I am developing a game where based on steps to be performed in a particular level, time and lives are pre-calculated. As the level gets completed, remaining lives and time are retrieved. Calculating score based on remaining time and lives will not work in this system, as the same level with smaller number of steps will have less time and lives than level with more steps. Can I count score based on time and lives efficiency? If yes, then how to calculated the score so that it is unbiased? I am stuck as it is my first game development. Any constructive criticism is appreciated.


1 Answer 1


Do nothing

Consider that you are deciding the time and lives to complete the level. If the margin the player has to complete the level remains the same, then the player would complete both the short and the long levels with about the same time and lives remaining. In which case there would no need to adjust for it, as the adjustment was already done when you decided how much time and lives to give the player.

Also, perhaps correctness is not the goal. You might want it to be easier or harder to get points on advanced levels.

Simple normalization

The idea is to use a proportion. Or percentage, if that makes it easier for you to reason about. So instead of using the remaining time directly, you use how much is the remaining time of the total time. And the same idea for lives.

This would be correct the time and lives you give the player grows linearly with the complexity of the game. Again, perhaps correctness is not the goal.

Complexity metric

You would have some complexity metric that tries to capture how hard the levels are. Then you do some computation over your complexity metric to come up with the time and lives.

An example of complexity metric would be the number of steps needed to solve the level. Yet, it does not consider all the alternative options the player might try, only the ideal path. The more things the player could do at any particular moment, the more work they might need to figure out the solution.

Submitted to your consideration: We know that any - standard - Rubik Cube can be solved in at most 20 steps. But that 20 does not capture all the possible moves that do not solve the Rubik Cube, nor the skill needed to know which moves to do in order to solve it.

So, the actual complexity of the game might actually be an exponential function of the number of steps needed to solve the level.

It might be a better idea to incorporate your complexity metric into the scoring directly.

So you can use the quickness to solve the level in complexity over time for scoring. In other words, the score could be based on the complexity of the level divided by the time the player took to complete it.


What do engineers do when they don't know? They fuck around and find out do basic research.

Which in this case would mean to get a set of test players to try the games and record how much time and lives they take to complete the levels, and compare it to the time and lives you would be given them.

This will give you some idea of:

  • How generous is the computations you are using to give them time and lives.
  • How does the time and lives used by the players scale with your complexity metric.

You might also discover that the player struggle in unexpected ways and times. Even might struggle more in levels you consider easier. Also, players might give you ideas.

And then, having your results, you can decide if you are going to adjust the computation, or the score, or both.

And then you repeat the experiment. This time adding more test players so you have a mix of people who have tried the game before and other that don't. And perhaps you do further refinements or stick with what you got.

Ultimately what you do is tweak values and pick something that works in practice. Even if you don't fully understand it, you would have empirical evidence that it works.

Other Alternatives

Perhaps it is better to add collectibles for points, or the game design allows for style points.

Or perhaps the game does not need score at all. Losing a live is already lost time, so a leader board based on time might be enough.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .