Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

Statistics are crap

Name: Anonymous 2006-03-14 15:10

How can the study of chance and probability ever be meaningful when so many unknown and completely random factors occur?  So many unnacounted for events can occur that can never be predicted by statistics, especially in relation to individuals and individual events that do not fit any sort of generalization or statistic. 

For example, this TV program asked some engineering students to build a giant web that would catch a plastic fly when launched from a trebuchet.  When it actually launched, the strands of the web were clearly too weak to catch it on impact, but it happened to catch on a supporting strand and dangle above the ground.  The calculations were for any part of the web to withstand the impact, yet they were saved by a "lucky" event when the fly could have easily fell to the ground.       

Name: Anonymous 2006-03-14 20:36

SCIIIIEEEEEENNNNNCNCCCCCE!

Name: Anonymous 2006-03-14 20:50

statistics are all made up

Name: Anonymous 2006-03-15 4:33

>>1

I agree.

Name: Anonymous 2006-03-15 4:36

>>3

37.2% of statistics are made up.  Also, 4.8% of Americans have an unhealthy obsession with statistics.

Name: Anonymous 2006-03-15 9:45

Never trust statistics that winston churchill did not fake

Name: Anonymous 2006-03-16 12:59

Why take a statistics class then?

Name: Anonymous 2006-03-16 13:10

all generalizations have exceptions. if they did not have exceptions, they would be called rules or laws.

generalizations are like predictions of what would happen. if 90% of the time the web catches the fly, then there is a 90% chance that a particular fly will be caught by the web.

Name: Anonymous 2006-03-16 20:50

Assuming an event will occur nine times out of ten, how likely is that event going to occur nine times out of ten given any number of occurences?  What do you call it when each instance is lower or higher than your predicted average of 9 out of 10?  What do you call a freak occurence in your experiment like a tornado or a riot or a meteor interrupting your results? 

Name: Anonymous 2006-03-16 21:31

PATTERN DIFRACTION!

Name: joetaproot 2006-03-17 1:44

those occurences are accounted for in the percentile, wouldnt they be? lets say for example, 75% chance of whatever outcome is solid chance, and the next 15% are freak occurences. you have 3/4 chance of that occurence due to chance, and the other 15% is random, so 90% chance of it happening, through chance or accident. some of the time anyway.

Name: joetaproot 2006-03-17 1:44

by random, i mean accidental*

Name: Anonymous 2006-03-17 15:24

What about a steady occurence of low results, like 20 in a row of "3 out of 10."  Wouldn't the rate go up and down rather than stay at an average of 90%?  What would you call this average fluctuation?

Name: Anonymous 2006-03-18 7:07

True randominess does not exist.

Name: Anonymous 2006-03-18 8:32 (sage)

>>14
LOL YES IT DOESD LOOK AT QUANTEM MECANICS STUPID!
YOU LOOSE!
I R D4 L33T WINZ0R!

Name: Anonymous 2006-03-18 10:35

>>13
A time PARADOX

SCIIIIEEEEEENNNNNCNCCCCCE!

Name: e !9ggCfw77Kg 2006-03-20 17:00

>>1
while unknown factors exist, there are many situations where their effects are negligible and allow for probabilities to be calculated accurately enough that they are representative over a large time period or sample.
if you want an example of a more complex situation: life insurance companies need to make a profit while still being competitive.  so they need to assess risks to set prices and maintain profit margins without overcharging.

Name: Anonymous 2006-03-21 7:05

What's unpredictable canot be foreseen.Would you be ready to gamble your life on a dice that should hit three or four just becoz statistics tell you they're the highest frequency of appearence?(no threats this is just scheme!)I wouldn't.

Name: Anonymous 2006-03-21 10:43

no, but thats because i know from statistics that it would be a retarded thing to do.  i wouldn't bet my life that the next roll of a die would not come up 1, but id bet 5$ it wouldnt if i was paid back $6.50, net $1.50, for each time it wasnt, and if i got to play 1000 rolls.

Name: Anonymous 2006-03-21 16:32

Many statistical studies actually do take randomess into account. After all, most random noise is Gaussian, and therefore, given that enough is known about the properties of the Gaussian noise, it can be filtered.

Name: Anonymous 2006-03-21 18:25

>>20
Most random noise is random

Don't change these.
Name: Email:
Entire Thread Thread List