If you believe more in evaluation than formal learning, you might shoot for a scoring system that follows a nice bell curve, with some winners, some loosers, and most lumped in the middle.
We like to pretend that academic scores/grades are instructional, but really they are motivational. The reason to work hard is to get a good score, as opposed to using the score to figure out how to improve.
We also like to pretend that scores are scientific, when in fact they represent a massive editorialization of what is important, and by how much. Shooting a big asteroid is 50 points. Shooting a little asteroid is 100 points. Dodging an asteroid is worth nothing.
Scores in educational simulations seem to work best if they are built around the academic standard of
- 69 and below = D = unacceptable
- 70-79 = C = good try but plenty of room for improvement
- 80-89 = B = solid
- 90 - 100 = A = great
This seems natural to those on the outside and a good cultural standard to follow (imagine using an educational simulation twice, the first time getting a score of 24, and the second time, 190), but for designers, squeezing the world of outcomes into that range is as natural as being a yoga master. And putting absolute numbers to actions should make anyone at least a little uncomfortable (try putting numeric values to the activities you did today).
Having scores unquestionably increases usage and focus. But the greater the reliance on scores for motivation, the more students will worry less about learning the material and more about gaming the system - finding logic shortcuts to exploit, and actually subvert the learning.