### 23 Dec. 2022 | by Peter Wong

The Valladao task is accomplished in a problem that features all three special chess moves, namely castling, en passant capture, and promotion. The mixture of three such distinct elements makes the theme an unusual one, but it’s a pleasing idea that’s adaptable across many genres. In Part 1 we examined a number of directmates that illustrate the task. Here we shall look at a helpmate, an endgame study, and a retro-analytical problem. You can find further examples on this site: (1) a proof game by Satoshi Hashimoto, Weekly Problem No.623, (2) a series-selfmate by Ian Shanahan, No.6 on his biographical page, and (3) a retro by John Keeble, in the blog Special conventions in retro-analytical problems.

Arturo Carra & Isidoro Zezza
2nd FIDE Tourney 1959, 1st Prize

Our first position is one of the earliest helpmates to bring about a Valladao. The black king will be mated on the eighth rank by the knight, which can cover one diagonal flight. Black must block four other flights, and this is possible in the castled position with d8-rook, d7-pawn, c7-knight, plus one more unit on b8. The quickest way to arrange the latter is by utilising a promoted piece, the path of which (on b1) needs to be cleared. 1.0-0-0 Ka2 2.b1=R Se4 3.Sc7+ c4 4.bxc3 e.p.+ Ka3 5.Rb8 Sd6. Black carries out all three thematic moves in this single-line problem.

The Problemist 2007, 1st-2nd Special Prize =

White to play and win

The Valladao task is rarely seen in endgame studies, and this example achieves it with splendid economy. 1.Kd2+? fails to handle the two advanced pawns – 1…g1=Q 2.Rxg1+ Kxg1 3.Kc2 a1=Q – so White begins with 1.0-0-0+! A long side-variation shows how Black gains another passed pawn on the h-file to no avail: 1…Kh2 2.Kb2 dxe4 3.Kxa2 Kh3 4.Kb3 Kxh4 5.Kc4 Kg4 6.Kd4 h4 7.Kxe4 h3 8.Rg1 Kh4 9.Kf3 h2 10.Kxg2 hxg1=Q+ 11.Kxg1. The main line proceeds with 1…g1=Q 2.Rxg1+ Kxg1 3.Kb2 dxe4. Now 4.f3? allows Black to draw with 4…e3! 5.f4 Kf2 6.f5 Kxe2 7.f6 Kd2 8.f7 e2 9.f8=Q e1=Q. Instead, 4.f4! exf3 e.p. (4…Kf2 5.f5 Kxe2 6.f6 is too slow for Black) 5.exf3 Kf2 6.f4 Kf3 7.f5 Kf4 8.f6 Kf5 9.f7 Kg6 10.f8=R! A nice underpromotion to finish the problem. White avoids (1) 10.f8=Q? a1=Q+! 11.Kxa1 stalemate, (2) 10.f8=B? which leaves Black in a kind of fortress, able to defend the h7-pawn permanently, and (3) 10.f8=S+? Kxh6! draws as well. White’s thematic e-pawn also performs an Excelsior.

Klaus Wenda
Die Schwalbe 1978

Helpmate in 2
Twin (b) Be3 to g1

This helpmate/retro invokes two problem conventions governing the special moves: (1) castling is permitted unless it can be proved illegal by retro-analysis, and (2) starting with an en passant capture is prohibited unless it can be proved that the previous move was a double-step by the opposing pawn. In the diagram position, Black’s whole force is present, so White’s last move wasn’t a capture, such as gxh7. Most of White’s units couldn’t have made the last move, including the king which would’ve been in an impossible check had it just come from d6 or d5. Only the d4-pawn could have played last, moving from either d3 or d2, and that was preceded by a bishop check like …Bf4-e3+. Since the white pawn didn’t necessarily make a double-step, it’s forbidden for Black to capture it en passant. On the other hand, there’s no indication that castling is illegal, hence the solution is 1.0-0-0 a7 2.Sd5 a8=Q.

In position (b), the e3-bishop starts on g1 instead. White’s last move still couldn’t have been any capture, like exd4, for the same reason as before. The difference is that now d3-d4 can be ruled out as the last move, because that would leave Black with no possible way of delivering the bishop check from g1. Only if the last move was d2-d4 could Black have given the check legally, with the discovered …Re3-c3+. Since we have shown that the previous move must have been the pawn double-step, the solution 1.cxd3 e.p.+ Kxb4 2.Bg8 hxg8=Q becomes valid. What about the sequence commencing with castling that solved part (a) – is it still viable? Since the d4-pawn was on d2 just a move ago, the pair of pawns on b2/d2 would have trapped the original c1-bishop, which never moved and was captured at home. That means the dark-squared white bishop on a5 is a promoted piece. As there are no missing black units, the only white pawns capable of marching to a dark promotion square without capturing are those from the b-, d-, f-, and h-files, but three of these pawns are still on the board. So White promoted the missing f-pawn, which must have passed f7 and attacked the e8-square. Therefore the black king had moved from e8 previously and can no longer castle.

The ingenious twinning shift of the bishop thus not only legalises the en passant capture but renders castling illegal, to ensure a unique solution in each part. These special moves are complemented by a pair of queen promotions in the actual play and a bishop promotion in the retro play.