Login with Facebook
 ADVANCED
Median vs. mean consensus grades
View Latest Posts in This Forum or All Forums
   Page 1 of 1.  
Follow replies to this topic? Notify me at the top of web site.
1

Email me.
 
 
By MikeJ
Jun 26, 2012
Consensus grades on the site are currently calculated using the mean opinion. Switching to the median opinion would make the estimate more robust to outliers, and would prevent this sort of nonsense:

mountainproject.com/v/debbie-d...

(Administrator decides he doesn't like a 5.11a climb so he gives it a 5.0 rating and screws up the consensus grade.)

FLAG
By Nick Wilder
Site Landlord
From The Bubble
Jun 26, 2012
Personal Photo
I deleted the bogus ratings, and the route comments related to them.

FLAG
By agd
Jun 26, 2012
alaska
If the sample size is large enough and its normally distributed then the mean and median should be the same. With that said, I'm not sure what the sample size is...

FLAG
By Darren Mabe
From Flagstaff, AZ
Jun 26, 2012
wham bam hand jam. Wrapping up the final moves of Twist of Fate, Oak Creek Canyon. <br /> <br />photo: Blake McCord
alexdavis wrote:
If the sample size is large enough and its normally distributed then the mean and median should be the same. With that said, I'm not sure what the sample size is...

beat me to it.
except there is nothing normal about the distribution on MP!

FLAG
By slim
Administrator
Jun 26, 2012
tomato, tomotto, kill mike amato.
i agree that the median is the way to go.

FLAG
By agd
Jun 26, 2012
alaska
Young Gun wrote:
Switching to the median opinion would make the estimate more robust to outliers, and would prevent this sort of nonsense: mountainproject.com/v/debbie-d... (Administrator decides he doesn't like a 5.11a climb so he gives it a 5.0 rating and screws up the consensus grade.)


agreed.

FLAG


Follow replies to this topic? Notify me at the top of web site.
1

Email me.
Page 1 of 1.