<2...g6> In the game V Zbandutto vs Botvinnik, 1924 played the same year, we can see <2...d5 3.c4 e6 4.Bg5 Nbd7 5.e3 c6 6.a3 Qb6?! 7.Qc2 dxc4?! 8.Bxc4 Bd6 ±>.

<4.e3> It is far more popular, and possibly better, to develop the other knight here: <4.Nc3> as seen in 2 games John Cochrane - Mahescandra, Kolkata (India) 1855, 1-0. These games played by Cochrane in India gave the name "Indian Defenses" to everything starting with 1.d4 Nf6... Because Mahescandra and others were always playing 1...Nf6 in reply to 1.d4, which appeared new at the time. India, so exotic with the spices, elephants, tigers, maharajas etc...

<6.Bd3> The most played move here, and possibly the way to secure a small opening advantage for White, is <6.Be2> as first played in 1925 (so one year after the actual game!) in either G Fontein vs Euwe, 1925 or in <Henri Gerard Weenink - Max Euwe, Amsterdam (Netherlands> 1925, 0-1>.

<6...Nc6> Maybe not the ideal spot to develop this knight here. It is more popular to put it on d7.

<8.dxe5?!> This trade is not indicated here. As mentioned by User: aw1988 above <8.d5! Nb4 9.Bb1 ⩲/=> Michael Greul (1885) - Harald Zippel (1744), Mittelfranken Team Championship 04-05 (Germany) 2005, round 7, 1-0.

<9.e4?!> This is not the best. One could suggest the simple <9.h3=> as in the game Gerit Lehmann (1427) - Wolfgang Rausch (2197), Frankfurt Championship (Germany) 2008, round 7, 0-1.

<9...Be6?!> Possibly better is <9...Bg4 ⩱>. Stockfish 8 - 64 bits POPCNT.

<10.Bg5> Simply <10.h3 a5=> Stockfish 8 - 64 bits POPCNT.

<10...Qd6?!> For a second time, better is <10...Bg4!=/ ⩱> Stockfish 8 - 64 bits POPCNT.

<11.Be2> Again better is <11.h3! Nh5=/ ⩲> Stockfish 8 - 64 bits POPCNT.

<11...Qc5?!> There is <11...Nd4! 12.Rc1 c6=> Stockfish 8 - 64 bits POPCNT.

<12.Nd5?> Premature a move! <12.Be3! Qb4 13.a3! Qe7 14.Nd5! Qd6 ⩲/=>. Note that 13...Qxb2?? 14.Na4-+ would lose the queen. Stockfish 8 - 64 bits POPCNT.

<12...Nxe4▢ ⩱> Botvinnik starts this second part with the advantage as Black.

<13.Nxc7 ∓/ ⩱> Andreev misses <13.Be3! Qd6 14.Qc2 Nf6 15.Rad1! Bxd5 16.cxd5 Nb4! 17.Qc4! b6 18.Bg5! e4 19.a3 exf3 20.Bxf3▢ Nbxd5! 21.Bxf6! Qxf6 22.Rxd5 c5 ⩱ / ∓> also 19.Nd2 Nbxd5! 20.Nxe4 Nxe4 21.Qxe4 Rae8 22.Qc4 Qe5 ⩱ Stockfish 8 - 64 bits POPCNT.

Black to play

click for larger view

<13...Rac8? 14.Nxe6 fxe6=/ ⩲> Botvinnik, in his turn, missed a beautiful tactics: <13...Nxf2! 14.Rxf2 e4▢ 15.Qd2> (or 15.Nxa8 exf3▢ 16.Bxf3 Qxg5! 17.Nc7 Bd4 18.Nxe6 Bxf2+ 19.Kxf2 fxe6 ∓) <exf3 16.Bxf3 Bd4! 17.Bh6 Rad8 18.Nxe6 fxe6 19.Bxf8 Rxf8 20.Qe2 Bxf2+ 21.Qxf2 Qxc4 ∓> Other move orders are possible. Also, note the possible intermediary move 16.Be3 Bd4! 17.Rxf3 Bg4! 18.Rg3 Bxe2 19.Qxe2 Rad8 ∓ Stockfish 8 - 64 bits POPCNT.

<15.Be3 Nd4!=> Stronger is <15.Bd3! Nxg5 16.Nxg5 Qe7! 17.Ne4=/ ⩲> Stockfish 8 - 64 bits POPCNT.

<20...Nc5?!> Gives back the advantage to White. Still, more than one moves were maintaining the equality : <20...Ra8 21.Be3> (or 21.Rfe1 Nc3!? (or 21...Nd6=, as well as 21...Nf6=) 22.bxc3▢ Rxa7=) <Nd6=>, or <21...Nd6=>, <21...Bh6=>. Stockfish 8 - 64 bits POPCNT.

<21.Ng5?!> Better is the trade line <21.Bxc5! Rxc5 22.Ng5! Rc6 23.Rfc1 Rfc8 24.Rxc6 Rxc6 ±> An improvement of the given line could be <22...Rfc8! 23.Nxe6 Rc2 24.b4 ⩲/ ±> Stockfish 8 - 64 bits POPCNT.

<21...Re8? 22.Rfc1 Rxc1▢ 23.Rxc1 ±> Time to be active! <21...Nd3! 22.Be3 Bh6! 23.Nxe6 Bxe3! 24.fxe3 Re8! 25.Ng5 Re2=> also possible is 24.Nxf8 Bxf2+ 25.Rxf2 Rxf2 26.Ne6! Kf7 27.Nd8+ Ke7 28.Nxb7 Rxb2 29.Rxb2 Nxb2=; or <21...Nd3! 22.Nxe6 Ra8 23.Be3 Bf6=> Stockfish 8 - 64 bits POPCNT.

<24.Rc2?!> Why not the more active <24.Rc7! h6 25.Ne4▢ Nxb2 26.Rxb7 Rc8! 27.g3 Rc2 ±> with a clear advantage for White, due to the passed pawn on the queen-side, the lack of mobility of the black bishop and the black pawn majority being difficult to exploit. Stockfish 8 - 64 bits POPCNT.

<25.Nxe4> There was also the interesting 25.Rc7 to consider.

<26.Rd2?! Bxb2=> Andreev should have transposed in the last variation given at the very end of my PART 2 comment, with <26.Be3 ⩲>. Now this game was back to equality.

<29...Kh6?> This is a sad place for the black king to be! White still having some tiny edge, the burden was on their shoulders so to speak. So, simply <29...Kg8=>. Stockfish 8 - 64 bits POPCNT.

<30.g4?!> Even better was <30.h4! Ra4 31.f3 g5! (escape route!) 32.hxg5 Kg6 33.Rxb7 Bxa3 34.Kh2! ±> Stockfish 8 - 64 bits POPCNT.

<30...Bxa3?? 31.g5+! Kh5 32.Rxh7+ Kg4+-> The losing move by Black. Automated play? Time pressure? Botvinnik failed to see the danger he is facing, with his king stuck on the edge of the board. The only move, as mentioned above by User: eristoff, is <30...g5! 31.Rd6! Rxa3 32.Rxe6+ Kg7 33.Nxg5 ±> Stockfish 8 - 64 bits POPCNT.

White to play, and win!

click for larger view

<33.h3+?? Kf4=> Incredible! Andreev lets the win slips away! Here, as spotted by User: eristoff, White is winning like this: <33.Rf7+!> and now the lesser evil, according to Stockfish 8 - 64 bits POPCNT, is to give away the bishop to divert the white knight from the defense of the g-pawn, while pinning the f-pawn with <33...Bc5 34.Nxc5 Kxg5 35.Nxe6+ Kh5 36.f3! g5 37.Nf8!+-> and there is nothing Black can do to survive, here.

<34.Rh4+ Kf5=> Possibly better was to move the knight away to c3, d2 or g3.

<36.Rg4 Kf6 ⩱ / => Better is <36.Rh7!=> Stockfish 8 - 64 bits POPCNT.

<38.h4 Ra5! ⩱> Maybe more commendable is <38.Rg5 Ra4! 39.Re5 Kf7 ⩱> Stockfish 8 - 64 bits POPCNT.

<39.Rg5?? Rxg5 40.hxg5-+> is weak, as User: Bob Loblaw expressed. More than weak, it is losing! White is a pawn down and had the weakest minor piece on the board... Black only has a small advantage after <39.Ng5! Kf6 ⩱ / => Stockfish 8 - 64 bits POPCNT.

<40...Bb4> Faster is <40...b5! 41.Kf1 Kf7 42,Ke2 e5!-+> etc. Stockfish 8 - 64 bits POPCNT.

An interesting game full of errors, but entertaining.