Standard deviation...?

Is anyone able to explain standard deviation to me without frying my brain?

How far from the mean the average data point lies. Example:

You have two data sets with 3 data points in each:

Data Set #1: 0, 5, 10
Data Set #2: -20, 5, 30

The mean for each data set is 5, but the standard deviation for data set #2 is much higher.

EDIT: And just for future reference, the standard deviation is the square root of the variance. So when players talk about high amounts of variance, they mean there's large fluctuations in value around what would be the average result.

Or if you like the picture explanation:

http://upload.wikimedia.org/wikipedi...deviations.svg

OK, so if those two data sets were poker session results in dollars for two different players although they've both made the same amount of money player 1 is the 'steadier' player with a lower standard deviation - am I right in saying then that the standard deviation for those two players would be 2.23 ($2.23)?

Player 1 is the steadier player.

However, the standard deviation for data set #1 is 5.
The standard deviation for data set #2 is 25.

OK, so where does this bit come into it?

Because just about everyone says "variance" when describing swings instead of "standard deviation".

http://en.wikipedia.org/wiki/Standard_deviation

Simplified it says that about 2/3 of the time(68%) a stat will fall within 1 sd of normal. Meaning the other 1/3 of the time it will be an abnormal deviate....... ohhhhhh!

But that abnormal deviation could be on either side of the normal top of the bell curve.

SO using the 80% of the time you get AA you will win, it suggests that 1/6 percent of the time you will get a horrible bad beat, (as opposed to a not so horrible bad beat).......

c9 vs dj


final battle, FIGHT!

I have no idea what he's talking about.

(neither do i)

ps I'm not on anyone's side, just wanted to post a picture from Mortal Kombat

All you need to know is that standard deviation = variance, and VARIANCE SUCKS! Got it?

There, now you've got it right.

In order for a standard deviation to be defined there has to be enough data so that a bell shaped curve can be seen on a graph when plotted. You can not have a sample size of three points and get a graph. Well you can but it will not be meaningful.

The range of possible points need to be taken into consideration. So taking the wiki link in C9's 1st response http://upload.wikimedia.org/wikipedi...deviations.svg there are 2 different graphs overlaid on the same chart, showing 2 different standard deviations.

So a standard deviation is some value that represents a deviation from the average over a sample data set. If we use trackers for our poker data set we approach a perfect data set and a Standard Deviation can be defined, or agreed upon over a large sampling.

But you could be sampling cars colors as a game with you kid while you drive between Kansas City and Tulsa. She counted maybe 1000 cars and plotted them on a color basis and found that 60% were either black or white. You might determine that the 60% number constitutes a standard deviation if the numbers show that the black/white ratio was approximately equal. If so then a standard deviation could be defined as 30% (either way on the graph) and a pink Mazda with Flames (you saw 1) is not within the standard deviation, but does represent one extreme on the graph, while a yellow bug with Tall tree painted on it (also saw 1) would represent the other extreme.

SO .....

Your head has exploded and your eyeballs are looking for a socket to hide in.:confus ed:

Point is that for the most part you need a bigger sample size and a definition of what you are plotting before a standard deviation can be determined and thus possibly useful.

Another try......

Truly simple answer, based on the responses given so far is NO. But the sample size of the answers given here is small, (but hopefully growing) and at this point a standard deviation can not be determined With any hope of it being useful info.

There may be ways to use a SD in a small sample size to make a reasonable decision. This would require using a value for a SD determined by some previous graphing with specific parameters and then following those parameters exactly. For example if you know that the average age of dying is 72, and you have a predetermined SD of (guessing) 6 years, and you are 39 then you can feel comfortable about not dying for a few more years.

If you understand that, you should get in the Insurance Biz.

You make it sound like SD is really hard. LOL.

Trying to prevent a brainfry.....


It isn't hard if you understand the basics of statistics, how samples are gathered, and plotted and some context of how the members of the data set are related.

So for example you can't really graph the chart of the color values of both oranges and apples and determine that within 1 standard deviation you will get a pomegranate......