Found what looks like an untested line in the Caro-Kann Tartakower (6.g3 Bd6 7.Qe2+) feedback welcom

6.g3 Bd6 7.Qe2+

The known theory here is mostly 6.Bc4, 6.c3 or 6.Nf3. The Qe2+ trick itself isn't new ( it's a known motif after 6.Bc4 Bd6 7.Qe2+ Be7 ) but nobody seems to have tried it in this g3 move order, where White hasn't committed the bishop yet.

Stockfish confirms 7...Be6 as Black's best reply, and it's stable from depth 10 to depth 50 — not a shallow-search artifact. Main line:

7...Be6 8.Bh3 (pinning) Qe7 9.Bxe6 Qxe6 10.Qxe6+ — and Black is left with a backward e6-pawn in a risk-free, pleasant endgame for White. 7...Qe7 instead is straight up bad for Black — big advantage for White right away.

Engine eval is only around +0.2, so it's not a refutation of anything — but it's a clean, low-risk structural edge with zero theory on the other side, and I've been scoring around 60% with it in practice (small sample, so take that with a grain of salt).

Curious if anyone's seen this before or has ideas on how Black should try to deviate earlier.

"Black is left with a backward e6-pawn in a risk-free, pleasant endgame for White."

• It is pleasant for black too: equal as can be.

Wrong forum ?

Interesting idea. The key drawback for Black seems to be the endgame after 8.Bh3 Qe7 9.Bxe6 Qxe6 10.Qxe6+, where Black is left with the backward e6-pawn and White gets a very comfortable, risk-free position.

The fact that 7...Be6 holds up from depth 10 all the way to 50 makes the idea even more convincing. Definitely a nice practical weapon.

"White gets a very comfortable, risk-free position"

• Black gets a comfortable risk-free position too.
  The backward pawn at e6 is no liability. Black can play Kf7 and Re8 to defend the pawn.
  Black has a good bishop Bd6 not on the color of his pawns b7-c6-e6. White has a bad bishop Bc1 on the same color as his pawn d4 and his pawns f2-g3-h2.

"a nice practical weapon"

• White trades his good bishop Bf1 for black's bad bishop Bc8, and white dissolves the black doubled pawns f7-f6.
  Even more: before the trade white has a queen side pawn majority and d4 is a candidate passed pawn in the endgame. After black recaptures ...fxe6 pawn d4 is no longer a candidate passed pawn.
  It is playable, but dead equal.

@tpr said ^

"White gets a very comfortable, risk-free position"

• Black gets a comfortable risk-free position too.
  The backward pawn at e6 is no liability. Black can play Kf7 and Re8 to defend the pawn.
  Black has a good bishop Bd6 not on the color of his pawns b7-c6-e6. White has a bad bishop Bc1 on the same color as his pawn d4 and his pawns f2-g3-h2.

"a nice practical weapon"

• White trades his good bishop...

I think we're discussing two different things here. I'm not claiming the position is objectively winning for White, only that it's a comfortable practical position.

Yes, ...Kf7 and ...Re8 defend e6, but "defended" doesn't automatically mean "not a weakness." White can still pressure it, while Black has to spend tempi coordinating around it.

As for the bishop trade, I don't think calling Bf1 "good" and Bc8 "bad" settles the position. After 8.Bh3, White is specifically using the bishop to force ...Qe7 and then exchanging it to remove Black's active dark-squared bishop. Whether that's objectively equal is debatable, but I don't think it's accurate to dismiss the resulting position as automatically "dead equal."

The +0.2 eval obviously isn't a refutation. My point is that 7.Qe2+ gives White a low-risk position with a useful practical imbalance and, more importantly, avoids the usual theory. That's why I called it a practical weapon, not an objective advantage.

"I'm not claiming the position is objectively winning for White"

• I am not claiming the position is objectively winning for white after the better 6 c3, 6 Bc4, 6 Be3 either, but white retains an advantage +0.2: black must fight for equality. With 6 g3 Bd6 7 Qe2+ you give black his equality 0.0.

"White can still pressure it"

• If white still had his light square bishop, then white could pressure pawn e6.

"exchanging it to remove Black's active dark-squared bishop"

• No black's active dark square bishop Bd6 stays, while white has spent his good bishop Bf1 to remove black's bad bishop Be6.

"The +0.2 eval obviously isn't a refutation"

• Normal lines 6 c3, 3 Bc4, 6 Be3 give white +0.2; 6 g3 Bd6 7 Qe2+ gives 0.0, dead equal.

"a useful practical imbalance"

• No there is no imbalance left: the black doubled pawns are resolved and pawn d4 is no longer a candidate passed pawn.

This is the position they are discussing and noop e6 is not a weakness at all.

I think this would serve value only as a surprise weapon that too only against higher rated players, as it is an easy draw for both sides. Personally I rarely play for draw, (I think I might have never played for a draw in my life.) I would say that black could lose if they try too hard for a win but honestly its not easy to make a mistake from this position in classical game.

TL:DR
It is a solid line but not at all ambitious one.

+0.2 is funny.

you could be shown 20 positions that Stockfish evaluates as +0.2 and I doubt half of those most could not say where the +0.2 is coming from, and, fire up a different engine and you would get a wider range of different values. Fact.

The point?

The point is that +0.2 is objectively equal. I don't think either side has to "fight for equality". The equality is in the position.

Way too much reliance on computers here and the "absolute truths" it purports to expound.

Better players still use the classic informant descriptions of "Equal/slightly better for white or Black/Unclear...etc." in conversation about chess. Why? Because chess is not a computer calculation. Chess is played by humans, in the end. We, humans, talk about it from a (false) position of expertise, as if we know. We toss in the comp evals to "make our case". Just silly.

Let us know when you sit at the post-mortem and say something silly like "Well, I had about a +0.3 advantage and you are fighting for equality." --- That, if you don't want to embarrass yourself, will never happen.

Doing it online does not count as there is no immediate feedback on how stupid it may sound from behind a keyboard.

Slapping a numeric value on it does not make it more right. If anything, it dehumanizes the entire event that happened betweeen two human players.

Just Silly.

Two AI-slop posters in one thread. Record?