Double Rounder

TODAY was a double-round day in the Czech Tour's Highlands Open at the Hotel Atom, Trebic.
My two games could hardly have been more different - game one being effectively decided in the opening, game two deep in the ending.

Spanton (1923) - Jan Kalveus (1854), Round 7
English Opening v Dutch Defence
1.c4 f5 2.Nc3 Nf6 3.g3 c6!?
Not a popular choice, but it has been played by a 2418.
4.Bg2 d6!?
Aiming for a Stonewall with 4...d5 looks more natural.
5.d4
By transposition we have reached a position that occurs 127 times in ChessBase's 2018 Mega database.

Black's fifth move is surprising but has a certain logic

5...Be6?!
Threatening the c4 pawn, but this awkward-looking move was not played in any of the database games.
6.d5 Bf7 7.Qb3 Qd7 8.Nh3 g6 9.Be3 Bg7 10.Ng5 0-0 11.Rd1 h6?
It cannot be nice as Black having a White knight hanging around on g5, but the text spends a tempo to make White play a move he wanted to play anyway.
12.Nxf7 Rxf7 13.c5
White already has a winning advantage, according to my main analysis engines Stockfish9 and Komodo9.

Neither player has connected rooks, but Black is busted

13...cxd5
The engines suggest 13...Ng4 or 13...Na6, but have White winning.
14.Nxd5 Nc6 15.Nxf6+ Bxf6 16.Bd5 (1-0, 44 moves)

My second game today reached a knight-and-pawn ending after 39 moves.

White has just captured a rook on d1 in Miroslav Zacharias (1727) - Spanton (1923), Round 8

It is often said a knight-and-pawn ending is the closest you can get to a pawn-ending. This is true in the sense that it is much closer than, say, a rook-and-pawn ending is to a pawn-ending.
But really there are some major differences, not least the fact that one player may sac his knight to get rid of the opponent's last pawn.
With the help of Stockfish9 and Komodo9, and later the Nalimov six-figure endgame tablebase, I will try to use John Nunn's endgame convention to annotate these moves.
39...Ke5?
Black is much better after 39...Nf5 or 39...Ke6, but the natural-looking text turns a probable Black win into a drawn position.
40.Kd2?
Again natural-looking, but the engines give 40.Nc6+! Kd5 41.Nb4+! Kd4 (or 41...Kc5 42.Nd3+ Kd4 43.Kd2) 42.Kd2 and Black cannot make progress.
40...Kd4?
I had seen the superior 40...Kd5, which threatens to trap White's knight, but then played the king too far forward.
41.Nc6+ Kc5 42.Ne7 Kb4?!
Better seems to be 42...Nb7, but probably not winning.
43.Kc2 Kc5
Dead drawn is 43...Nf5? 44.Nxf5 gxf5 45.Kb2
44.Kc3 b4+ 45.Kd3 Nb5 46.Ke4
46.Ng8 is possibly the more sure-fire draw, eg 46...Nd4 47.Nf6 Nxb3 48.Nxh7 Kb5 49.Nf8 Nc5+ 50.Kc2 b3+ 51.Kc3 g5 52.Nh7 Ka4 53.Kb2 Nd3+ 54.Kb1
46...Nd4 47.Nd5 Nxb3 48.Nxb4?!
This may draw, but Komodo9 strongly prefers 48.Ke5. Stockfish9 reckons both moves draw.
48...Kxb4 49.Kxf4 Kc5 50.Kg5 Kd6 51.Kh6?
Almost certainly drawing was 51.Kf6, eg 51...Nd4 52.h4 Kd7 (Komodo9's best shot, but not good enough) 53.h5 gxh5 54.Kg5 Ke6 55.Kxh5 (tablebase draw)
51...Ke6 52.g4 Nd4 53.f4
MZ offered a draw in my time.
53...Kf6 54.Kxh7

Black to play and win

****
****
****
****
****
54...Ne6?
I was torn between this and the correct 54...Ne2 (54...Nf3 also wins)
55.h4?
On 55.f5, I planned 55...Ng5+ 56.Kf6 Nf7+ 57.Kh7 g5, and wins. But we both missed that White draws in this line thanks to 56.Kg8!
If I had played 54...Ne2, then 55.f5 loses trivially to 55...g5
55...Nxf4 56.Kh6 Ne6 57.g5+ Nxg5! 58.h5
Or 58.gxh5 Kf5 etc.
58...Nf7+ 0-1