Re: Announcing: Fyleet, Crobe, Sangraal

[Gareth Rees <garethr@cre.canon.co.uk>]

J R Partington <pmt6jrp@gps.leeds.ac.uk> wrote:
> No, you've missed the point. The sort of puzzle that we have in mind
> is one where there is a unique route through a maze, which is told to
> you in a cryptic message by a character in the game (and different
> players of the game will be given different routes and hence different
> cryptic messages).

It's often possible to arrange things so that the game can *tell*
whether or not the player has really solved the puzzle, or is just
guessing.

For example, suppose an adventure game had the "find one counterfeit
coin out of twelve in three weighings on a balance" puzzle [1].  The
writer of the game could use the following implementation:

  * keep track of which possibilities the player has ruled out in the
    weighings so far;

  * when the player sets up a weighing, choose at random a result that
    is consistent with the remaining possibilities;

  * at the end when the player has revealed which coin they think is
    counterfeit, let them win only if there is a single possibility
    remaining and they named it correctly.  If there are multiple
    possibilities remaining, the game reveals one of the possibilities
    the player didn't name and thus the player loses.

With this implementation, saving the game does the player no good.
Neither does guessing randomly.  So the player must really solve the
puzzle.

This is obviously a bit more work for the author -- and not all puzzles
can easily be made to fit this kind of implementation strategy -- but I
think it gives a much better impression of fairness to the player than
just denying the player the opportunity to save the game.

[1] Don't do this!  It's been done before.

-- 
Gareth Rees