 |
|
| Science Forum Index » Statistics - Math Forum » Number to represent both Avg and StdDev... |
|
Page 1 of 1 |
|
| Author |
Message |
| paroxsitic... |
 Posted: Fri Nov 06, 2009 2:01 am |
|
|
|
Guest
|
Whats the best way I can represent data with ONE NUMBER taking account
the average and also the deviation.
What I have been using is the Avg - (StdDev - AvgStdDev)
This is for NFL fantasy football analysis, Where AVG is the average
difficulty of their schedule (based on 0-100). The StdDev is the
standard dev. of the difficulty of their schedule. And AvgStdDev is
the Average standard deviation of all the other team's std dev's. |
|
|
| Back to top |
|
|
|
| Russell... |
| Posted: Fri Nov 06, 2009 5:57 am |
|
|
|
Guest
|
On Nov 6, 7:01 am, paroxsitic <paroxsi... at (no spam) gmail.com> wrote:
[quote]Whats the best way I can represent data with ONE NUMBER taking account
the average and also the deviation.
What I have been using is the Avg - (StdDev - AvgStdDev)
This is for NFL fantasy football analysis, Where AVG is the average
difficulty of their schedule (based on 0-100). The StdDev is the
standard dev. of the difficulty of their schedule. And AvgStdDev is
the Average standard deviation of all the other team's std dev's.
[/quote]
There probably isn't one "best way". The quantity
you want to calculate is what best represents the
concept you're trying to study. By definition
you're always going to lose some information when
you do this, but you gain the simplicity of looking
at a single quantity. The canonical statistical
quantity that combines mean and standard deviation
is probably the coeffecient of variation, SD/MEAN,
but that may not be related to what you're interested
in. This may not seems helpful, but if
Avg - (StdDev - AvgStdDev) tells you something that
makes sense to you in the context of your application,
fantasy football, I don't know any reason off the
cuff to suggest anything different. If you have not
tried these, look at Avg / (StdDev - AvgStdDev) or
(StdDev - AvgStdDev) / Avg, which are kind of
hibreds of what you've been using and the coeffecient
of variation and see if either works better, but
beyond that I have no suggestions.
Cheers,
Russell |
|
|
| Back to top |
|
|
|
|
|
All times are GMT - 5 Hours
The time now is Thu Dec 10, 2009 8:47 am
|
|