Round 1

The Short System with 4.Nf3 and 5.Be2 is quiet but venomous: White wants c4 or b4 and slow space on the queenside. Sebastian answers 6...a5, which takes b4 away from White for a long time and provokes 7.a4, turning the a4 pawn into a target later on. Almost every club player would push 6...c5 here without thinking. The grandmaster first looks at what the opponent wants, then takes it away.

Same picture with 7...f6: the classical lever against e5, played at once rather than prepared with 7...c5. After 8.c3 fxe5 9.dxe5 Bc5 the bishop is active, the knight is heading for e7 and f5, and White has to hold e5 with pieces. Two moves that show what preparation looks like at this level: not the obvious plan, but the one the opponent likes least.

White has centralised the knights on d4 and g5, and both look at e6. The pawn is undefended, and Nxe6 would fork queen and rook. It looks dangerous, and this is exactly where the game is decided. There is only one move that keeps the balance: 13...Bxd4. Anything else, for instance 13...Nc5 or 13...Qb8, leaves White a good pawn better.

Sebastian gives up the handsome b6 bishop for the knight and thereby removes one of the two attackers. After 14.cxd4 comes the second only move: 14...h6. The knight has to retreat, and with 15.Nf3 White has spent two tempi on an attack that gained nothing. The position is level after that, with a slight edge for the more active black pieces.

The lesson: when an enemy piece arrives with a threat, first count how many attackers point at the target square, then exchange the attacker you most want gone. Here that was the centralised knight on d4, not the one that jumped to g5.

With 15...b6 and 16...c5 Black has opened the centre, and after 18...Bg4 19.h3 the question that shapes so many Caro-Kann games arises: keep the bishop pair or remove the f3 knight? I would have narrowly preferred 19...Bh5, keeping the bishop on the h5 to d1 diagonal with all options open. Sebastian chooses 19...Bxf3, a good move, not the best one. The difference is a quarter of a pawn, so a matter of taste.

The reason for the exchange is strategic: the f3 knight is the key defender of d4 and e5. Without it, after 20.Bxf3 Nf5 21.Bg4 cxd4 22.Bxd4 Nc5 White has to cover d4 with the bishop, and that bishop immediately becomes exchange material. After 23.Bxf5 Rxf5 the position has simplified: White has two islands with a4 and e5, Black the isolated d5 pawn. Level, but clear. For a team leader on board one, a clear path to simplification is often worth more than a tenth of a pawn.

The phase from move 24 is the only one where I have anything to say about Sebastian's moves, and it is not much: 24...Qf8 instead of the immediate 24...Rf4, and 26...Rc7 instead of 26...Rf7. Neither gives away anything real. White uses it for a liquidation with 27.Bxc5 that at first sight wins a pawn: 27...bxc5 28.Rxc5 Rxc5 29.Rxc5 Qxc5 30.Qxf4.

White really is a pawn up now. But Sebastian had seen the whole path: 30...d4 creates a passed pawn, and after 31.Qe4 Qc1+ 32.Kh2 Qxb2 material is level again while only the white queen holds back the d-pawn. Every black move in this sequence is the best one. 33.Qa8+ starts the perpetual check, and both sides take the draw.

It is a lesson in counting material: what matters is not the interim balance after 30.Qxf4 but the position at the end of the forced sequence. Whoever has seen it in advance can allow the liquidation without a moment of nerves.