Quote:
Originally Posted by Virago
To develop a system with minimal margin of error is big permutational math.
-
No way to make it better without making it longer
-
criticism on your current system: the system you have is only good for choosing a #1 spot and it only works if you have a runoff with the #1 choices from each bracket (which you did not). again, none of the ranks up top are backed mathematically. mathmatically speaking you put a bunch of different shaped blocks in a bucket closed it shook it around for awhile opened it up and randomly picked blocks from the top of the bucket declaring them #1 though #20 as you chose each
-
You use of wildcards is pointless. the use of wildcards is only applicable to systems involving multiple tiers and your system only has one tier. You use them as "honorable mention" places which is even more pointless is a system where even the #2 spot (had you followed through to the end of your chosen system) has no basis in mathematical proof
-
maybe the most effecient kind of system for what you're trying to do (using simple math) would be to have multiple tiers where anyone who has better than x% more than the the median ('median' not 'mean/average') number of votes in their flight moves on to the next round with no limitation on how many entries can proceed to the next tier from any one flight. the lower your x the more accurate the results, but the longer the entire process. the value of x could be changed between tiers but not between flights. x could also have a negative value
-
Continue that process until you have your final 20 (could take several tiers) and host one last poll. then you have the ability to declare a 1st through 20th place based on the results of that single last poll. simple, but so long as you don't have too many instances where many besides the top [100% -(50%+x%) ] are "contenders" in their flight until the last round (and if they are, you can always go back and edit the entire tier's results with a new x value) it should be decidedly more accurate than what you have here
-
a longer process but a tons more accurate
-
and there could be a more elegant way to do it. this is only one of an infinite number of ways but it should give you an idea of what is needed to be done
|
i think at this point you need to draw pictures and stuff. If you went over my head I think you went over the head of 90% of the people here.
I haven't pulled out the trig calculator in a long time, i wanted to keep it somewhat simple. It's a poll.
I can appreciate the time and effort you put into writing that, but
Quote:
|
maybe the most effecient kind of system for what you're trying to do (using simple math) would be to have multiple tiers where anyone who has better than x% more than the the median ('median' not 'mean/average') number of votes in their flight moves on to the next round with no limitation on how many entries can proceed to the next tier from any one flight. the lower your x the more accurate the results, but the longer the entire process. the value of x could be changed between tiers but not between flights. x could also have a negative value
|
Sorry man that is way too complicated, next time maybe i will just do the poll on my site and do all 320 at 1 time.
I follow (i think) on what you saying, how in the hell am I suppose to explain to everyone voting how it works.