Finding the Invalid Move

Yep, still 3 invalid. Though if the point is to find the invalid moves, then I'd prefer to have a unique answer - if not unique board position. :P

Remellion wrote:

Na1+, Na2+, Nb1+, Rg8+, Rh7+, Rh8+

4R3/4R3/8/8/8/1RN4k/6R1/K4B2 w - - 0 1

this almost cooks, except Rg8+ is technically Rgg8+.

Almost, but not quite a cook; I realised that and thought it would make a good try. Although by convention it's written "Rgg8+", "Rg8+" is unambiguous since there's only 1 way to check with a rook moving to g7 which is kind of unsatisfying.

So, let's get some ambiguity back in the picture. No numbers means more possibilities, yes?

(Minimum invalid?) axb*+, bxc*+, cxd*+, dxe*+, exf*+, fxg*+, gxh*+

@Remellion -- I have found a way to make all of them valid. :P

White:

*Pawns: a7, b4, c6, d3, e7, f4, g6

*Rooks: a6, d1

*Bishop: h2

*Queen: h6

*King: b1

Black:

*King: d6

*Other pieces: b8, c5, d7, e4, f8, g5, h7

All of the moves would be valid, but 2 of the moves are promotions. (bxc8=B+ and exf8=B+)

Bravo! Yes, with promotions all are legal, hence "minimum invalid" = zero. Without promotions axb* is illegal, and I was hoping someone would fall for that.

Two in descriptive. Despite the vagueness, I believe there is enough information to decide (and prove!) each.

(1 invalid) KXN mate, KXB mate, KXR mate, KXQ mate
(1 invalid) R-R1ch, N-N1ch, B-B1ch, Q-Q1ch, K-K1ch

@Remellion

The first puzzle's invalid move is KxQ#.

White:

*King on f6, Bishop on b2 and Knight on h6.

Black:

*King on h8, pawns on h7 and f7, rook on e7, knight on g5 and bishop on f5.

The 2nd is cooked.

R-R1ch and K-K1ch seem like the least flexible moves. Here's a position that eliminates only R-R1ch:

And here's one that eliminates only K-K1ch:

As for the first, I can get all but KxQ by fixing GuessWhoIAm's position:

The Q case has too many problems to be valid. Something like this allows both KxQ mate and two other K captures for discovered mates:

The problem is, if I place an R on f4 or e5, that's an impossible check, while a B on either square guards the other square. Not to mention that an N can block the line of the wB.

hxg!+, fxg!+, hxg@+,fxg@+, fxe!+, fxe@+, dxe!+,dxe@+, exd$+,gxh$+

same symbol=same number

What's the minimum number of illegal moves?

Fun stuff guys!

Here’s my own puzzle (white to play, 1 invalid):

Kxd2, Kxe1, Kxe3, Kxf1, Kxf2

nice revival

also

it could be Kxf1 being illegal

throw a white king on e2, black knights on d2 e1 e3 and f2

and a black bishop goes from h3 to f1 check

proof:

for f1 not to be protected (and instead something else)

then there need to be non-knights on d2 and e3

bishop on d2 protects e1 and e3, bishop on e3 protects d2 and f2 each make two others illegal

rook on d2 protects f2, and has to move last to check

then rook on e3 protects e1, and has to move last to check

but black cant move two pieces in a row, so f1 has to be protected OR if it is not protected two other squares need to be protected (e1 and e3, d2 and f2)

edit: thus f1 is the only illegality that can exist alone, and that is the only arrangement, with only one other black last move (Bg2-f1+)

Kxf1 is correct! However, two things: pawns can be captured too, guarding only the adjacent squares diagonally in front of them. So if there was a black pawn on d2, it guards e1 but not e3. Also, if the move mentioned in your proof is invalid, then your logic might not be true for that case. For example, if Kxf2 is invalid, then the rook on d2 guarding f2 doesn’t mean anything (but it still gives check).

There is a very close try with Kxe1 being invalid. The piece on e3 must be a rook and the piece on f1 must be a bishop. But the double check by bishop and rook is illegal, so Kxf1 is the only illegal move.

how could i forget pawns...

also, your post has some errors that make it confusing in terms of the square names

well, i mean if i find a position that fits an illegal move, then i can post it w/o proof and leave it to others to maybe cook/prove it

Sorry for the errors and any confusion caused. For some reason I shifted the original position 1 file to the right and ended up mixing the squares.

Anyway, another one to try (1 invalid): O-O-O, Kxd1, Kxd2, Kxf1, Kxf2

well 0-0-0 and Kxd1 obviously cant coexist (if a piece is on d1, castling is impossible)

so which one...

black queen bishop pawn AND rook on d2 prevent castling, but a knight doesnt!

but then the knight on d2 defends f1, so this try fails

thus 0-0-0 must be the invalid one

but wait

d1 must be a bishop (pawns arent on the first rank, knight defends f2, rook and queen defend d2)

same logic for f1

but then using the same logic, d2 and f2 must be bishops (or pawns)

but then, that creates illegal double check!

But when did I say white to move? Pawns can work then.

nonono

it would be double check

Like this:

aah

how could i be so dumb