David Edery on how hard (or easy) you should make your games:

Too many of us are still holding onto design philosophies that were born in the days of quarter-gobbling arcade games. Too many developers get most of their design feedback from QA teams made up of hardcore gamers who have played a game way more than most normal people ever will. Making a game “just hard enough” (be that very hard or very easy, depending on the person playing) is one of the primary keys to fun — and, I think, an under-appreciated way to significantly increase sales. It deserves more attention from our industry, even as we search for ways to incorporate meaningful, educational, and remarkable consequences back into our games.

I’ve long been a fan of the approach of having multiple difficulty levels at once in the same place, using things like optional badges, multiple levels of success and bonus objectives. The simplest form can be found in most racing games, which allow you to pass a race in 1st, 2nd or 3rd place.

Medals, optional missing objectives, secrets, collectibles, level (and game) completion percentages - all of these allow you to have more than one level of difficulty on the same map at the same time, which can substantially reduce QA time and other design problems that come from a situation where you need to run the same content more than once during testing because the rules have changed in some way. If advanced players have the same experience as regular players, except that they skip less, a lot less can go wrong.

David Sirlin’s excellent analysis of Donkey Kong Country 2’s secrets was the first writing that got me thinking this way. Time and time again, working on small games with tight deadlines and short QA cycles, we took advantage of this technique.

This is not to say that it’s impossible to do dynamic difficulty well. People smarter than me are already working on better automated ways of adjusting difficulty in real time and presumably, they’ve solved the QA problem. I wonder how they’ll solve the emotional problem. Some people love being frustrated by games and some people hate them. Until game systems can detect how mad you are, the system will have to err in one direction or the other.

A fixed difficulty with a range of levels of success is the best of both worlds. Instead of dynamically adjusting difficulty is that it allows the player to decide for themselves how difficult they want the game to be, in real time, in a highly contextualized way. If the one section is too frustrating, then they can ignore the side missions and just get things done. If another is going really well, they can reach for the gold. If it’s going poorly but they are still enjoying themselves, they can reach for the gold anyway.

Plus, it makes it easier to compare the size of your achievements.