The opening position after the book resembles a Petroff Defence, with a slight advantage for White. In game 33 a sharp tactical sequence led to a double-rook-and-pawns endgame by move 20, with White in control of the d-file. Around moves 35-36 SF allowed Komodo to win a pawn, which the latter failed to capitalize on (KRx+1vsKrx endgames being notoriously hard to win). The middlegame lasted ten moves longer in game 34, and led to a B+N vs BB endgame. Once again White had an extra pawn but White's knight was offside which allowed Black enough time to establish a firm blockade of the potential passer.
Game 35 went the whole length, with Komodo reporting 0.00 and SF a minuscule Black advantage throughout. Scheduled for playoff, although it's hard to imagine that anything but a draw could result from the final near-symmetric position.
In game 36 Komodo didn't see a minuscule Black advantage, so a quick GM draw was the result.
Another line with an early exchange on e5. Not particularly fun to play but interesting to see how the engines handle the semi-open positions with a static center of this kind. The position didn't stay semi-open for long in game 37 thanks to sharp play by both sides. As often is the case in engines games, a sharp middlegame led to a level endgame, with Black having a distant passed pawn on the queenside but no real winning chances.
In game 38 SF took its time to build up his position before opening it some more, so Komodo had to try opening it himself before things got much worse. That in itself didn't help Black much as the following middlegame was evaluated by both engines as advantageous for White; however, the best SF could think of is exchange-exchange-exchange into a B+N+some pawns endgame with an extra pawn for White. Komodo cleverly eliminated the queenside pawns after which the draw was not far off, all the pawns being on the same side of the board.
A line where White goes for an early b4 push rather than the more common kingside fianchetto. Not a particularly threatening line, IMO; without the g2 bishop the b-pawn advance doesn't have much sting.
Game 39 saw another semi-open, semi-symmetric stolid position with an e4/e5 center, similar to quite a few of the previous games in this match. Around moves 24-25 SF went for a repetition of moves, which Komodo eventually avoided (see the effect of the twofold-repetition fix from around the New Year on SF's eval here), nevertheless it couldn't avoid several exchanges and failed to make any meaningful progress.
Game 40 saw some enterprising play on Komodo's part, to avoid the above-mentioned near-symmetry it didn't hesitate to allow for some pawn weaknesses in its camp. It ended up with a dangerous-looking (to a human, that is) passed a-pawn which didn't not hep to avoid yet another draw.
Round 21
E16 - Queen's Indian Defence with ...Bb4+
A more mainstream opening than what we have seen so far in this match - a Queen's Indian/Bogo-Indian hybrid, although an offbeat line.
Game 41 featured some solid play with numerous exchanges; SF was not afraid of pushing pawns in front of his king to get Komodo's offside knight stranded. Consequently, much of the middlegame revolved around White's effort to liberate it and Black's counters to that. By move 40 a KNPPPP endgame arose, with K having a passed pawn on the kingside. SF did not hesitate to give up a pawn and yet get a draw.
Game 42 was a bit livelier, with Komodo going for pawn weaknesses in its camp balanced by active piece play, which must have led SF to overrate its chances for a while. On move 45 K sacrificed a knight for an attack on the white king; two moves later SF had to part with a rook in order to avoid a perpetual check. In the resulting complex endgame with queens on the board the material balance favored Black: a rook and two pawns for two minors. Although SF succeeded in drawing, the impression one gets is that it was barely hanging on.
Although the line starts out as an English it transposes early on into a stolid Gruenfeld with e3. After the creative exertions of the previous round Komodo opted for a quick draw in game 43. With white pieces in game 44 SF went for a more aggressive b4, only to give Black a somewhat more pleasant position. The game ended in a sharp position that deserved being played out on the board. Having seen the 35...Bc5 36.gxf5 Rxf5 37.e4 dxe4 38.Nxe4 Rxa4 39.Nf6+ Rxf6 40.exf6 Qxg2+ 41.Kxg2 line the competitors chose not to, though.
Game 45 was another early draw. Komodo went in for complications on move 14 and won a pawn; the price was a broken-up king's cover with the resulting kingside counterplay for Black.
Game 46 was an even shorter, featureless draw. I think if I ever run another such match I'll have to pick the opening lines manually.
Round 24
A46 - Polish Defence cum Colle-Zukertort System
Not a fan of the Colle but it's tons better than the Exchange French. Leads to the kind of positional play Komodo thrives on. In game 47 it proceeded fine until SF uncorked a startling 20...Qg5! Seven moves later an opposite-colored bishops ending was reached; although the queens were also present, the game never strayed far from a draw. Game 48 saw Komodo close the queenside early rather than give SF hanging pawns or an IQP, which allowed White to try for a kingside attack. Black parried the attack, forced the creation of hanging pawns after all and proceeded to win one of the them; however, in the resulting endgame his extra pawn proved indefensible thanks to the unfortunate placing of the black pieces - after 35.Be5 White was threatening to win at least an exchange with 36.Rc3. Komodo had to settle for spoiling its kingside pawn structure and relinquishing the extra pawn to reach a somewhat inferior (although still drawn) rook ending.