A fortress fit for a queen
The FIDE chess.com Isle of Man Grand Prix Swiss – yes, that’s the full name! – is going on at the moment. It’s quite a remarkable event: a large Swiss tournament where the winner qualifies directly for the Candidates. Besides a horde of 2650+ grandmasters, there are also wildcards and qualifiers from online tournaments, meaning that, in theory at least, you or I could have a lucky night at home playing online blitz, score a bit more luck on the Isle of Man, stay strong while the seven other candidates withdraw with food poisoning in Yekaterinburg next March, and end up playing Magnus Carlsen for the World Championship next year.
In practice, despite the multiple pathways to the Candidates tournament, the usual suspects seem to reappear, even via the World Cup route. But the Isle of Man event is a notable outlier, in that it’s quite possible someone outside of the top 20 will earn a spot. What makes the tournament even more exciting is that the massive prize fund attracts the guys who’ve already qualified, such as Fabiano Caruana, or who don’t need to, like Magnus Carlsen. It’s a fabulous spectacle, only dampened by the fact that the Australian time difference ruins my sleep.
Beforehand, I predicted that Carlsen wouldn’t win. It’s a pretty safe bet: in such large events, the chances that any one player wins are tiny, even for the world champ. So far he languishes on 2/3, but Caruana, the world number two, shares the lead on 3/3. Still, it’s a better-than-even bet that neither he nor his co-leader will eventually win the title. There’s just so much variability in a large Swiss, which makes it all the more exciting, I guess.
In round 3, Fabiano defeated former Candidate Alexei Shirov. It was an enthralling battle, but I wonder whether Shirov gave up just a little too early. There was a fantastic possibility right at the death:
Here, Alexei played 52.Qd3? and resigned after 52…Qb8!, threatening d1=Q followed by Qxb2#. And 53.Bb3 Rg3! fails, so Shirov called it a day.
But what about 52.Qd5? The idea is to play Qd7 followed by Bb3 if necessary. It seems that this should be refuted by the same move 52…Qb8, but then 53.Qxg2!!:
Black can get a second queen, but White has his own plan thanks to his domination on the light squares. After 53…d1=Q 54.gxh7!!,
…White has the devious threat of Bg6(!!), locking down the black kingside for good. In that case, White wouldn’t even mind an exchange of queens,
…because the sole black queen can’t do enough damage on her own without the help of her partner.
Black might therefore try 54…Kxh7 to rule out this fortress possibility. But now the point of 52.Qd5 (rather than the immediate 52.Qxg2) becomes clear: the temporary absence of defenders around the black king allows for 55.h6!!:
Capturing the pawn either way leads to mate, so Black must play 55…g5, and White captures Black’s last remaining pawn. White will probably lose his h-pawn, but he can set up a very nice arrangement by keeping his queen on the g-file and his bishop on b3, taunting Black to make progress:
If Black could exchange queens, he easily wins by bringing his king to c1 and targeting the pawn on b2. But I can’t see a way for Black to force the queens off, and meanwhile White can be very annoying with his constant pestering of the black king. I suspect it’s a draw.
I haven’t checked this thoroughly with an engine, but from a practical perspective, it would have been interesting to see how things would have played out. How many queens are too many?