Hmmm. Hope he’s not jacked it in. Must be a ton of work.

Keep at it JH.
Looks very interesting. Triangles Are Good.

Bump!

Any news from Jack Hughes (and other rhyming queries if you have them) about the Semi-Slav / Triangle rep for Chessable...?

Jack Hughes wrote on 05/19/20 at 10:15:30:


I didn't think that 24. Qf2 Qc7 gave White any advantage, so I focused on 24. Bf1. However, after 25...Na4 it does appear that Black equalizes, though 26. Rd4 might be a tad more challenging than 26. Nd6.

In the line with 14...b4, 21. b3 bothered me more than 21. a3. However, I took a second look, and I think Black has equality there as well. Overall, I think these two lines together can indeed be a solid alternative to the main line (especially since Black isn't having a fun time after 16. a4 or 16. f4).

On the other hand, if 16. a4 and 17. Ne2 is the only line that actually bothers you, then I would suggest looking into 19...Bb4!?, where I think Black has good chances to equalize. I'm not fond of 19...Nd7 - a dangerous line you didn't mention is 21. Bf1 Nf8 22. e5!?, which led to a win for White in a correspondence game.

Hiruma666 wrote on 05/19/20 at 11:39:15:


My main line here runs 18. Ne4 Rhf8 19. Rfd1 Nf5 20. Rxd8 Rxd8 21. Rd1 Nd4 22. Bxd4 cxd4 23. b4 f5 24. Nd2 e5 25. Bd5+ Kf8 26. b5 Nc7 27. Qxa7 Nxd5 28. cxd5 Qd6 29. Nc4 Qxd5 30. Qxb6 e4 31. a4 Be6,

reaching a sharp and double-edged position. I think Black should hold here, for instance:

32. h4 h6 33. f4 Kg8! =
32. h3 Rd7 33. Rc1 d3 34. Nd2 Kf7 35.Qe3 Qa2 36. f3 Qxa4 37. fxe4 Qxb5 38. exf5 Bd5 39. Re1 Rb7 =

But you know what they say - long analysis, wrong analysis, etc. Hiruma, what were your thoughts on this line?

Syzygy wrote on 05/18/20 at 21:04:14:

Yeah, that's the line I'm talking about. The long line you gave is pretty much the mainline in my file, except that I focused more on 24. Qf2 than 24. Bf1. After the latter, my mainline is 24... Qe5 25. Nf5 Na4 26. Nd6 Re7 27. Rac1 Rd8 28. Nf5 Red7 29. Rxd7 Rxd7 30. c4 Kh7 31. cxb5 cxb5 with full equality.

Against the 12. dxe5 move order black's independent option is 14... b4, in which I am even more confident than 12. h3 exd4. One important line I got was 15. Na4 Bd4 16. Re1 c5 17. Be3 Rc8 18. Bxd4 cxd4 19. Qd2 Re8 20. f3 Nh5 21. a3 b3 =.

My main reason for disliking the mainline 12... Re8 is mostly subjective: I feel like giving up the dark-squared bishop is a big positional concession, and in general, I find it easier to understand why black gets away with it in these 12... exd4 and 14... b4 lines. Especially as an author who wants to be able to justify his choices rather than just list them this is important to me.

Besides that subjective preference, there were a few lines that bothered me at least a little in the mainline. Theoretically speaking, the only line where I struggled to convince Leela that black is okay after 16. a4 Qe7 17. Ne2 Bd6 18. Nd4 g6 19. Nf3 Nd7 20. Rfd1 Rad8 21. Bg5 f6 22. Be3 Bc5 23. Bf4 (I suppose this was the line I had in mind when contrasting Stockfish and Leela evaluations). For what it's worth, if I am to recommend the mainline in the course I might be inclined to go for 19... Nxe4!? which leads after 20. Rfe1 Bb4 21. Bg5 Nxg5 22. Rxe7 Nxf3+ 23. gxf3 Rxe7 to an endgame where certain versions of Leela are quite happy with black, even though Stockfish pretty much hates it.
Syzygy wrote on 05/18/20 at 21:04:14:

Now that is extremely curious. I'll PM you some links.

Hi everyone, sorry for the delayed reply. I've been so inundated with work on the course that it's been difficult for me to keep up. I'll break my post down into several points.

- I'm really not scared of the 'Marshall endgame' after 11... Qxd1 (as far as I'm concerned whenever a move like 16. h4 becomes an option - immobilising the kingside majority and fixing another pawn on the dark-squares - the time to play for a win is over) but I think I like Syzygy suggestion even more, especially considering my goals when designing this repertoire. Nice catch!

- I double-checked my files and it turned I was sort of misremembering them in regards to 16. a4. Certainly, there were a lot of lines where Leela was evaluating the position in white's favour and giving ideas that Stockfish was missing, but at least early on in the lines, Stockfish was claiming a modest white edge. I agree with Syzygy about the positional risk black is running when ...c6-c5 is difficult to get in and the dark-squared bishops liable to quickly disappear.

- I have spent some time checking this 8. Bd3 Bb7 9. 0-0 a6!? and so far I agree that it looks very decent for black. I still can't make up my mind what the most critical line is but nothing seems too scary. I still haven't checked 12... Ne8, but given that it has been the unanimous choice in recent correspondence I would expect that what you say is true.

- As far as I can tell, black's objectively strongest reply to this 11. e4 e5 12. h3 line is probably 12... exd4. Neither Stockfish nor Leela is too thrilled at first but I found that by clicking through the correspondence lines they more or less came around. Leela still had some minor worries (~0.20-~0.30 range), but only minor ones. What your thoughts on this line?

- I've taken a look at 12... Rd8 and I agree that this is a very good line for black. I will change my recommendation in the course accordingly. The reason I had concentrated on 18... Nf5 was that the correspondence games had all featured that move, but if I had gone to the trouble of actually using my brain it would have occurred to me how similar the position is to move 19 of the 12... Nh6 mainline. Once you start analysing the two positions in tandem the reply 18... Rhf8 sticks out like a sore thumb. I probably shouldn't be too hard on myself, because I suspect that this is a sin most of us fall into from time to time, but it's rather embarrassing all the same.

- That's rather annoying about the links to the forums. A few days ago I linked them to beta-testers before sharing the course and they were able to access them just fine. If you'd like, Szygy, I'd be more than happy to send copy-pastes of relevant posts to you through PM (or perhaps email - some of my posts are too long for ChessPub messaging.

Syzygy wrote on 05/13/20 at 19:49:44:

For what it's worth I'm inclined to agree. Almost all of the lines Hiruma mentioned were already in my files with over 100 nodes of analysis and supporting that conclusion. Amusingly enough the only one I'd missed was the 11. Nf3 endgame, but after looking at it I don't think that it's anything to fear. Hiruma gave 12.Rxd1 Rb5 13.Rd4 Be6 14.Re4 but I feel that instead 13... Nxe5 14. Nxe5 Rxe5 15. Rxc4 h5 should be fine for black. Black has the bishop-pair and white's dark-squared bishop is 'bad' from a classical point of view and with the pawn on h5 black is already well placed to deal with white's kingside pawn majority. It reminds me a bit of endgames you'll get in the Marshall Attack.

On this topic, I've actually asked about this line for the next season of Jan's Opening Clinic, so when he gets to it he might offer some interesting thoughts. One thing I note in my question is that there has been a marked trend away from 8. Nfd2 in recent high-level OTB games, which seems like a promising sign for black.

I do, however, share Hiruma's worry about the sharper lines being a lot of work for black, so as part of my beta-testing I'm planning to organise some practice games starting after 12... Bb4 where I advise people to prepare as much as they want to before the game. I'm very interested to see how these games pan out. (And, to make myself sound like a broken record, want as many volunteers for this as possible - join the Lichess team!)
Syzygy wrote on 05/13/20 at 19:49:44:

This part of your post is very interesting to me. I had noticed that Carlsen played 9. Bd3 a6!? against Giri and at the time thought that it looked fairly interesting but unnecessary, and I didn't analyse it very deeply. I was confident that black is completely fine after 9... 0-0 10. 0-0 Bb7 11. a3 a5 (and still am) and thought that if black is fine after having weakened the queenside like that then 11. e4 should be even less threatening. At the time I had done some minimal analysis with Leela and Stockfish, which concurred with this assessment, so I thought nothing of it.

Only in recent weeks have I gone deeper and found just how challenging this 11. e4 is. The reasoning that black avoids weakening the queenside misses the point that ...a7-a5 prepares ...Ba6, when the positionally favourable exchange of light-squared bishops is on the cards. I guess by the time I realised this it had just been too long since I'd seen the Giri-Carlsen game and I didn't really think to go all the way back to 9... a6 as a potential improvement. I will look into it more deeply and get back to you.

Your suggestion against 9. Be2 is even newer to me. I had noticed that Ding Liren played 10... a6 11. Ng5 Be7 and could have sworn that I had analysed his continuation of 12. e4 h6!? to a slight but manageable white advantage, but checking my files I can't actually find any analysis of it! I will take a closer look.

One point I would like to note is that, even if your recommendations are easier equalisers from a theoretical perspective (obviously I'm not deep enough yet to say), Lichess statistics suggest that the traditional mainlines with 9... 0-0 and 10... Bb7 will probably give black players a better practical score at club level - assuming, that is, that black players do their homework. Obviously, I haven't checked your recommendations deeply enough to know en if your recommendations are easier equalisers from a theoretical point of view. There are a lot of really popular ways for white to pretty much just lose straight out of the opening! In such scenarios, my general approach is going to be to switch my main recommendation at the fourth or fifth level of depth of coverage in the course. As an example, this is how I currently intend to approach this thread's topic of 9. Ba5 in the Marshall Gambit -- the super-exciting sacrifices of 10... e5 in the third level, the rock-solid equaliser of 10... f6 in the fourth and fifth.

By the way, does your assessment of 16. a4 in the Bd3 Anti-Meran indicate that you have started using Leela for analysis? As mentioned in the Chessable thread that's a line where I've found Stockfish spitting out zero after zero but Leela much more pessimistic.

Finally, if you want to get a better idea of the approach I'm taking to the course, along with a few of the lines that I think are critical, then please feel more than free to check out the course forums (https://www.chessable.com/discussion/course/the-sharpest-semi-slav/26289/)! Make sure to check out Section 4 ('What You Can Do To Help: Engine Analysis') of the 'Welcome Beta-Testers!' thread before checking out any of the 'Variations for Analysis' threads for an explanation of why the latter are so sparse. I thought that these forums would be inaccessible for an as-yet-unpublished course but it turns out that anyone with the link can see them.

Thank you so much for all your help Szygy. You might not be an official beta-tester but to me you have served as an honorary one, and a truly outstanding one at that!

Great to have you hear Hiruma, I think that we can all agree your first post makes for an excellent debut in our little community! In regards to your lines.

- I am very pleased to see that you have assessed 23. a3 as white's most critical try, because this was my conclusion in my own research as well. Though I am generally reluctant to draw this conclusion, my assessment of these lines was that Stockfish understands them better than Leela, with the latter often claiming a serious white advantage in positions that on subsequent analysis turned out to be equal or even better for black. As an example, the absolute mainline of my files in terms of nodes devoted to it was 24...Rd3 25.Rxd3 Rxd3 26.Qc2! (stronger than the cutest Be5!?) 26...Qxc4 27.Nd2 Qd5 28.Be2 Bb7 29.f3 Rd4! (this was a move Leela didn't like, but came around to after being force-fed the Stockfish lines)30. Bxd4 Qxd4+ 31. Kh1 Nc7 32. Ne4?! (this was Leela's first choice, but this was the point at which Stockfish switched from its trademark 0.00 to claiming a black advantage - Leela tried no less than seven other moves for white here but in all cases was either already happy with black or easily convinced after a few more moves) Nd5 33. Bb5 f5 34. Nc3 Nf4 35. Rd1 Qe3 36. Qd2 Qxd2 37. Rxd2 Nd5 38. Bc4 Nd6 when suddenly Leela starts agreeing that black is better.

- I spent a few hours looking at 12... Rd8, but wasn't able to make it work for black. Unfortunately, after 16. Rfe1 c5 17. Qa4 Kf7 18. Rad1 I concentrated my analysis on 18... Nf5. I haven't had the chance to properly check 18... Rhf8 yet but given how happy Hiruma666 and Syzgy are then I am inclined to suspect that it is a real improvement after which black is doing very well. I will report back after I have taken a proper look. Incidentally, let me emphasise that I am not wedded to any particular reply. When I can be convinced that one line is a clear improvement over my intended recommendation then I will change my intended recommendation. I want the course to be as good as I can possibly make it and to knowingly recommend inferior lines would be in clear violation of that desire.

As an aside, I have finalised my team of beta-testers and am going to start sharing the course with the successful applicants. I am really excited about the team that I have assembled and think that together we will be able to produce a really great course! Unfortunately, I have had some unanticipated technical difficulties in sharing the course with more than five people and am going to wait until I can get that sorted before I share it with individual team members.

Additionally, if there is anyone reading this who is interested in the course and wishes to help out in some capacity but who is not part of the beta-testing team, I would like to announce the creation of 'The Sharpest Semi-Slav' team on Lichess. I am planning to use this team to organise practice games between beta-testers and other interested parties from pre-selected starting positions that may appear in the course. It is my hope that these practice games will enable me to get a better understanding of the human, practical assessment of the lines that I am recommending or considering to recommend. To that end, contributions from anyone - beta-tester or otherwise - will be highly invaluable. Anyone interested can find the link to the team https://lichess.org/team/the-sharpest-semi-slav-on-chessable-official-group.

Finally, Hiruma666 and I have had a great discussion on Chessable about the 12... Bb4 Catalan line that has been discussed so productively in this thread, and I suspect that there is much of value that he can contribute to this discussion. However, this is the Semi-Slav board, and I believe that all this wonderful discussion has been rather misplaced in this thread. I am wondering if it would be possible for a forum moderator to create a new thread in the Catalan subforum (titled something like 'Open Variation with 5... a6 6. 0-0 Nc6 7. e3 Rb8!? - What is the state of play?') and move all relevant posts to that thread.

23...Bb7 24. h3 Nb8 25.Qxa7 Bxf3 26.Bxf3 Nc6 27.Bxc6 Rxc6 28.Qa4 Rcc8  29 Qb5 Rb8 30.b4 Kg8 might just be a draw by repetition if White grabs on c5?