# Forced selfmate in n-moves

I assume most people on PSE know a bit of chess

With white to move, this is checkmate (1. Qb7#):

That's boring. Let's try something more interesting. With black to move, black can FORCE white to checkmate (1... Qxb7+ 2. Qxb7#):

We can go another step. In this position, white can force black to force white to checkmate (1. Qxb7+ Qxb7+ 2. Qxb7#):

I'm going to call this a forced checkmate in 2 (an abuse of traditional problems where you have to force checkmate in 2).

I can do forced checkmate in 12 (1... Qxb7+ 2. Qxb7+ Qxb7+ 3. Qxb7+ Qxb7+ 4. Qxb7+ Qxb7+ 5. Qxb7+ Qxb7+ 6. Qxb7+ Qxb7+ 7. Qxb7#):

# Can you do better?

NOTES:

• There must be exactly one line after the first move, otherwise it is not strictly forced and I have to think too much.
• The position must be theoretically possible (so not 10 queens of one colour, 11 bishops of one colour, etc.)
• @humn you can just put s (small), m (medium), l (large and default), or h (huge) after the 5 letter imgur address. m works best for me – Beastly Gerbil Mar 9 '17 at 21:59
• This puzzle must have been fun to concoct, amounting to the creation/solution of these examples and others along the way – humn Mar 10 '17 at 1:46
• related, cross-site dupe: chess.stackexchange.com/q/8789/9025 – Herb Mar 10 '17 at 2:38
• @ais523 But selfmates can have multiple branches. Also, selfmates would only be forced checkmate in (odd #). – boboquack Mar 10 '17 at 23:21
• I have a question - why not 2 ... Qbxb7? – Wen1now Mar 11 '17 at 7:31

Yes, I can do better:

I can do forced checkmate in 13 (1... Qbxb7+ 2. Qxb7+ Qxb7+ 3. Qxb7+ Qxb7+ 4. Qxb7+ Qxb7+ 5. Qxb7+ Qxb7+ 6. Qxb7+ Qxb7+ 7. Qxb7 Qxb7#):