LeeRoth wrote on 04/17/20 at 22:18:21:

I replied to the middle part about 3. Nc3 earlier.

I think "Slow Slav" is not just anything with an early e2-e3, but this line in particular: 1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.e3 Bg5 5.Nc3 e6 6.Nh4. There is also 6.Bd3, but that line has been considered equal from ancient times. Recent games don't change that assessment. You are correct that a lot of strong GMs are using 3.e3. For some reason I didn't notice that before. Now I spent some more time in the ChessBase tree, looking at how those strong GMs play it. I put the "main line" (if 3.e3 has such a thing) in bold. One thing to keep in mind after 3.e3 Bf5 is the related opening 1.d4 d5 2.Bf4 Nf6 3.e3 c5 4.c3 Nc6 5.Nd2 e6. If you think black is doing great against the London system then you will feel right at home in the 3.e3 Slav.

1.d4 d5 2.c4 c6 3.e3 Bf5
This is the important reply, avoiding transpositions.
(a) The big move is of course 3...Nf6, played 7x more often, when can happen 4.Nc3 e6 (4...a6) (4...g6), or 4.Nf3 Bf5 (4...Bg4) (4...e6) (4...a6) (4...g6).
(b) Another option for black is 3...e6. Black can follow up with the Stonewall, but usually just plays ...Ng8-f6, in which case I don't understand the move order.
4.Nc3
(a) 4.Nf3 e6 5.Nc3 see 4.Nc3.
(b) 4.Qb3 Qc7 5.cxd5 cxd5 see 4.cxd5.
(c) What the GMs are mostly playing is 4.cxd5 cxd5 5.Qb3 Qc7. This probably will transpose to the similar line with Ng1-f3 / ...Ng8-f6 in. If white wants to play this line, 3.e3 seems like a good move order. I have to say, I don't get the point of 6.Bb5+ at all, it must be one of those silly-looking computer moves based on a random tactic that humans, especially me, can easily overlook.
4...e6
Only move, although in the database 4...Nf6?! has been played 5x more often. That alone is a good reason to play 3.e3 against the small fry.
5.Nf3
The line 5.g4 Bg6 6.h4 is a specialty of Z. Kozul, and 6.Nge2 is another try. If black is worried about this then 3...Nf6 is the way to go.
5...Nd7
Again avoiding the transposition, although in the database 5...Nf6 has been played 20x more often, allowing the Slow Slav with 6.Nh4.
6.Bd3
(a) 6.Be2 h6 7.Bd3 has been tried, directed against ...Bf5-g6. 7...Bxd3 8.Qxd3 f5!? (ambitious, in practice black always plays 8...Ngf6, with a useful extra ...h6 in an equalish line) 9.cxd5 (9.g4!? works out okay for black 9...dxc4 10.Qxc4 Qf6 11.gxf5 Qxf5 12.Qe2 Ngf6 13.e4 Qh5 = Stockfish) 9...cxd5 10.g4!? g6 11.gxf5 gxf5 12.Bd2 Ngf6 = Stockfish. g2-g4 is not necessary for white, but is an attempt to punish black's extra ...h6 move. Otherwise, it will look very much like 6.Bd3. I suppose a disadvantage of ...f5 after ...h6 is black can no longer play ...Ng8-h6. An advantage is ...g7-g5 is supported.
(b) 6.Qb3 as I gave earlier is not especially good, e.g. 6...Qb6 7.c5 Qc7 8.Nh4 Bg4! as Jack Hughes correctly pointed out. White has done poorly from here, the most recent example being 9.h3 Bh5 10.g4 Be7 11.Ng2 Bg6 12.Nf4 b6 13.cxb6 Qxb6 14.Qd1 e5 15.dxe5 Nxe5 (1/2, 30) Akesson - Rosberg, Vasteras op 2017.
6...Bxd3
6...Bg6 is also a good move, when white can break with 7.e4, or with 7.Bxg6 hxg6 8.e4, or develop with 7.O-O Ngf6 8.b3 Bb4 9.Bb2, or with 7.O-O Ngf6 8.Qe2 Bb4 9.Bd2 a5.
7.Qxd3 f5
Both players should seriously look at 7...f5. It's not especially popular, but scores rather better than 7...Ngf6, which just transposes to a standard ...Bf5 Slav line, e.g. 8.O-O Bb4 9.Bd2 a5. Here black should also equalize sooner rather than later, so it seems black has a good choice on move seven.

7...f5 is rare enough, I don't see a main line. A couple of interesting examples:

(a) 8.O-O Bd6 9.b3 Ngf6 10.cxd5 cxd5 11.Nb5 Bb8 12.Ba3 Ne4 13.g3 a6 14.Nc3 g5 15.Nd2 h5 16.f3 Nxc3 17.Qxc3 Nf6 18.e4 dxe4 19.fxe4 Ba7 (1/2,78 on time?!) Cuenca Jimenez - Carlsen, Banter Blitz 2020

(b) 8.Bd2 Qf6! 9.Ne2 Bd6 10.Qb3 Rb8 11.Bb4 Bc7 12.Qa3 dxc4 13.Rc1 Ne7 14.Bd6 Bxd6 15.Qxd6 Nc8 16.Qc7 Ncb6 (0-1, 52) Miton - Hector, DEN-chT-1112 2011

(c) I think white should play 8.cxd5 to force 8...cxd5. Otherwise, Hector's 8...Qf6 will allow black to answer with the positionally desirable ...e6xd5. After 8...Qf6, black's further scheme is ...Bd6, ...Nh6, ...O-O, and then marching on the kingside. But 8.cxd5 cxd5 also looks quite solid for black, and black is doing fine there in the database.

I don't agree with the idea that the bishop does little on f4 in the Exchange Slav. Above all from those squares it takes away the c7 and b8-squares from the black queen and rooks (the usefulness of the former being seen in the very line we are talking about). Black is often even willing to play ...Bd6 just to get rid of this bishop, offering white the long-term positional trump of the 'better' bishop, and sometimes white even rejects the offer!
Of course all methods of study time allocation have their drawbacks, and the drawback in mine is definitely that out of fashion ideas could take me by surprise. I for one certainly do not pretend to know what the correct way to balance study of old lines and cutting edge theory.
Which players people should pay attention to of course varies from person to person based upon a large variety of factors. As someone living in Australia where there are no 2550 players active on the tournament circuit your line of reasoning certainly does not apply to me, but for you it sounds like it makes sense. In terms of personal preparation I'm willing to admit that I pay far more attention to top level chess than would be ideal, but on the other hand doing so gives me a much greater enjoyment of top level chess and opening study than a more improvement oriented approach - the latter would probably even require that I stop spending so much damn time on the openings, and that is a price I am definitely not willing to pay!

Well,  3.e3 can’t be too bad, as a lot of strong GMs use it to reach the Slow Slav.  Not sure why.  I always just thought that 3.e3/4.Nf3 and 3.Nf3/4.e3 transposed into one another.

In his Strategic Opening Repertoire, Watson recommended 3.Nc3 and 4.e3 as a way to reach the Meran.  Why not go 3.e3 first as a way of avoiding the Slav Gambit (3.Nc3 dxc4) and the Winawer Counter Gambit (3.Nc3 e5)?  Not sure.

Anyway, after 3.e3, it looks like any of 3..Nf6, 3..Bf5, 3..e6 or 3..g6 should transpose to regular lines after 4.Nf3.  White has to be careful about trying to deviate with 4.Bd3 due to 4..e5! but he can try 4.cxd5 cxd5 5.Qb3 Qc7 6.Bb5 if he wants something different.

I'm afraid I am going to have to plead ignorance on this line - in fact when I read your comment my first thought was that you might be confusing it with the line 1. d4 d5 2. c4 c6 3. Nc3 Nf6 4. e3 Bf5 5. cx5 cxd5 6. Qb3, which has an obvious and concrete justification. I guess part of the explanation of my ignorance would be that I usually only study lines when I see them played either against me or in high level modern games (I don't know my classics that well), and this line seems like one of those that is historically significant but hardly ever played nowadays (in my database of recent games 5. cxd5 was played 62 times compared to 1147 with 5. Nc3, and the highest rated such player was rated 2555. To be honest it does look strategically ridiculous to me, and I stand by my assessment of the examples I gave as showing the misplacement of a queen's bishop behind the pawn chain in the Exchange Slav. But maybe I would be a bit more flexible in my thinking and the line would make more sense to me if I actually studied it. One thing I could say in my defence is that the mainline is apparently 6... Qc7 7. Bd2 with the idea of a later Bb4, which is ruled out by the early Nc3. But considering that 7. Nc3 is apparently not completely stupid it wouldn't be much of a defence.
I still maintain that the position after 7... cxd5 is nothing white should aim for (mainly because I know when I'm outmatched, and Stockfish outmatches me by quite a lot!), but in my attempt to justify that claim I learned something new. Thank you for that.

an ordinary chessplayer wrote on 04/17/20 at 03:05:58:

Also Summerscale (and Johnsen) in A Killer Chess Opening Repertoire.

You really haven't seen this line?

1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.e3 Bf5 5.cxd5 cxd5 6.Qb3

Played for the win by many strong grandmasters, B. Larsen, M. Tal, V. Salov, amongst others; notably played a lot by Slav expert P. Nikolic; and a particular specialty of Mark Hebden.

Played in a world championship match!
https://www.chessgames.com/perl/chessgame?gid=1013152

Advocated for white in a recent repertoire book, and mentioned here on ChessPub. Maybe one of the other posters can recall which/where.

Mentioned in quite a lot of books.

• Martin (1990) Trends in the Slav, game 88
• Markov/Schipkov (1994) The Slav, pg.62
• Silman/Donaldson (1996) The Slav, pg.62
• Flear (2005) Starting Out Slav and Semi-Slav, pg.70-71
• Vigorito (2008) Play the Semi-Slav, pg.237-242
• Sakaev (2012) Complete Slav I, pg.189-190

There is also the trappy sideline 6.Nc3 e6 7.Ne5.

Played in a world championship match!
https://www.chessgames.com/perl/chessgame?gid=1013149

The big idea is 7...Nc6? 8.g4! +/-. I have won three games with 8.g4, including one with black! After 1.d4 d5 2.c4 c6 3.cxd5 cxd5 4.Nf3 Nf6 5.Nc3 Nc6, twice I faced 6.Bg5?! Ne4 7.Bf4. The first time I sacrificed a pawn somewhere, and had to work hard for the win. The second time I was prepared. 7...e6! (Vigus gives 7...g6 as best. I agree 7...g6 is good, but 7...e6 is also good.) 8.e3? g5! -/+.

I was quite surprised to find that modern theory has completely defanged this trap. 1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.e3 Bf5 5.cxd5 cxd5 6.Nc3 Nc6! (?! according to the old theory) 7.Ne5 Bd7! this move equalizes. See De Firmian (2008) MCO-15, pg.486, quoting Milov - M. Gurevich, Corsica 2006.
https://www.chessgames.com/perl/chessgame?gid=1437647

Thanks for advancing the discussion!

1.d4 d5 2.c4 c6 3.Nf3 dxc4.
It's a little off topic, except in the sense that white is explicitly avoiding it with 3.e3, but I don't dispute your points. I always play 3.Nf3 Nf6 4.Nc3 myself, but 3.e3 might have practical value for Slow Slav players. It's the kind of early transposition that interests me.

1.d4 d5 2.c4 c6 3.e3 Bf5 4.Nf3 e6 5.Nc3 Nd7 6.Qb3 Rb8 7.cxd5.
Maybe 7...cxd5 is an improvement. I'm a little concerned that black is committed to ...Nb8-d7 instead of ...Nb8-c6. It's true white's c1-bishop is behind the e3-pawn, but that's a legitimate line in the Exchange Variation. One thing in black's favor is, by leaving out ...Ng8-f6, defending the queenside has been smoother, e.g. 8.Bd2 a6!? (either now or on move 9) 9.Rc1, and without b5 available white's only queenside plan is 10.Na4 and 11.Bb4. But I don't know if white will actually get anywhere with it. Or white can wait a bit with 10.Be2 and 11.O-O before trying Na4. And again, Nh4 might happen.

1.d4 d5 2.c4 c6 3.e3 Bf5 4.Nf3 e6 5.Nc3 Nd7 6.Qb3 Qb6.
You might be right that an early Qb3 doesn't mix well with Nh4. I don't know that much about the Slow Slav since I don't play it as white, and as black I switched to the QGA before it became trendy. But your 7.Nh4 is a lemon. 7.c5 Qc7 8.Nh4 must be the way to play, and whether black can avoid a "normal" Slow Slav position is still an open question. I'm just poking around in the ChessBase opening tree here, and the early statistics are all hugely in favor of black (white scores like 30%!). But when I continue with normal-looking moves for white, the statistics become slightly in favor of white (usual 53% or better).