World Cup Baku 10 September to 4 October 2015

[Jonathan Rogers]

Well, pretty much anyone who reaches the QF could cause more surprises. Remember Kasimdzhanov in Tripoli 2004, suddenly beating Ivanchuk, Grischuk, Topalov and Adams, each next round supposedly the one where he would bow out ... and Topalov had looked in absolute top form in that event.

[Jonathan Rogers]

Adams & Caruana out, two matches going to play-off (Giri-Woj and Karjakin-Andreikin).

Topa and Wei Yi have some chance to level their matches, MVL will try but probably not convert a piece for two pawns v So and Eljanov-Jakovenko will prob end in a draw and a play-off.

Agreed. I've been very impressed by Topalov so far today.

[Jonathan Rogers]

sigh ... I really must stop commenting. Topalov will probably allow ...Qe7 and ...Ng4+ next!

[Christopher Kreuzer]

Commentator's curse!

[JustinHorton]

How does Vachier-Lagrave try to make progress? Knight to e7 and try and bring the king round?

[Tim Harding]

Topalov saves a few rating points by offering a draw in a lost position. Svidler agonised for a couple of minutes and then settled for qualifying for the next round.
I originally wrote "think MVL probably cannot win but he can certainly try" but after looking at the commentary for a while (White played Kd2) maybe it can be won. It will certainly be an instructive example for future endgame books in that case.

[LawrenceCooper]

sigh ... I really must stop commenting. Topalov will probably allow ...Qe7 and ...Ng4+ next!

He found another way to be eliminated

[AustinElliott]

I originally wrote "think MVL probably cannot win but he can certainly try" but after looking at the commentary for a while (White played Kd2) maybe it can be won. It will certainly be an instructive example for future endgame books in that case.

After 55. Ke3, the chess24 engine showed ...Kc4, Kc5 and Kb5 all as highly advantageous for Black (and presumably largely equivalent), but NOT ...Kc3, which to club patzers such as myself might look the most obvious move. Anyone better at endings than me (so - most if not all people reading) care to enlighten me as to why 55 ..Kc3 doesn't work? I presume it lets the white pawns loose, or allows White access to e4 with the King after a judicious e5, but curious to know.

PS I suppose this might have been covered in the commentary, but our WiFi is too slow for the video feed.

[Christopher Kreuzer]

I suspect the 'level' evaluations are actually lost (and being mis-evaluated by the engines), unless it is some complex form of opposition that holds a draw. MVL will probably win this (after 60.f5). The variations would probably repay close study.

[MartinCarpenter]

The lines from the Chess24 engine did seem thoroughly confused for several moves. Not sure why this specific position seemed so hard for it.

Still it just hit -88, which seems to indicate it has finally fallen into range of things soluble by its powers of calculation Then a few moves later its rather scarily declaring mate in 39.

[Roger de Coverly]

I presume it lets the white pawns loose, or allows White access to e4 with the King after a judicious e5, but curious to know.

chessbomb gives this line as drawing

[FEN "8/8/2n3p1/6Pp/2k1PP1P/4K3/8/8 b - - 12 55"] 55... Kc3 56. e5 Ne7 57. e6 Nf5+ 58. Ke4 Nxh4 59. e7 Nf5 60. e8=N h4 61. Nf6 Kd2 62. Kf3 Kd3 63. Kg4 Ke3 64. Nd7 h3 65. Kxh3 Kxf4 66. Nf8 Kxg5 67. Nxg6 Kxg6 68. Kh2 Ne3 69. Kg1 Kf7 70. Kh1 Nf5 71. Kg2

[AustinElliott]

The position and lines after White's 55th were understandable in some respects - f5 for White was always losing to Black being able to plonk the Knight on e5, so that White's f-pawn got stuck on f6 (i.e. f5 Ne5 f6) when the e-pawn was also then blockaded by the Knight - Black would then just use King opposition to force White back and take the e-pawn.

What I didn't really get was what the winning plan would be for Black absent White going for f5 - thus if Black didn't play ..Kc3 (see prev), and White just shuttled the King between e3 and d3. I'm guessing this was where MVL's ...Nd8 - f7 - d6 manoeuvre is significant, re-routing the Knight to where it attacks the e-pawn. I guess the trick is (i) spotting this Knight manoeuvre; and (ii) being able to see that the Knight can do this without allowing an f5 advance at some stage.

Thanks Roger for that drawing line after ..Kc3. I am reminded once again of why I hate Knight endings almost as much as Rook ones...!

EDIT: MVL is being interviewed and said he rejected 55. ..Kc3 because of e5

[Christopher Kreuzer]

Cripes. Wei Yi has messed up his winning position. He can't play endings, it seems. Magnus will be taking note of that.

[Tim Harding]

Cripes. Wei Yi has messed up his winning position. He can't play endings, it seems. Magnus will be taking note of that.

In the position after 60 Rd3 the online engine at chess24 says equal 0.00
The Lomonosov tablebase says White mates in 36.

After 62 Kg5 (immediate Kf4 was also good) the online engine wakes up and sees the win.

[Christopher Kreuzer]

Oh, my apologies to Wei Yi then!! (did he play it perfectly or not?)