High Times

[Turboski]

Admins - Curious how some of the super high times mess with the numbers and averages?

For instance, I'll forget I'm playing (phone call, kids, job, wife, meteor strike, cat, dog) and come back and finish a map and it says I finished in 10,000+ seconds.

Any thoughts on 'cleaning' up the database to rid it of times that are clearly on par with this type of thing? Plus, the cheater low scores? Are there enough real games played where the outlying high times don't skew the results enough to matter?

[boldly-going]

I have the same thing quite often, and then I see my time be at the end of the bell curve. Ideally the bell curve should make it so it doesn't skew so bad as its out from the standard deviation but it's been so long since I did bell curves that I can't remember how outliers affect it... I still feel guilty when it happens though like I'm messing with the average lol.

[admin]

After we get about 6 months of real world solve data (this will happen around February) we will implement the finalized data system that should fix most of these issues. Averages will not be skewed by super long solve times as we will move to a median instead, where extremes at either end will have no effect on the median value.