Board Thread:Wiki - Running & Policies/@comment-3225604-20171127232502

"Variable" difficulty is one of the most hotly debated difficulty ratings in this wiki. However, some levels that are not "variable" have every reason to become "variable". Here I propose we should grade a level "variable" if the standard deviation of difficulty exceeds 2.1 instead of 3. This means it is common for a player to find a level >2 difficulty ratings away from the median difficulty. I would also probably suggest that it must receive at least 10 votes of any graded difficulty for the standard deviation value to be valid.

There are currently 3 candidates that fulfill this criteria:
 * Level 313 has a standard deviation of 2.22-2.29 and receiving 20 votes of a graded difficulty. A good example of votes distributing equally among different difficulty labels, by that I mean across very easy to extremely hard. This makes the level a good candidate to be variable.
 * Level 880 has a standard deviation of 2.27 and receiving 45 votes of a graded difficulty. It seems the polarization of difficulty is a consensus for this level. It is alleged that our reaction time may alter the difficulty of this level considerably. You can find about 9 users voting for easy while about 11 voting for extremely hard, showing the polarization of this level.
 * Level 2227 has a standard deviation of 2.19. A less convincing candidate I would say as it only receives 11 votes a graded difficulty. However, votes are scattered across difficulty ratings and I think this would make a good candidate for "variable".

Notable candidates that do not fulfill this criteria:
 * Level 463 has a standard deviation of 1.94-1.98

Of course, if a level receives 60% or higher support votes for variable. The rating would be assigned.

 