (1972).

Here are some sources on the Nimzo-Larsen:

A short video by FM Grant Szuveges

Chessgames.com collections, such as willfly's

IM Cyrus Lakdawala's book (most of it is dedicated to 1. b3 e5, but there are chapters of interest: 'Reversed Nimzo-Indian', 'Reti set-ups', 'Reversed London, Torre and Trompowsky')

IM Ilya Odessky's book

I'm not sure how good they are, though

wow thx for that input. i will answer that sooner or later

You're welcome

Upon further reflection, I've realized that 2. e3 is a bit preciser way of attaining the same positions, the difference being in sidelines.

1. Nf3 c5 2. e3 d5/Nf6 3. b3 Nf6/d5 - proceed as above.

2... Nc6 3. Bb5 (a pure reverse Nimzo-Indian, not wasting time on the fianchetto).

2... d6 (or 2... Nf6 3. b3 d6) can now be met with 3. d4 cxd4 4. exd4.

The problem with 2. b3 is that 2... d6 (preparing e5) is rather testing (though likely a bit worse objectively than d5): e.g. 3. d4 cxd4 4. Nxd4 e5, and the knight has to go back to f3 because c2 is still occupied by the pawn (unlike in 2. c4 d6 3. d4 cxd4 4. Nxd4 e5 5. Nc2). 3. c4 (followed by e3, d3, Be2) or 3. Nc3 (followed by e4, Bc4, d3) are equally decent ways out, but they lead to rather closed positions.
_______________________________

Another variation with good database results is 2. c3 (preparing d4) - the Reverse Slav. The extra tempo seems quite helpful to White, I don't know why. If you play the (Semi-)Slav with Black, you can explore its mirror image and give it a go with White

spewmachine seems like a fitting handle for that line. Just play 2. g3

^

I don't quite understand the idea behind 2. g3. Is it a better move order for the Symmetrical English / Reti (early c4, delayed d4), or is it to set up the Reverse Grunfeld (the preparation for d4, so that Black d5 pawn is attacked by the fianchettoed bishop when this pawn break is made*), or is it something else (I don't think you're going for the outdated KIA)?

* 1. Nf3 c5 2. g3 d5 3. Bg2 Nc6 4. d4.

Black can play 2... Nc6 to prepare e5 instead, but then White can play c4 (Symm. English) or d4. In the latter case, White can play the Reverse Exchange Grunfeld if necessary:

1. Nf3 c5 2. g3 Nc6 3. d4 cxd4 4. Nxd4 e5 5. Nxc6.

Is that the point? Or what are you and RoundTower going to do vs 2... Nc6?

Imo 2. c4 > 2. e4 > 2. e3 ~ 2. g3 in terms of theoretical advantage unless proved otherwise. I might try to alternate e3 (Nimzo-Larsen) and g3 (Reverse Grunfeld) in blitz as surprise weapons and see what happens.

Here's my best implementation so far (I'm socially awkward so I'm playing vs computers yet). Many of the above setup ideas (Bxc6, Ne5, f4, d3, Nd2) appeared there.

It's strange that the comp, usually good at tactics, made the 11... Nxe5?? blunder. Even stranger is that 12. exf5 is preferred by Stockfish (to 12. Bxb8 Qxb8 13. exf5 Qxf4), though in the game I just didn't see the hanging exchange

After that moment, I was very imprecise because of my usual time scramble (at least thank goodness for that computer blitz now has an increment on chess.com, 3+2 instead of 5+0), in particular, 23. fxe6 wins way faster than my instinctive 23. f6.

(But how many did I lose? )
actually, I guess the Reverse Bogo-Indian (1. Nf3 c5 2. e3 d5 3. Bb5+) is not bad either as a surprise weapon (as long as White plays like Black in the normal Bogo-Indian), though it's reeeally offbeat I'm yet to find how to exploit the fact that White has an extra tempo and Black's knight is not on f6 yet (as opposed to White's 3. Nf3 in the Bogo).