 Pages

 1 Charles Johnson  Mar 30, 2014 12:37:30pm
 2 b_sharp  Mar 30, 2014 2:44:14pm

Should

For mathapobes, hold

be mathophobes?

 3 b_sharp  Mar 30, 2014 2:49:36pm

The sentence

“The variance is determined by the following equation:”

The variance is the square of the Stddev. so it’s the math before the square root.

 4 b_sharp  Mar 30, 2014 2:54:37pm

A real life example of the importance of the stddev is the 1998 global temperature anomaly which is 2.5 stddevs from the linear 10 year moving trend line. This means it was highly unusual and should not be used as a starting point for any linear trend line.

 5 b_sharp  Mar 30, 2014 2:57:52pm

Excellent place to start with the math Obdi.

 6 Fairly Sure I'm Still Obdicut  Mar 30, 2014 3:24:53pm

re: #5 b_sharp

Excellent place to start with the math Obdi.

Thanks, fixed.

re: #4 b_sharp

A real life example of the importance of the stddev is the 1998 global temperature anomaly which is 2.5 stddevs from the linear 10 year moving trend line. This means it was highly unusual and should not be used as a starting point for any linear trend line.

Yeah, excellent point. One great use of standard dev is when someone gives you just one number from a set, you can ask what the z value is.

Speaking of which, I should mention that standard deviations are measured in Zs.

 7 EPR-radar  Mar 30, 2014 4:33:38pm

Well done. Some minor nits about the math:

1) The mean in the equation is the x with a bar over it, not X as stated below the equation.

2) If possible, it might be useful to have the notion be something like sum_i (x_i - xbar)^2 to show that the x_i (i.e., the sample points) are the things that vary from term to term in the sum.

3) (probably too technical). The N-1 in the denominator makes this the formula for the sample standard deviation as opposed to the population standard deviation (which would have N in the denominator).

 8 b_sharp  Mar 30, 2014 5:00:07pm

Well done. Some minor nits about the math:

1) The mean in the equation is the x with a bar over it, not X as stated below the equation.

2) If possible, it might be useful to have the notion be something like sum_i (x_i - xbar)^2 to show that the x_i (i.e., the sample points) are the things that vary from term to term in the sum.

3) (probably too technical). The N-1 in the denominator makes this the formula for the sample standard deviation as opposed to the population standard deviation (which would have N in the denominator).

Obdi was fairly clear about 1) and 2) in his description. Without using a program with math symbols it becomes a tad tough to match images.

 9 EPR-radar  Mar 30, 2014 5:09:12pm

re: #8 b_sharp

Obdi was fairly clear about 1) and 2) in his description. Without using a program with math symbols it becomes a tad tough to match images.

Agreed 100%. I’m just enough of a geek that I read the equations first, and then see if the text matches how I read the equations.

 10 b_sharp  Mar 30, 2014 5:53:08pm

Agreed 100%. I’m just enough of a geek that I read the equations first, and then see if the text matches how I read the equations.

GEEK!!
GEEK!!

cough, cough.

I’m kind of a geek too.
A really old one.

Not as good at stats as you and obdi though.

 11 wheat-doggha -- oo bird outside my window  Mar 30, 2014 7:35:55pm

Well done! This math teacher approves.

 12 abolitionist  Mar 30, 2014 10:35:11pm

Nitpic: The graphic image is grossly inaccurate; here is a better one: en.wikipedia.org

 13 Fairly Sure I'm Still Obdicut  Mar 31, 2014 4:05:18am

re: #12 abolitionist

Nitpic: The graphic image is grossly inaccurate; here is a better one: en.wikipedia.org

Shit, I’m bad at visual estimation.

 14 iossarian  Mar 31, 2014 8:27:18am

Here’s an interesting thing (I think) about the normal curve: you can think of it as the way that lots of individual binary (yes/no) decisions yield an aggregated range of outcomes.

In the case of, say, people’s weight, in a very simple model, say we all start off at the same weight, and we all decide every day whether to eat salad or hamburgers for lunch. Let’s say that overall it’s just a 50/50 choice. To put it another way, people literally flip a coin every day to decide what to eat.

Assuming that eating a hamburger makes you slightly heavier and eating salad makes you slightly lighter, if everyone behaves this way then what you end up with is a normal curve of weights.