correct reply to 1.b3 gambit by black

What is the best way to reply to this annoying gambit? do you take it and defend like crazy or do you sidestep it with 4.nf3? and if 4...f6 how should white proceed?

Tell me if I'm correct here. After Ne7, if white is going to capture the pawn on e5, he must capture it first and swap the bishop for the c6 knight, otherwise the e5 pawn will be a poisoned pawn when black chases away the b5 bishop. Did I calculate that correctly?

if you take the pawn, black has very nasty compensation, with a6 bxc6 nxc6 , for example bb2 qg5! or bg3 h5! h4 d4 is very unconfortable for white to proceed.

Just take the pawn... the engine says both choices are about equal, but taking the pawn is a long term advantage. White should be able to play a precise followup considering he's the nimzo larsen player.

nope, i played a fellow master who threw this at me in an otb game. and the resulting position for white are MISERABLE. IF you play perfectly, white gets... a 0.1 edge. if you as much as deviate even slightly, you at least -0.5 down and the position doesnt get any easier to play.

nice lines!

This is a bit like one of White's more troublesome lines against Owen's Defence, only there White has also thrown in Bd3 with the extra move. My own instinct here in the Nimzo-Larsen would likely be to decline the gambit (accepting does look pretty miserable) and maybe play it something like this:

after which Qd2 (or maybe Qd3) looks fine. White has the classic pressure on the a1-h8 diagonal making it a gnat's awkward for Black to complete development, and can hope to use the half-open d-file too.

i play it with black, but i always forget to play the right moves lol

Correction to #7: Qd1 seems like a better retreat after more thought. To meet ..., Qg5 with Nf3.

I think you're being misled by the statistics. The very fact black knows this obscure nimzo-larsen line deep will correlate with him winning - for reasons that go beyond the position itself. There are places black can go wrong as well, it's just not happening frequently in practice because most black players are following their prep very precisely, as we'd expect from anyone playing this line. The solution is for the nimzo-larsen player to be prepped just as well. But I think rarely will this be a problem in practice

Initial glance at the engine shows this as a possible line. It’s equal, but you’re playing the Nimzo-Larsen against fellow masters players in over the board tournaments so something tells me you don’t mind equal positions as White. Might be worth giving a glance to as the accepted form of the gambit is very highly favored by all engines right now. I first heard of it a year and a half ago back, around the same time that the Endgame Offer Caro-Kann started gaining traction as another Leela recommendation.

Same. I need to do a deep dive into the lines as Black again as I didn’t maintain that part of my repertoire and now I’m playing positional chess against the Nimzo-Larsen while I get the lines back in my head

the problem isnt e5, its f6 after d4 e4 nfd2 f5 or even nfd2 h5!? be2 bg4

Checking to see if I got your variations correct. 4… f6 5. d4 lines, correct? It seems similar but different, as long as 6. Ne2 rather than 6. Nfd2?!

@darkunorthodox88 I have been thinking about this off and on since my initial snap reply. I now think that my first thoughts were complete unutterable dross. On closer examination, Black is simply better in those lines. Still worse, the more I look at the positions after 4..., Ne7 the more I feel that Black's game has more "playability" than White's even where the engines are saying it's equal or trivially better for White to the tune of +0.04. This has surprised me mightily because Black has some interesting "tries" against lines with Nc3 and Ne2 against Owen's Defence. I guess the extra move White has made gives Black something extra to try and undermine. All of which makes me wonder whether maybe 3. Nf3 (I seem to recall this a van Geet idea, and do recall it was played by Gerard Welling against David Bronstein sometime back in the 1990s) might be the answer. Or maybe 3. c4 as long as one doesn't end up on the wrong side of a Larsen-Spassky game.

i know what you mean, the similarity between this and the owens line with nge2 nc3 bd3. but they ultimately very different. 
and im also inclined to agree that on a practical level , 4.nf3 seems preferable over memorizing all these lines for a 0.1 edge lol esp since if white deviates even a little, he is already worse.

check out this line for example

the problem is

I’m still recommending switching to 5. Nf3 against the 4… Ne7 systems, playing 5. d4 against the 4… f6 systems only.

Why not just play 1...d5 in responce to 1 b3?