Consensus ratings glitch?
|
On this route: mountainproject.com/v/salt-… |
|
Ummmm, probably just rounding? |
|
The equation rounds down rather than up when a .5 difference is realized. |
|
We recently stopped rounding and moved to a median-based system. Medians are the middle value of an ordered list. With just two votes, a median is one or the other... |
|
Well that's because The opinion based subjective system known as Opinion based ratings Is extremely precise. It takes into account the first voters point system and since he has 17 more points he gets 22% more of the vote. But if the second climber has more pink points then he will get 42% more of the share. If you are an old traddie you get an extra 2% but if you are crusty you get -5%. |
|
Nick Wilder wrote:We recently stopped rounding and moved to a median-based system. Medians are the middle value of an ordered list. With just two votes, a median is one or the other...While I think that's an improvement (& I've advocated this change), it does get a bit funny sometimes w/ even number of votes. Wikipedia entry says: The median of a finite list of numbers can be found by arranging all the observations from lowest value to highest value and picking the middle one...If there is an even number of observations, then there is no single middle value; the median is then usually defined to be the mean of the two middle values |
|
There are at least 3 algorithms for "rounding" the median. Pick one. |
|
Does it matter that the downgrader just did it on TR? I'm sure that should be incorporated into the algorithm somewhere... ;-) |
|
That was a Top Rope FLASH!!!! ;) |
|
It's pretty elementary that the mean of an even numbered list is the mean of the two middle values: if you say you are using the median, you should be doing that, NOT rounding down (or up, or picking the left or right value). |
|
teece303 wrote:It's pretty elementary that the mean of an even numbered list is the mean of the two middle values: if you say you are using the median, you should be doing that, NOT rounding down (or up, or picking the left or right value).+1 |
|
Here is the consensus and all individual ratings from Jungle Blues from Jupiter: Consensus: 5.10c Rick Vermeil 5.11b tbol 5.11a/b Ken Duncan 5.11a Nate Harris 5.10c How do you get a 10c consensus out of those numbers? Does the original poster of the route (Nate Harris in this case) carry more weight? Thanks, David |
|
Yes, Nate Harris is the route submitter, and there are not (yet) enough opinions to overrule him. |
|
how many are needed to overrule? |