You are here
Puzzle points
Forums:
After the first two rounds of the Sudoku GP, I found that some puzzles' point values didn't reflect their real difficulty. Some puzzles' point values are much higher than their real difficulty, such as the Extra Regions Sudoku and the Irregular Sudoku in Sudoku GP2, resulting in 8 competitors with 1000+ points and 42 competitors with (the total) 719+ points.
In my opinion, the point values should reflect puzzles' real difficulty, but should not be too overvalued to scare competitors. The GP should be a test of brain power, not psychology.
I suggest the authors contact more testers and estimate the puzzles' difficulty more precisely. The total points can change, but the scale should remain at 10 pts/min and not change.
Efficiency should be considered
When determining the point values of puzzles, testers should consider not only the actual time they spend on each puzzle, but also their efficiency of earning points. This is because not all testers are top solvers and some testers' "efficiency" of earning points is lower than 10 pts/min. If efficiency is not considered, it must result in some puzzles having point values higher or lower than their actual difficulty.
For example, a tester spends 8 minutes solving a puzzle. If he is a top solver around the world and can get 10 points per minute in GP or WSC/WPC, then his calculated point value can be 80 points. However, if he can only get 7 points per minute in competitions, then his calculated point value should only be 56 points. And then, use the average of all testers' calculated point values as its final point value.
Trial should also be considered
Moreover, because GP and WSC/WPC don't care how competitors get the solution, sometimes a competitor may use trial to get the solution, which significantly reduces his/her solving time. Therefore, to balance, point values should be reduced a little if the puzzle is easy via trial, even it's hard via logic.