Once the "generator" for prompt-selection mechanism of Numb3r Stations had been decided, listening to a number station, there was still a LOT to be done to make it both usable and fun. It's a very clear example of the difference between design and development, and I'm very thankful I got Craig's help on it.

Using three digits (which I call the Key) to pick a prompt wasn't as straightforward as I'd initially imagined. 

You could just add the digits until you got a single number, but that was not very "spy-like". Other methods were too involved, and often encountered problems if any of the Key's digits were zero. I played around with geometric shapes, "choice trees" and various cyphers. I got a mathematician friend involved who described what I was attempting as "creating a fun hash function", and that rang alarm bells that maybe it wasn't as fun as I thought.

After some research into code decryption, the best method I could come up with was dividing a sequence of numbers up into sections depending on the digits in the key, then selecting a number within a specific section using they Key as the guide. But again, this ran into the trouble if the digits were too high - it became busywork, a tedious exercise of counting rows.

Fortunately it was at this point that Craig came in and solved the issues with two very elegant fixes.

The first, which should have been obvious from the start, was simply to say "zeros count as 10s", which solved "the zero problem" in one fell swoop. A classic example that as a designer I couldn't tell the wood from the trees and had gone way too far down the rabbit hole.

The second was to have multiple sequences of numbers instead of just one for the whole game, and have the Key select one of them. These sequences then became a 10x10 grid (the one time pad), and the second and third digit of they Key chose a column and a position to guide the player to the prompt. 

Simple, less like busy work and more like decoding, and crucially it opened up avenues to ratchet up the tension during gameplay.

You can't have a spy game without tension.

The decoding method Craig came up with meant we could add mechanics on top so there was a bit more game instead of just journaling prompts.

We decided to come up with a 'success system' which, counterintuitively, made success increase the chances of your character being caught.

This made thematic sense as the more successful a spy would be, the more likely it is counterintelligence will be looking for them. It added a bit of spice to what was otherwise a simple writing exercise, increasing tension as the game went on, and it allowed us to add the "Burn the Evidence" mechanic which gives players a valve to release some of that tension.

So finally we had prompt selection and mechanics to add some oomph to the game - but writing is a major part of the fun right? That's where we came up with codes, which I'll talk about in the next entry.