Loshinsky 100 - about a "Loshinsky-magnet" problem

On the occasion of Lev Loshinsky's coming centenary (Jan.17th), it may be profitable and instructive to return to a problem by Lev Loshinsky which was mentioned on this blog. In fact, the problem quoted there was a version of Lev Loshinsky's magnet masterpiece.

In 1947, Loshinsky published his first prizewinner showing the magnet theme:

Лошинский, Лев Ильич Шахматы в СССР 1947 1st Prize #3 12 + 10

1. Qb1! (2. Sh5+ Kxe4 3. Re3#) 1... Rd5 2. Rd4 (3. exd5, Sh5#) 2... Rxg3 3. exd5# 1... Rd6 2. Rd5 (3. Bxe5, Sh5#) 2... Rxg3 3. Bxe5# 1... Rd7 2. Rd6 (3. Rf6, Sh5#) 2... Rxg3 3. Rf6# 1... Rc4 2. Rc3 (3. Sh5#)         2... Sd3,Rxe4,Rxg3 3. Sxd3,Qxe4,fxg3# 1... Rb4 2. Rb3 (3. Sh5#) 1... Ra4 2. Ra3 (3. Sh5#) Other variations: 1... Rxe4 2. Rf3+ gxf3 3. Qxe4# 1... Rxg3 2. fxg3+ Kxe4 3. Rf4#

Let’s focus on the construction. The white Queen plays an ambush key, hiding behind a white Rook which threatens a double check mate on third white move. Since White’s second move can’t be avoided, Black has but one defensive motive to parry the threat: grant a distant flight to his King by moving his Rook away.

The main challenge in the Loshinsky magnet theme is to ensure a good reasoning so that White’s second move must follow Black’s first move. Here the whole idea is carried by the need to have a second threat, besides 3.Sh5#, which will be fulfilled when the white Knight g3 is captured at black second move.

Therefore after 1…Rd5 the weakening is the black Rook is en prise by white Pawn e4, hence 2.Rd4 threatens 3.exd5#. 1…Rd6 interferes black Queen b8, so 2.Rd5 threatens 3.Bxe5# Finally 1…Rd7 interferes black Bishop c8 allowing 2.Rd6 threatening 3.Rf6#. The by-play is also interesting, with the white Rook performing three additional equipollent moves on the third rank following black Rook’s moves on the fourth rank.

This is certainly a masterpiece, displaying an original conception in a convincing form. However, even masterpieces can be improved after a thorough analysis, and that’s what another great Russian composer – Andrey Lobusov – might have certainly thought.

• The first point is the underused white Rook f1, whose sole mission seems to stop black Rook g1 from giving white King a check. The attempt to put black Rook on g2 fails because Black can play 2…Rxf2, providing the flight e3. But is the black Rook the only possible way to attack white Sg3? Why not try a black Pawn h4 and add another black Pawn g5 in order to block this square firmly?
• The second slight improvement is to get rid of black Pawn b6. Yes, the black Queen must be prevented from getting control over the file, but then why not switch the position of black Queen b8 and black Bishop c8 and put the latter on c7 instead of b8?

These two were just minor adjustments and do not significantly improve the original version. The last change is more complex: white Sc1 is slightly underused, so it would be better placed on e1. This would automatically include the shift on white Pawn e2 to g2, in order to avoid a simple mate. We must provide also for 1.Se2#. Therefore the black Rook must be posted on f2, so white Pawn f2 must be traded for two white Pawns on d2 and h2.

And here comes the discovery: because of the black Rook on f2, we must provide a reply against 2…Rxd2, which can be achieved only if the white Queen is placed on c2 instead of b1. It was normal to switch its place from out of play square b5 to the more active d1 (as 1.e3?? is no longer possible) and we finally reach Andrey Lobusov’s version:

Лошинский, Лев Ильич 1 чемпионат СССР 1947 1st Place #3     11 + 11

1. Qc2! (2. Sh5+ Kxe4 3.Re3#) 1... Rd5 2. Rd4 (3.exd5/Sh5#)2... Rxd2 3. Qxd2# 1... Rd6 2. Rd5 (3.Bxe5/Sh5#)2... Sd3,Rxd2 3. Sxd3,Qxd2# 1... Rd7 2. Rd6 (3.Rf6/Sh5#) 2... Sd3,Rxd2 3. Sxd3,Qxd2# Other variations: 1... Rc4 2. Rc3 (3. Sh5#)  1... Rb4 2. Rb3 (3. Sh5#)  1... Ra4 2. Ra3 (3. Sh5#):      2... Sd3 3. Sxd3#      2... Rxe4 3. Qxe4#      2... hxg3 3. hxg3# 1... hxg3 2. hxg3+ Kxe4 3. Re3# 1... Rxd3 2. Qxd3 (3.Qe3/Sh5#) 2... Sxd3 3. Sxd3#

Is this version better than Lev Loshinsky’s original? The answer is not easy and depends a lot on personal taste. Nowadays, a new – although secondary – variation enriching the initial contents would be considered a significant improvement and most composers would probably prefer Lobusov version.

However, one must also consider that the solver’s perspective was a very important factor in ‘40s. There are three unprovided strong defenses (black captures on d3 and g3), which will make solver’s job much easier. Loshinsky preferred to prepare answers to these defenses and the set variation 1…Rxg3 2.fxg3+ Kxe4 3.Rf4# - which is lost in Lobusov version, must particularly not be missed so easily.

Even in the computer era, let’s not underestimate the technical skills of the giant composers from the past!