To answer <DaltriDiluvi>'s question, the Schliemann Deferred is rarely seen precisely because 5.d4! exd4 6.e5, as in this game, is considered very strong. Stockfish 16, depth 56 assesses that position as +0.67. Not a good time for Black, though he might still hold with perfect play. ChessBase's correspondence database shows that White scores 68.5% in 269 games from that position.

The regular Schliemann (the immediate 3...f5!?) is considered sounder precisely because 4.d4!?, the thematic response to Black's flank adventure, loses material to 4...fxe4 5.Nxe5 Nxe5 6.dxe5 c6! White's main lines against the regular Schliemann are 4.Nc3 and 4.d3. Stockfish 16, depth 77 assesses 4.d3 fxe4 5.dxe4 Nf6 6.O-O as +0.31. SF 16, depth 75 gives the alternative 4.Nc3 fxe4 5.Nxe4 Nf6! 6.Qe2 d5 as +0.33. Both comfortably within the drawing margin. The Schliemann Deferred is a worse opening, but obscure, and Pragg gambled that Vidit wasn't prepared for it. That seems to have been the case.

So if the Schliemann Deferred is so good for White, why did Vidit lose? Consistent with <sudoplatov>'s comment, Stockfish 16 considers 8.Qxd4 better (+0.89) than Vidit's 8.Nxd4 (+0.71). It says that his very next move, 9.Nxf5, was a lemon allowing Black to equalize with 9...Nxb3 10.axb3 d6!▢ The point is 11.ex6 Qd7!=.

<'The regular Schliemann (the immediate 3...f5!?) is considered sounder precisely because 4.d4!?, the thematic response to Black's flank adventure, loses material to 4...fxe4 5.Nxe5 Nxe5 6.dxe5 c6! '>

click for larger view

The piece sac 7.Nc3 scores reasonably well OTB. Even I have had a few good wins with it. G Chandler vs B Tauren, 1989 It has good OTB shock value.

It still amazes me that top GMs, very well prepared for a high-level tournament, with seconds and engines, can go astray so early in the opening. If 8.Qxd4 is book, why did Vidit play 8.Nxd4_? And then 9.Nxf5_? Did Vidit not know the opening at all? Or was it part of this prep, on which he got out-preped by Pragg?

<If 8.Qxd4 is book, why did Vidit play 8.Nxd4>

Possibly to step out of Pragg's prep. Play an OK move to side step outside the mainline where a T.N. may be lurking. He took 18 minutes deciding on it.

Yes FSR post is good, however as FSR says Jon Hammer's assessment "The game could go either way depending on whether White's advanced pawn turns out to be strong or weak. " is of more value than any computer evaluation. If it was too unbalanced for even him to call it then a S.F. evaluation is next to worthless.

GM surya Shekhar ganguly , who is second of Vidit and in padt has been second to Anand earlier , knew this opening , however , apparently , he did not show it to Vidit.

<On Twitter, grandmaster Srinath Narayanan tweeted: “Surya had shown me this variation a few years ago. I wonder if he had shown this variation to Vidit.”>

Stepping out of a prep to avoid a TN makes sense, but generally only if you have a prep yourself which apparently Vidit didn't: he thought for 18 minutes (thanks for the info) and made suboptimal moves afterwards. In that case it's probably wiser to go by the book which has a big advantage: saving time which you can use when the opponent diverges. Also there's a chance he's bluffing, so calling his bluff right away is better especially on a dubious defence like the Schliemann.

Totally agree on your comments of humans & psychology vs computer.

The version of SF that was turned inside out by A0, which I think was SF8, is completely different from current versions. Even six years ago the comment that SF doesn't understand the concept of counterplay was a gross overstatement at minimum, but it's quite false today.