Sun Stones has plenty of powerful analytics built into it, so after launch we should be able to tune difficulty to really serve our audience well. But in these final weeks before launch, we're still making blind guesses about difficulty. I thought it might be interesting for our readers to get some insight into our difficulty tuning process, as it's played out these past few weeks.
1 - We want the game to be accessible. This means that we need to avoid places where players cannot proceed.
2 - To ensure that players understand our mechanics, we want to very explicitly introduce our two main mechanics (discarding stones, and creating new stones.)
3 - Creating individual teaching levels for each mechanic means a lot of very simple levels.
4 - Too many simple levels at the outset give the impression that our game is trivial.
5 - Any sort of difficulty spike in the initial levels turns away some significant % of our trial users.
6 - Repeat steps #4 and #5...
We were stuck in this loop for quite a while - how do we get to the good stuff quickly, without moving too fast for some users?
The biggest part of our solution was to break up our mechanics into pods - rather than allowing players to use all our mechanics, we put in place a system to "lock" some mechanics out. This allowed us to introduce just two key concepts, then build complexity around them, before our next teaching sequence. This gives our game more of a shark-fin difficulty flow, rather than an upward-curve. We hope this will create a good experience for all our users - not just one group or the other.
One sort of odd consequence for the "locking" mechanism is that if you go back and play the early puzzles after unlocking more mechanics - you can often get better scores, or earn more achievements. Will hard-core players be upset that they can't earn every trophy on their first play-through? Will players actually go back and discover that they have more flexibility later on? It's not clear.
As a way of selling this mechanism, we've implemented in-game achievements for each unlock. These are closely modeled after Xbox-live achievements, and they create a pretty strong reaction in play-testers who are familiar with that sort of reinforcement. Our assumption is that people who will go back and re-play early levels with new mechanics are probably the sort of folk who recognize achievements - but maybe we'll be proven wrong?
Team is hard at work already, after a long Thanksgiving break.