Standard Deviation

[mmanne]

I just want to check how I'm calculating the SD of my own play, it seems high, but it could be because I'm mixing all of my types of play together. I have a spreadsheet that keeps my time playing and amount won. I've used it to get my mean win.

To get SD, I do
(Sum((Session $ won/Session Hours) - (Avg $ won by hour * Session Hours))^2)/Sessions
This number should give me the variance, so I ^.5 to get the SD.

Does this seem right to everyone here?

Thanks,
Matt

[uuDevil]

You can find some good info here. (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Number=602323&page=&view=&sb =5&o=&vc=1) Look at the posted replies for details.

[BruceZ]

(Sum((Session $ won/Session Hours) - (Avg $ won by hour * Session Hours))^2)/Sessions
This number should give me the variance, so I ^.5 to get the SD.


No, this isn't right. You don't divide by Session Hours until after you take the square. You want to take Session $ won minus expected $ won. The variance should be:

(Sum(Session $ won - (Avg $ won by hour * Session Hours))^2/Session Hours)/Sessions

You could also use the formula from Mason's Essay (http://twoplustwo.com/mmessay8.html) in the essay section. This is the same formula in a different form.

If you want, I can email you a spreadsheet which correctly computes the SD using both of these formulas.