Weekly Problems 2014-B

The key 1.Bxc7! seems to be defeated by 1…0-0, but here we can prove that castling is illegal by using retro-analysis, where we deduce the play that must have occurred prior to the diagram position. The white bishop on d8 could not have reached that square by normal means because of the black pawns on c7 and e7, which have never moved. The bishop is therefore a promoted pawn that arrived on d8 from d7. When on d7 the pawn attacked e8, meaning the black king must have moved previously to answer or avoid the check. So Black, having moved the king before, cannot castle in response to the key. The threat is 2.Rd8 mate, and if 1…Kf8 then 2.Rd8+ Kg7 3.Be5.

Nigel Nettheim: A neat and simple example of retro-analysis in a directmate.
Dennis Hale: A nice retro idea – technically, the d8-bishop is termed an obtrusive piece.

The key 1.Qh4! entails an unobvious threat, 2.Qe1. White has placed the queen in an ambush position behind the black rook, and if 1…Rg4 (or 1…Rg3, etc.) then 2.Qxe7. 1…Rf5 disables both the threat and 2.Qxe7 by cutting off the f7-rook’s control of f4, but allows 2.Rxe7 due to self-block. 1…Sf5 blocks the same square, enabling the white bishop to interfere with the f7-rook: 2.Bf6. The analogous self-block 1…Sd5 permits the white knight to interfere with the d2-rook: 2.Sd3. 1…d5 defends by creating a flight on d6, and is answered by 2.Bxc7. Also, 1…Bd4 2.Qxd4 and 1…Sh~ 2.Qf4.

Dennis Hale: Very difficult to solve. The white queen leaps into a nest of black men, seemingly disadvantageously restricting her mobility. Particularly pleasing are the variations 1…Sf5 2.Bf6, 1…Sd5 2.Sd3, and 1…d5 2.Bxc7. A delightful problem.

A prominent set variation, 1…Rc3+ 2.Qxc3, is abandoned by the key 1.Qb8!, which creates a battery on the b-file and threatens 2.B5-any. Now 1…Rc3+ provokes a new mate, 2.Bd3. The two white rooks can be captured in four different ways, generating two pairs of dual-free variations in spite of the multiple threats: 1…Kxa5 2.Bxc4, 1…Kxc5 2.Bxa4, 1…Rxa5 2.Rxc4, and 1…Rxc5 2.Rxa4. So the two black rooks are captured in four different ways as well!

Dennis Hale: The key is not easy to see as it removes the set counter to 1…Rc3+ (2.Qxc3) and replaces it with 2.Bd3. The problem is a comprehensive exercise in dual avoidance.
Nigel Nettheim: Extremely neat, with four sacrifices, marvellous symmetry, and a check by Black with a changed mate. (It does not matter at all that there are several threats.) Bravo!

Dennis Hale has discovered that the economy of this problem could be improved slightly. Shift the white king and the h4-pawn to e3 and e4 respectively, and replace the b2-bishop with a white pawn. The play works exactly as before, including 1…Kxc5 2.Bxa4 since the white king now covers d4.

Black has only two possible moves, and both are provided with set play: 1…Sxc2 2.Qb2 S~ 3.Qxh2 and 1…Sxb3 2.Qb2 S~ 3.Qb8. If 1.Kxh2? then 1…Sxc2!, or if 1.b4? then 1…Sb3! The key 1.Qb2! creates another block position. Now after 1…Sxc2 White has only one available waiting move 2.b4, forcing 2…S~ 3.Qxh2. Similarly after 1…Sxb3 White can maintain the block only with 2.Kxh2, leading to 2…S~ 3.Qb8. Technically a mutate, though the changes between the set and actual play are not as dramatic as that seen in many Whyatt problems.

Nigel Nettheim: The tries 1.Qd2? (2.Qxh2, 2.Qe3) 1...Sf5! and 1.Qb4? (2.Qb8, 2.Qc5) 1...Sxe6! show that the battle between the queen and the d4-knight is won by the knight. Instead of impatiently threatening either of those mates on the b8-h2 diagonal, White must more subtly step back with 1.Qb2!, keeping both those mates in mind for the future. So the battle between the queen and the a1-knight is won by the queen. The restraint shown in giving up an immediate attack and instead stepping back to lie in wait is the most appealing feature for me.
Dennis Hale: A delicate waiter with the queen mating at one end of the b8-h2 diagonal in the first variation, and the other end of that diagonal in the second. A very fine problem.

The excellent key 1.Qf2! concedes two flights to the black king and threatens 2.Sd4. When the king accesses these flights, White uses the f4-knight to execute two double-checks, discovered from different directions: 1…Ke5 2.Sg6 and 1…Kf6 2.Sd5. The Q + S battery, initially controlled by the a2-rook and a8-bishop, fires again after the b4-knight interferes with one of these line-pieces, allowing White to shut off the other one: 1…Sc6 2.Se2 and 1…Sc2 2.Sd5. Lastly, 1…Rd5 2.Se2 (somewhat distractingly similar to one of the main variations) and 1…Bg7 2.Sxg7.

Nigel Nettheim: The key 1.Qf2! is not hard to find. The variations are good, especially the flight ones. The a2-rook stops 1.Se2. The h3-pawn is not needed for soundness, but it stops the too-strong defence 1...Rxh2+ which would almost give away the key.
Dennis Hale: The problem is a real eye-opener in line-openings and line-closures.

Two important set variations are 1…c4 2.Qa5 and 1…e4 2.Qxg5. After the key 1.Sxd4!, which threatens 2.Sxc2, the two pawn defences yield new mates: 1…c4 2.Ra5 and 1…e4 2.Rf5. Further, the two queen mates seen in the set play are transferred to another pair of defences: 1…cxd4 2.Qa5 and 1…exd4 2.Qxg5. Such a combination of changed play and mate transference with respect to two thematic defences is named the Rukhlis theme. Three black knight variations make up the by-play: 1…Sxd4/Sb4/Se1 2.Se3, 1…Se6/Sf3 2.e4, and 1…Se4 2.Bf7.

In this miniature, the waiting key 1.Bf3! leads to various attractive mating configurations. 1…Kf5 2.Se7+ generates two model mates with 2…Ke6 3.Bd5 and 2…Kg5 3.Qc1. The distant self-block 1…Bg5 permits 2.Qe7+ Kf5 3.Qe4, while other bishop moves unguard f6: 1…B~ 2.Qe7+ Kf5 3.Qxf6 – another model mate. Lastly, 1…f5 is answered by 2.Kc4, threatening 3.Bd5, and if 2…f4 then 3.Bg4. The latter variation ends with a mirror mate, in which the mated king is surrounded by eight empty squares. Thanks to Michael McDowell who discovered this “forgotten gem” while researching in the library of the British Chess Problem Society.

After the key 1.Qb5!, White threatens 2.Qd3. All moves by the d5-knight defend by creating a potential flight, and a random placement 1…Sd~ is answered by 2.Qf5. The knight has three specific moves that prevent the secondary threat of 2.Qf5, but they commit new errors that enable White to mate in other ways: 1…Se7 2.Qe5 (self-interference), 1…Sf6 2.Sd6 (another self-interference), and 1…Se3 2.f3 (self-block). One minor variation is provided by the other black knight: 1…Sb4/Sc5 2.Sc5.

Nigel Nettheim: 1...Sd~ allows 2.Qf5, except for three “corrections” 1...Se7, 1...Sf6 and 1...Se3. The first two each shut off one of the black rooks, while the third is handled in a contrasting way. The set play is retained after the key. The neat responses to the moves of the d5-knight are the best feature.
Dennis Hale: One small step for a queen, one giant step for the white army. I particularly like the variation 1…Se3 2.f3.

Initially, almost every possible black move has been set with a mating reply, the exception being 1…dxc6. The waiting key 1.Rxb6! covers c6 so that 1…dxc6 is met by 2.Bxc6. Besides completing the block position, the fine key also adds two new variations: 1…cxb6+ 2.c7 (cross-check) and 1…c4/cxd4 2.Rb5 (switchback). Five remaining variations are unchanged from the set play: 1…dxe6 2.Rd8, 1…d6 2.Se3, 1…b3 2.c4, 1…S~ 2.Sf4, and 1…Ba7 2.Sxc7.

Nigel Nettheim: White provides for 1…dxc6 with the surprising 1.Rxb6! The following capture with check (1...cxb6+ 2.c7) makes this an excellent key. The giving up of an en passant capture by 1...b3 2.c4 provides an appealing variation.
Dennis Hale: Arguably the best variation is 1…cxb6+ 2.c7; Black's battery is successfully assaulted by White's battery. I also like the three defences involving Black's d7-pawn. An enjoyable problem.

Two important set variations are 1…Sf6+ 2.Sgxf6 and 1…d3 2.Se3. The key 1.Qc8!, by putting an extra guard on c4, threatens 2.Rxb5. Since the queen has abandoned control of e5, the set play no longer works, but now that c5 is attacked by the same piece, the play changes to 1…Sf6+ Sexf6 and 1…d3 2.Sc3. There’s only one other variation – 1…Sb~ 2.Qd7 – in this straightforward illustration of free change, i.e. a problem with changed mates but which is not a complete block.

Nigel Nettheim: The key was a bit hard to find, perhaps precisely because it is not specially brilliant! It overprotects c4 (thus threatening 2.Rxb5) and c5, but leaves e5 under protected, changing the set 1...Sf6+ 2.Sgxf6 to 2.Sexf6 and 1…d3 2.Se3 to 2.Sc3.
Dennis Hale: Particularly pleasing are the variations, 1…d3 2.Sc3 and 1…Sf6+ 2.Sexf6.

One set variation, 1…d1=Q 2.Bg2, is discarded by the key, 1.Qf4! (waiting). The black king has two flights, and accessing them gives 1…Kf1 2.Qxf2 and 1…Kd1 2.Qxd2/Qxg4. If Black moves a pawn, White exploits the resulting self-block in different ways depending on Black’s choice of promotion piece: 1…f1=Q 2.Qxd2, 1…f1=S 2.Qxg4, 1…d1=Q 2.Qxf2, and 1…d1=S 2.Bd3. Three nice pairs of variations, in which Black plays repeatedly to d1 and f1. The thematic try 1.Qd4? (waiting) produces similar play, and involves one changed mate – 1…d1=S 2.Qd3, but it is defeated by 1…f1=S!

Nigel Nettheim: Giving up discovered check makes the key surprising. 1.Qd4? was tempting, but it is answered by 1...f1=S! The scheme is very neat. The white king is positioned to prevent 1.Qb5+ Kd1 2.Qf1, and at the same time allows 1...d1=S 2.Bd3. Similarly, the g4-pawn prevents 1.Qh5+ Kf1 2.Qd1 and at the same time avoids a dual with 1...f1=S 2.Qf3. The duals after bishop promotions are of little consequence.

If White retracts a nondescript move, such as -1.Kd5-e6, there is no mate-in-two, since the subsequent 1.Ke6 (threat: 2.a8=Q) is answered by 1…0-0! Instead, White must retract in such a way that the resulting position is one in which it can be shown – by retro-analysis – that Black cannot castle.

The solution is to retract -1.b6xBa7. The black bishop on a7 must have been promoted, since the original black-squared bishop couldn’t have escaped from f8, as it was locked in by the e7- and g7-pawns. Only Black’s a-pawn could have made this promotion, and it occurred on e1 or g1 – otherwise the new bishop would have been trapped on the first rank by White’s b2- and d2-pawns. To promote on e1 or g1, the black a-pawn needed to make six captures, while the black b-pawn required four captures to reach f3, for a total of ten captures. White is missing eleven units, but that includes the original c1-bishop which was captured at home (due to the b2- and d2-pawns) and hence it was not taken by one of the black pawns. That means White’s remaining ten missing units were all captured by the two black pawns. How did White’s h-pawn, in particular, reach the capturing path of either black pawn? This could not be possible unless the white pawn had promoted first. Such a promotion required the white pawn to have passed d7 or f7, in either case attacking e8. So the black king must have moved previously to answer or avoid the white pawn’s check.

We have thus proved that the black king has moved before and castling is now illegal. White can accordingly mate in two by playing 1.bxa7! – the same move that was just retracted – and there is no defence against 2.a8=Q.

Dennis Hale: A witty problem. White will not be deterred!

Well done to Dennis, Nigel Nettheim and Andy Sag, all of whom submitted correct solutions to this retro.

Since 1.Ra2? stalemates, White waits for the black bishop to move before checking on the a-file. Only the key 1.Rc2! (waiting) manages to cope with every bishop defence, retaining all of the set play: 1…B~ 2.Ra2+ Ba5 3.Rxa5, 1…Bb6 2.Ra2+ Ba7 3.Rxa7, and 1…Bb4 2.Ra2+ Ba3 3.Rxa3. Not 1.Rb1? Bc3!, 1.Rb3? Bb4!, 1.Rb7? Bb6!, or 1.Re2? Bd2! Another good try is 1.Bc7? leading to 1…Ka7 2.Ra2 Ka6 3.Rxa5 (ideal mate), but 1…Bb6! 2.Ra2+ Ba7. The actual play is simple but only five pieces are used and there are plenty of tries.

Nigel Nettheim: This problem reminds us that a three-mover is not necessarily harder to solve than a two-mover, and might even be a lot easier. Following No.200, the relief was welcome, rather like being given an “early mark” at school.
Dennis Hale: Not an easy problem to solve.

The key 1.Sd2! threatens 2.Qxe4. When the black queen moves, it brings about a remarkable number of dual-free variations. Two groups of ‘random’ moves lose control of e5 or the d6-d4 line: 1…Qe5/Qc6/Qb7/Qa8/Qg6 2.Qe5 and 1…Qxd3/Qe3/Qe2 2.Qd6. The correction 1…Qd5 self-blocks, allowing 2.Qe3. Other correction moves are answered by the queen’s capture, and in three cases this involves opening the R + B battery: 1…Qxe6 2.Bxe6, 1…Qxf3+ 2.Bxf3, 1…Qf5 2.Bxf5, 1…Qxg4+ 2.Rxg4, and 1…Qf4 2.Rxf4. Also, 1…Re7/Sc5 2.Qc4.

Nigel Nettheim: The two candidates (1.Sd2 and 1.Sd6) have an unsubtle threat (2.Qxe4), but the point is the choice between them. That turns out to depend upon Black’s 1...Qxd3, which I found difficult to spot. As far as I can see, the a4-pawn should be removed. The a6-knight is also extraneous, but it provides a fairly good variation 1...Sc5 2.Qc4. Such a piece may be tolerated or not according to personal preference, and here I find it quite OK.

The diagram is a complete block position, with every black move set with a mating reply: 1…exf2 2.e3, 1…Sb~ 2.Ra4, 1…Sc4 2.Rd3, 1…Sc~ 2.Se6 and 1…g3 2.Sf3. White cannot maintain the block with a simple waiting move, e.g. 1.Raa3? Sc4! Instead the key is 1.Rc2! (waiting) which generates two changes: 1…exf2 2.Bxf2 and 1…Sc4 2.Bc3. The other variations are as set: 1…Sb~ 2.Ra4, 1…Sc~ 2.Se6, and 1…g3 2.Sf3. A fine mutate in which the position after the key would also work as a sound two-mover, solved by reversing the key-move, 1.Rc3!

Nigel Nettheim: The two changed mates (for 1...Sc4 and 1...exf2) are excellent.
Ian Shanahan: The problem seems cumbersome, but closer scrutiny shows it to be as economical as possible. I saw immediately that it was a complete block – indeed, a mutate – which proved difficult to solve. Bravo!

The Novotny try 1.Sg5+? cuts off both the g6-rook and h6-bishop, provoking 1…Bxg5 2.Sxg3 and 1…Rxg5 2.Sd2 (or 2.Re3), but 1…Qxg5! is adequate. The key 1.Ba1!, threatening 2.Re5, lures the black queen to a defensive position where it becomes overloaded: 1…Qf6 2.Sg5+ Qxg5 3.Re5, while the Novotny play is retained without the dual: 2…Bxg5 3.Sg3 and 2…Rxg5 2.Sd2 (1…Qh8 2.Sg5+ is similar). If 1…Bf4, the bishop also becomes overloaded: 2.Sd2+ Bxd2 3.Re5. 1…Be3 obtains a flight-square on d3, but results in a dual mate: 2.Re5+ Kd3 3.Rexf3/Rfxe3. (And 1…Bg7 2.Sd2.)

Nigel Nettheim: It is a little unfortunate that the b4-pawn and a3-pawn serve only to rule out 1.Bc3 and 1.Bb2, so they point towards the key. An alternative would be to replace the a3-pawn with a white b2-pawn. Then the key would be 1.Bc3! and the variation 1...bxc3 2.Re5+ Kd4 3.bxc3 would be added. (The d7-pawn could be removed too, adding a short variation 1...Qxd6 2.Sxd6 which, however, the composer might not have fancied).

After the key 1.Qd7! (threat: 2.Qxd4), Black defends by closing the white queen’s line to d4. These defences on d5 and d6 by Black’s rooks and bishops result in two pairs of Grimshaw interferences: 1…Bd5 2.Sc5, 1…Rd5 2.f4, 1…Bd6 2.c5, and 1…Rd6 2.Sf4. Here the four white mating moves are also quite harmonious – two battery mates are delivered by the pawns, which play to the same squares used by the pair of mating knights. However, that these thematic mates also follow other black defences is a little distracting: 1…Be5 2.Sc5 and 1…Bb6/Rf4 2.c5. There's by-play with 1…Sc3 2.Sb2.

Nigel Nettheim: The theme is mutual interference (Grimshaw) between the black rooks and bishops. The variation 1...Rf4 2.c5, with interference by the white pawn, is nice too. If the a2-knight were removed, the variation 1...Sc3 2.Sb2 would be lost, but the position would become a little cleaner, with less distraction from the theme; its removal might be a matter of personal preference.

The give-and-take key 1.Se6! (waiting) offers two new flights to the black king but removes an existing one on f4. Moving to the flight on e4 allows an attractive model mate: 1…Ke4 2.Qxh7, while accepting the sacrificed knight on the other flight yields 1…Kxe6 2.Qf7. The other variations show straightforward unguards, but they are unified by the “close-encounter” queen mates: 1…B~ 2.Qg6, 1…S~ 2.Qg4, and 1…d5 2.Qe5.

Nigel Nettheim: The strong defence 1…Kf4 must be provided for, and the only way is to prevent it. But the try 1.Sh3? allows 1…Ke4! when there is no mate because 2.Qxh7 leaves d4 unguarded. The sacrifice 1.Se6! guards that square so that 1…Ke4 2.Qxh7 is now mate, and that is the standout variation. It is nice that 1…e6 and 1…e5 were provided with 2.Qf6, yet with the key White gives up that resource.

The flight-giving key 1.Qb4! unguards the e4-knight but controls d2, threatening 2.Sc5. If the black king takes the flight – giving a discovered check – White answers with a battery mate: 1…Kxe4+ 2.Sxf3. More battery play follows with 1…Qd1+ 2.Sc2, 1…Qf7+ 2.Se6 (two cross-checks), and 1…Qf5 2.Sxf5; in these variations White exploits the pin of the g3-bishop after the black queen has moved. The half-pin on the third rank is made complete by 1…Bxd6 2.Sf2, when a bishop move leaves the black queen pinned and unable to prevent the mate. Lastly, the self-block 1…Qxe4 permits 2.Qc3.

Nigel Nettheim: The key offers a sacrifice that gives check (1…Kxe4+ 2.Sxf3). The h4-pawn prevents the cook 1.Rxg3.

The thematic try 1.Bg4? prepares for 1…Kb5 2.Bf3 Ka5 3.Be2 followed by 4.Ra6 mate, but Black is stalemated on the third move! The actual play sees White arranging a similar mating net, but with an additional plan to provide Black with a spare move: 1.Bh3! Kb5 2.Bxg2 Ka5 3.Bf1 g2 4.Ra6. So by removing a black pawn, the white bishop releases another one and stalemate is averted. A pointed idea shown in miniature form.

Set play is provided for every black move in the diagram: 1…S~ 2.Qxe4/Sd3, 1…Se3 2.Sd3/fxe3, 1…Sxd6 2.Sd3, 1…Sb2/Se5 2.Qxe4, 1...e3 2.fxe3, 1…g3 2.fxg3/hxg3, and 1…B~ 2.Rf5. Unusually, such a complete block position is solved by a key that entails a threat: 1.Qc6! (2.Qxe4), making this an example of a block-threat. The queen has abandoned the half-pin on the fourth rank, effecting two changed mates: 1…Sxd6 2.Qxd6 and 1…e3 2.Sd3 (also a mate transference), while a set dual is removed: 1…Se3 2.fxe3. Taking the sacrificed queen produces another new mate: 1…dxc6 2.Sxe6. Lastly, 1…Bd5/Bf5 still allows 2.Rf5.

The key 1.Sd5! (waiting) is delightful in granting three flights to the black king. Since the king has access to all four diagonal flights now and each brings about a different mate, the star-flights theme is shown: 1…Kc6 2.Sa7 (model mate), 1…Kxa6 2.Sc7 (sideboard model), 1…Ka4 2.Sc3 (pure mate but not a model because the white pieces on e4 and c8 are not involved), and 1…Kc4 2.Sxd6.

Nigel Nettheim: Two clues make this easy to solve. (a) The b2-pawn can be of use only if the black king moves towards it, so the key piece must be the b6-knight. (b) The strong defence 1...Kxa6 requires 2.Sc7. Thus 1.Sd5! Then follows the “star” theme, in which the black king may move to each of the four corners of its field.

The occupied diagonal represents the Berlin Wall, and when the black pieces move they evoke the Wall’s falling stones. The long-range key is 1.Bh6!, a waiting move. The h8-knight facilitates two nice bishop mates: 1…Sf7 2.gxf7 and 1…Sxg6 2.Bxg6. The other black units are forced to either unguard a square – 1…Sg~ 2.Qe6 and 1…d3 2.Se3 – or open a line for White – 1…e4 2.Qd5, 1…c2 2.Rf3, and 1…b1=Q 2.Qf2.

Composer: The key-piece moves from the beginning of a diagonal to its end, in a straight line parallel to the Wall. The knights on g7 and h8 stand for the horses of the Berlin Quadriga, which was built on top of the Brandenburg Gate in 1793. Later they became a symbol of German national identity.
Nigel Nettheim: No mate is set for 1…d3 or 1…bxc1=Q, and both those black moves require the c1-bishop to move to an appropriate square. The a1-bishop is not needed for the problem, but it is needed to complete the Berlin Wall, and attractive pictorial effects are one of Molham’s specialties, as is seen in his autobiography on this site.

The try 1.Bf8? crosses over the critical square e7 to prepare for a self-interference, 2.Ke7, followed by 2…Kc5 3.Ke6 mate, but Black is stalemated by the try. To set up a similar mate without allowing such an escape, White executes a pericritical manoeuvre in which the bishop goes around the critical square instead of passing over it. 1.Bf4! Kc5 2.Bd2 (attacking b4 to keep the black king trapped on c5 and d5) Kd5 3.Bh6 Kc5. Now 4.Bf8+ brings the bishop to its target square, but unlike the try 1.Bf8?, the move is timed correctly to avoid stalemate, forcing 4…Kd5 5.Ke7 Kc5 6.Ke6. A splendid miniature.

Composer: There is also a thematic (intentional) pericritical try: 1.Bc7? Kc5 2.Ba5 Kd5 3.Bd8 Kc5 4.Be7+ Kd5 5.?? The white bishop obstructs the white king on the critical square!

The square-vacating key 1.Bb3! threatens 2.Qd5. When Black defends with the d7-bishop and its adjacent pawns, they cut off one another’s line of action: 1…Bc6 2.Qa7, 1…c6 2.Qa4, 1…Be6 2.Qh8, and 1…e6 2.Rxg4. Hence the problem demonstrates two Pawn-Grimshaws – pairs of mutual self-interferences by a bishop and a pawn. Two secondary variations are 1…c2 2.Qa1 and 1…Rg5 2.Qe4. The key-bishop must be placed carefully, e.g. not 1.Ba2? c2! or 1.Bf7/Bg8? Be6!

Nigel Nettheim: The d5-bishop must vacate its square while keeping c4 guarded, and the key is the only way to do this without disadvantage. Black’s occupations of c6 and e6 provide the main (symmetrical) variations, the bishop and pawns mutually interfering, thus showing two “Grimshaws”.
Imanol Zurutuza: A very nice problem: Pickabish theme doubled!

All possible black moves have been given set mates, including 1…B~ 2.Qe6/Bd7 – a dual separated by 1…Be6 2.Qxe6 and 1…Bd7 2.Bxd7 – as well as 1…Sh~ 2.Rf4. A fine key 1.Qd3! (waiting) exposes the white king to a check while creating a Q + R battery. The black bishop variations are changed to 1…B~ 2.Bd7 (dual removed) and 1…Be6+ 2.Rc4. Another change is 1…Sh~ 2.Qf3, with the correction move 1…Sf4 forcing 2.Re5. The black rook also shows correction play: 1…R~ 2.Sh6 and 1…Rxg4 2.hxg4. Two more variations, 1…Sd~ 2.Se3 and 1…c5 2.Qd5, round off a good mutate.

Nigel Nettheim: The two rather symmetrical cases of loss of control of a square in a set mate (1…Be6 2.Qxe6 and 1…Sf4 2.Rxf4) being changed to self-blocking of that square in the changed mate (2.Rc4 and 2.Re5) are a very nice feature. The construction seems excellent.
Geoff Foster: A very nice problem, with two new battery mates, including a surprising cross-check.