Interesting charts on Elo spread

#21Dark_DcPosted 1/9/2013 11:38:20 AM
gg_easy posted...

You could rejig the categories under which the data is grouped and change the modal category to be anything. I would bet that if you took the same data set and made the bar chart have categories ..., 1050-1150, 1150-1250, 1250-1350, ... the modal category would now be 1250-1350, and then you might say: 'most people are 1250-1350' elo, a "free" 50 elo increase for most people, which is of course nonsense.

You can say that most people are above 1200 elo, or most people are below 1300 elo, or take the interquartile range. The modal category doesn't tell you much about the population itself, because you can change it by representing the data differently.


Ehhhhhhhhh....

You need the other data like SD, mean, etc. to really say this. Because theoretically you can change any graph to suit your needs but the whole point is deciding what the appropriate categories would be. This chart seems a fair assessment and makes a lot of sense given the start point and such. If you did 1200-1500 they'd fall there, if you did 800-1500 they'd fall there, etc. but it wouldn't change the fact that the mean would be, lets say, 1234.
---
The snozberries taste like snozberries!
LoL: dantheautomator
#22EDumeyPosted 1/9/2013 11:38:51 AM
I think the word you're looking for is plurality.
---
http://i.imgur.com/ZhLrT.jpg
http://i.imgur.com/H6Frl.jpg
#23gg_easyPosted 1/9/2013 11:54:19 AM
Dark_Dc posted...
gg_easy posted...

You could rejig the categories under which the data is grouped and change the modal category to be anything. I would bet that if you took the same data set and made the bar chart have categories ..., 1050-1150, 1150-1250, 1250-1350, ... the modal category would now be 1250-1350, and then you might say: 'most people are 1250-1350' elo, a "free" 50 elo increase for most people, which is of course nonsense.

You can say that most people are above 1200 elo, or most people are below 1300 elo, or take the interquartile range. The modal category doesn't tell you much about the population itself, because you can change it by representing the data differently.


Ehhhhhhhhh....

You need the other data like SD, mean, etc. to really say this. Because theoretically you can change any graph to suit your needs but the whole point is deciding what the appropriate categories would be. ... If you did 1200-1500 they'd fall there, if you did 800-1500 they'd fall there, etc. but it wouldn't change the fact that the mean would be, lets say, 1234.


I think you just said what I said. Except you didn't need the other data like SD, mean to say it either.
---
easy
#24Dark_DcPosted 1/10/2013 6:01:26 AM
gg_easy posted...
Dark_Dc posted...
gg_easy posted...

You could rejig the categories under which the data is grouped and change the modal category to be anything. I would bet that if you took the same data set and made the bar chart have categories ..., 1050-1150, 1150-1250, 1250-1350, ... the modal category would now be 1250-1350, and then you might say: 'most people are 1250-1350' elo, a "free" 50 elo increase for most people, which is of course nonsense.

You can say that most people are above 1200 elo, or most people are below 1300 elo, or take the interquartile range. The modal category doesn't tell you much about the population itself, because you can change it by representing the data differently.


Ehhhhhhhhh....

You need the other data like SD, mean, etc. to really say this. Because theoretically you can change any graph to suit your needs but the whole point is deciding what the appropriate categories would be. ... If you did 1200-1500 they'd fall there, if you did 800-1500 they'd fall there, etc. but it wouldn't change the fact that the mean would be, lets say, 1234.


I think you just said what I said. Except you didn't need the other data like SD, mean to say it either.


Did I? my bad. I thought you were making a different point lol
---
The snozberries taste like snozberries!
LoL: dantheautomator