The following comment came on this page by email, from `Ulrich'. Ulrich,
thanks for the comment.
I have a remark on the shatranj problem 'Check!'.
There is a shorter solution to the problem than depicted here.
On move 4, the white knight should go to f7 instead of c6, which
would cover the escape square on g5. This gives 4. ...Ne4 (just
delaying one move) and then 5. Rxe4+, Kf6 6. Nd5 mate.
In addition, if white plays 4. Nc6+, then 4. ...Ne4 is bad because
it follows: 5. Rxe4+ (covering the pawn on g4!) and then 5. ...Kf6,
6. Nd5+, Kg5 and 7. Ee3 mate, allowing the most useless piece the
final strike :) This is still shorter than the reported line.
Given that you repaired the problem with the c7 pawn already, my
guess would be that the author did not intend this as a problem
in todays sense but rather as a nice series of checks.