Ducati Monster Forum

Kitchen Sink => No Moto Content => Topic started by: SacDuc on December 21, 2010, 12:59:11 PM

Title: Math Question
Post by: SacDuc on December 21, 2010, 12:59:11 PM


I suck at math so please pardon if I have my terminology all messed up.

Let's say you have an event that you believe is a 99:1 shot. How many independent trials would you have to run to be 95% sure that there is actually a 1% chance of this event happening?

Is there a formula where you could plug in odds of something happening (e.g. - 3:1 or 25%) the percentage of accuracy you need to confirm those odds to (presuming you only get perfect accuracy with an infinite number of trials) and get the number of trials you need to run.

Let me know if I need to clarify.

Thanks.

sac
Title: Re: Math Question
Post by: derby on December 21, 2010, 01:07:35 PM
well, technically, you could run the same 99:1 trial 100 times and get the same result every time.

Title: Re: Math Question
Post by: il d00d on December 21, 2010, 01:11:38 PM
I am not a math nerd, but remembering back to my statistic class probability always assumes an infinite number of "tries" -  if the probability of something happening is 10:1, in infinite amount of tries, this should be the actual outcome.  If you flipped a coin 10 times you might come up heads seven times, tails three, but the head to tail ratio would change the more you flipped the coin.  So, are you asking how many times, given a probability, you would have to try something to have the outcome come within 95% of the probability?  If so, I think that is a standard deviation problem, and I will be damned if I can remember how to do those.
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 01:13:09 PM
Quote from: derby on December 21, 2010, 01:07:35 PM
well, technically, you could run the same 99:1 trial 100 times and get the same result every time.




Yes you could. But if you did run 100 trials of a 99:1 shot, how accurate are the results? That's the question.


sac
Title: Re: Math Question
Post by: Monster Dave on December 21, 2010, 01:16:39 PM
Your question:

Quote from: Sắc Dục on December 21, 2010, 12:59:11 PM

I suck at math so please pardon if I have my terminology all messed up.

Let's say you have an event that you believe is a 99:1 shot. How many independent trials would you have to run to be 95% sure that there is actually a 1% chance of this event happening?

Is there a formula where you could plug in odds of something happening (e.g. - 3:1 or 25%) the percentage of accuracy you need to confirm those odds to (presuming you only get perfect accuracy with an infinite number of trials) and get the number of trials you need to run.

Let me know if I need to clarify.

Thanks.

sac

What I read:

I suck at math so please pardon if I have my terminology all messed up.

Let's say you have an event that you believe is a 99:1 shot. How many blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah

blah blah blah blah blah blah blah

Thanks.








[laugh]

Explanation:

I saw "math" and the start of a word problem......


......my brain just shut down.  [laugh]
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 01:19:36 PM
Quote from: il d00d on December 21, 2010, 01:11:38 PM
I am not a math nerd, but remembering back to my statistic class probability always assumes an infinite number of "tries" -  if the probability of something happening is 10:1, in infinite amount of tries, this should be the actual outcome.  If you flipped a coin 10 times you might come up heads seven times, tails three, but the head to tail ratio would change the more you flipped the coin.  So, are you asking how many times, given a probability, you would have to try something to have the outcome come within 95% of the probability?  If so, I think that is a standard deviation problem, and I will be damned if I can remember how to do those.


This is exactly what I am asking. It is indeed a standard deviation problem. Which is something I know make the beast with two backs all about.

sac
Title: Re: Math Question
Post by: zooom on December 21, 2010, 01:30:19 PM
THE ANSWER IS 42




nuff said...
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 01:32:09 PM
Quote from: zooom on December 21, 2010, 01:30:19 PM
THE ANSWER IS 42




nuff said...


Incorrect.


The answer is ... 4? (http://www.youtube.com/watch?v=4ZfpwfQ58Ds#normal)
Title: Re: Math Question
Post by: akmnstr on December 21, 2010, 01:36:51 PM
As you have written your question you really cannot come up with an answer.  However if you were to change it slightly and say you would like to 99% certain of the outcome + or - 1% (or any amount you choose) then you could do a trial, use the outcome of that trial to do a power analysis (it would take to long to explain) and then you would have an estimate with an upper and lower limit.  

On the other hand, you might need to take a stats class.  
Title: Re: Math Question
Post by: Rameses on December 21, 2010, 01:42:44 PM

The 95% certainty that you're referring to is known as the confidence level.

You're talking about a data set with an infinite population, since there is no finite limit to how many times the test can be repeated.

In order to calculate your sample size, you need one more parameter.  Confidence interval.

Confidence interval is a margin of error on your 1% chance.  Because there isn't a finite data set, you can never be 95% sure that the chance of the expected result is 1%.  You can only be 95% sure that the expected result is between 0% and 2% (which would be a confidence interval of 2%).

I think I remembered all of that right.  It's been a long time.

Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 01:47:21 PM
Quote from: akmnstr on December 21, 2010, 01:36:51 PM
As you have written your question you really cannot come up with an answer.  However if you were to change it slightly and say you would like to 99% certain of the outcome + or - 1% (or any amount you choose) then you could do a trial, use the outcome of that trial to do a power analysis (it would take to long to explain) and then you would have an estimate with an upper and lower limit.  

On the other hand, you might need to take a stats class.  


But I don't want to take a stats class. I build houses for a living. I have no practical use for this knowledge. This is just the shit I think about when I'm awake at 3am. And no, I wasn't high.

Is this a valid question:

I ran 100,000 trials and 90,000 came up x and 10,000 came up y. How sure can I be that the ratio of x to y is actually 9:1?

Show you work.


sac
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 01:49:19 PM
Quote from: Rameses on December 21, 2010, 01:42:44 PM
The 95% certainty that you're referring to is known as the confidence level.

You're talking about a data set with an infinite population, since there is no finite limit to how many times the test can be repeated.

In order to calculate your sample size, you need one more parameter.  Confidence interval.

Confidence interval is a margin of error on your 1% chance.  Because there isn't a finite data set, you can never be 95% sure that the chance of the expected result is 1%.  You can only be 95% sure that the expected result is between 0% and 2% (which would be a confidence interval of 2%).

I think I remembered all of that right.  It's been a long time.



Okay. So how do you calculate how many trials it takes to get to a 95% confidence level?

sac
Title: Re: Math Question
Post by: akmnstr on December 21, 2010, 01:49:38 PM
I forgot to mention.  Don't forget about Type 1 and Type 2 error.  










just messing with ya
Title: Re: Math Question
Post by: Monster Dave on December 21, 2010, 01:52:38 PM
[bang]
Title: Re: Math Question
Post by: akmnstr on December 21, 2010, 01:53:49 PM
Quote from: Sắc Dục on December 21, 2010, 01:47:21 PM

But I don't want to take a stats class. I build houses for a living. I have no practical use for this knowledge. This is just the shit I think about when I'm awake at 3am. And no, I wasn't high.

Is this a valid question:

I ran 100,000 trials and 90,000 came up x and 10,000 came up y. How sure can I be that the ratio of x to y is actually 9:1?

Show you work.



sac

From what you did you can be certain that the ratio of your sample is 9:1.  If you took 10,000 samples of 10 each you could then average them and calculate a confidence interval around that average and that could be used to describe the population that you are taking your samples from.  
Title: Re: Math Question
Post by: Grampa on December 21, 2010, 01:54:30 PM
The answer is...... vagina.
Title: Re: Math Question
Post by: akmnstr on December 21, 2010, 01:57:32 PM
Quote from: Sắc Dục on December 21, 2010, 01:47:21 PM

But I don't want to take a stats class. I build houses for a living. I have no practical use for this knowledge. This is just the shit I think about when I'm awake at 3am. And no, I wasn't high.

Is this a valid question:

I ran 100,000 trials and 90,000 came up x and 10,000 came up y. How sure can I be that the ratio of x to y is actually 9:1?

Show you work.


sac

On the other hand, next time this happens, try to think of the sexiest woman you know naked, wank off, and go back to sleep!
Title: Re: Math Question
Post by: Speedbag on December 21, 2010, 01:59:20 PM
67% of all statistics are made up on the spot.  ;)
Title: Re: Math Question
Post by: Rameses on December 21, 2010, 02:00:22 PM
Quote from: Sắc Dục on December 21, 2010, 01:49:19 PM
Okay. So how do you calculate how many trials it takes to get to a 95% confidence level?

sac


You still haven't defined a confidence interval.

Or were you saying to go with the one that I used as an example?

Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 02:03:19 PM


You people are not helping. And after all of the well thought out advice and deep knowledge I give to the members of this board I would expect that . . . no . . . wait . . . nevermind. You guys have every right to be complete dicks.

Where the hell is that one rocket scientist dude. He knows this stuff.

sac
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 02:05:50 PM
Quote from: Rameses on December 21, 2010, 02:00:22 PM

You still haven't defined a confidence interval.

Or were you saying to go with the one that I used as an example?




Sure let's go with your example for the moment.

sac
Title: Re: Math Question
Post by: Rameses on December 21, 2010, 02:14:09 PM
Quote from: Sắc Dục on December 21, 2010, 02:05:50 PM

Sure let's go with your example for the moment.

sac



Okay, let's see here...

...ninety-fff...  ....times theee....

mmkayy

...divide by fif....   carry the.....

allllright

...cross reference with th.


Yep.  Zoom was right.  It's 42.





Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 02:21:33 PM


I hate you all.


sac
Title: Re: Math Question
Post by: akmnstr on December 21, 2010, 02:24:12 PM
Quote from: Sắc Dục on December 21, 2010, 02:03:19 PM

You people are not helping. And after all of the well thought out advice and deep knowledge I give to the members of this board I would expect that . . . no . . . wait . . . nevermind. You guys have every right to be complete dicks.

Where the hell is that one rocket scientist dude. He knows this stuff.

sac

SAC I just got my copy of  Zar Biostatistical Analysis off the shelf behind me to look up the exact formulas you need.  The trouble is that my keyboard does not have the math symbols needed to write the formulas but if you look on pages 18 and 31 the formulas are there.  Yup, that should be all ya need to figure this one out.  

It is always better to help a friend with a math problem figure it out for them-self (with a little guidance) than to just give them the answer.  
Title: Re: Math Question
Post by: akmnstr on December 21, 2010, 02:25:21 PM
Quote from: Sắc Dục on December 21, 2010, 02:21:33 PM

I hate you all.


sac

Well if that is how your going to be, I'm taking my stats book and going home. 
Title: Re: Math Question
Post by: Rameses on December 21, 2010, 02:26:32 PM
Quote from: akmnstr on December 21, 2010, 02:24:12 PM

It is always better to help a friend with a math problem figure it out for them-self (with a little guidance) than to just give them the answer.  


[laugh] [laugh] [laugh]


Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 02:28:52 PM

Restate the question:

Let's say some tells you that there is a 9:1 chance that when you start a thread on the DMF some douche bag will threadjack it. Assuming it is possible to start an infinite amount of individual threads, how many threads would you have to start before you could say with 95% certainty that there is (or is not) a 9:1 chance that someone will say something douchey?


sac
Title: Re: Math Question
Post by: El Matador on December 21, 2010, 02:40:32 PM
Quote from: Sắc Dục on December 21, 2010, 02:03:19 PM

Where the hell is that one rocket scientist dude. He knows this stuff.

sac

*Raises hand*

I'm probably not the one you're thinking of but I'm Aerospace Engineering. It would take more energy to explain how your question is intrinsically wrong than it does for me to point and laugh. So there ;D

Title: Re: Math Question
Post by: ducpainter on December 21, 2010, 02:46:06 PM
One thread
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 02:48:25 PM
Quote from: El Matador on December 21, 2010, 02:40:32 PM
*Raises hand*

I'm probably not the one you're thinking of but I'm Aerospace Engineering. It would take more energy to explain how your question is intrinsically wrong than it does for me to point and laugh. So there ;D




Screw you too!   ;D

I found what I was looking for:
http://talkstats.com/showthread.php?t=12567 (http://talkstats.com/showthread.php?t=12567)

sac
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 02:59:32 PM


This was also helpful:

http://www.intmath.com/Counting-probability/12_Binomial-probability-distributions.php (http://www.intmath.com/Counting-probability/12_Binomial-probability-distributions.php)

When I can't sleep tonight I will set myself to really making sense of that.

sac



/or I'll just punch the clown and go back to sleep
Title: Re: Math Question
Post by: causeofkaos on December 21, 2010, 03:05:23 PM
punch the clown 42 times then goto sleep.
Title: Re: Math Question
Post by: Rameses on December 21, 2010, 03:26:36 PM
Quote from: Sắc Dục on December 21, 2010, 02:28:52 PM
Restate the question:

Let's say some tells you that there is a 9:1 chance that when you start a thread on the DMF some douche bag will threadjack it. Assuming it is possible to start an infinite amount of individual threads, how many threads would you have to start before you could say with 95% certainty that there is (or is not) a 9:1 chance that someone will say something douchey?


sac


Can we lay odds on who it'll be that says something douchey??




(And when did it change from 99% to 90%?)

Title: Re: Math Question
Post by: GAAN on December 21, 2010, 03:43:29 PM
Quote from: Sắc Dục on December 21, 2010, 01:47:21 PM

This is just the shit I think about when I'm awake at 3am. And no, I wasn't high.

sac

well

then that's your answer

think about it whilst high and then there is your answer
Title: Re: Math Question
Post by: Speeddog on December 21, 2010, 04:01:22 PM
Quote from: Sắc Dục on December 21, 2010, 01:47:21 PM
~~~SNIP~~~
Is this a valid question:

I ran 100,000 trials and 90,000 came up x and 10,000 came up y. How sure can I be that the ratio of x to y is actually 9:1?

Show you work.

sac

How big is the population that you're sampling from?

Because if it's 100,000, well, you know for sure the ratio is 9:1.
If it's 100,000,000,000,000,000,000,000,000,000 then your sample is not so good.
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 04:18:47 PM
Quote from: Speeddog on December 21, 2010, 04:01:22 PM
How big is the population that you're sampling from?

Because if it's 100,000, well, you know for sure the ratio is 9:1.
If it's 100,000,000,000,000,000,000,000,000,000 then your sample is not so good.

The number of trials that can be run is infinite. Like flipping a coin. Each flip is an independent trial and there can be infinite number of them. So how many times do I have to flip the coin before I can be 95% sure that the heads to tails ratio is 1:1? Being 100% sure through the use of repeated trials is of course impossible.

sac
Title: Re: Math Question
Post by: Triple J on December 21, 2010, 04:33:21 PM
Quote from: Sắc Dục on December 21, 2010, 04:18:47 PM
The number of trials that can be run is infinite. Like flipping a coin. Each flip is an independent trial and there can be infinite number of them. So how many times do I have to flip the coin before I can be 95% sure that the heads to tails ratio is 1:1? Being 100% sure through the use of repeated trials is of course impossible.

sac

Easy.

It takes the same number of trials as it does licks to get to the center of a Tootsie Roll Pop.
Title: Re: Math Question
Post by: mitt on December 21, 2010, 04:33:58 PM
How did I miss this thread, I do this all the time for predicting pass / fail - I don't have all my notes with me, but let me think about it.

the general eqn is

P = 1-p^n         

p = individual probability of event happening (in your case 1%)
n = number of tests (what you want to know)
P = chance of finding at least one event (in your case you want to find it at least 95% of the time)

Title: Re: Math Question
Post by: derby on December 21, 2010, 04:40:55 PM
Quote from: Sắc Dục on December 21, 2010, 04:18:47 PM
The number of trials that can be run is infinite. Like flipping a coin. Each flip is an independent trial and there can be infinite number of them. So how many times do I have to flip the coin before I can be 95% sure that the heads to tails ratio is 1:1? Being 100% sure through the use of repeated trials is of course impossible.

sac

http://en.wikipedia.org/wiki/Law_of_large_numbers (http://en.wikipedia.org/wiki/Law_of_large_numbers)
Title: Re: Math Question
Post by: mitt on December 21, 2010, 04:49:24 PM
So, I am pretty sure your answer is 299 times.

n = ln(.95)/ln(1-.01) = 289.07


mitt
Title: Re: Math Question
Post by: zooom on December 21, 2010, 05:12:56 PM
Quote from: bobspapa on December 21, 2010, 01:54:30 PM
The answer is...... vagina.

or boobies!
Title: Re: Math Question
Post by: Speedbag on December 21, 2010, 05:52:17 PM
Yesssssss.

I hear Ducati plans to make a hypermotard.
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 05:58:25 PM
Quote from: mitt on December 21, 2010, 04:33:58 PM
How did I miss this thread, I do this all the time for predicting pass / fail - I don't have all my notes with me, but let me think about it.

the general eqn is

P = 1-p^n         

p = individual probability of event happening (in your case 1%)
n = number of tests (what you want to know)
P = chance of finding at least one event (in your case you want to find it at least 95% of the time)




Cool! Thanks mitt! Now I might actually sleep tonight.

sac


/the rest of you can go to hell!   ;D
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 05:59:15 PM
Quote from: Speedbag on December 21, 2010, 05:52:17 PM
Yesssssss.

I hear Ducati plans to make a hypermotard.


Did you hear that Bacon Junkie got a 999?

sac



/because he did
Title: Re: Math Question
Post by: lethe on December 21, 2010, 06:00:59 PM
Quote from: Sắc Dục on December 21, 2010, 05:59:15 PM

Did you hear that Bacon Junkie got a 999?

sac



/because he did
the math just don't add up on that one
???
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 06:07:13 PM
Quote from: lethe on December 21, 2010, 06:00:59 PM
the math just don't add up on that one
???


Sure it does. Look, take 2000 miles (driven) divided by 3 days (total trip time) minus 142 mph (indicated) and then multiply the whole thing by the square root of retarded. And the answer is 999.

Easy.


sac
Title: Re: Math Question
Post by: GAAN on December 21, 2010, 06:13:43 PM
Quote from: Sắc Dục on December 21, 2010, 05:58:25 PM

Cool! Thanks mitt! Now I might actually sleep tonight.

sac


/the rest of you can go to hell!   ;D

C18H21NO4 + C16H13ClN2O + C20H21N = Sn0rE

Title: Re: Math Question
Post by: lethe on December 21, 2010, 06:16:23 PM
Quote from: Sắc Dục on December 21, 2010, 06:07:13 PM

Sure it does. Look, take 2000 miles (driven) divided by 3 days (total trip time) minus 142 mph (indicated) and then multiply the whole thing by the square root of retarded. And the answer is 999.

Easy.


sac
retarded equals 47.98070339340654

which is roughly 6 more than.....42
which was the sixth post after your first in this thread....what does it mean?
Title: Re: Math Question
Post by: mitt on December 21, 2010, 06:43:17 PM
I am not a statistician, but I play one at work, so the answer is not guaranteed.    ;D

If you access to excel, there are some good functions and help in there.  Your link was good too, and goes above what I usually do.  

An easy example of what are the odds (confidence) that you get at least one 5 rolling the dice.  

1 side probabilty is 1/6 = 16.7%
any other side is 5/6 = 83.3%

rolls   odds of at least one 5 side
1        1 - .833^1 = .167 = %16.7 (makes sense since it is the same as original 1 side above)
2        1 - .833^2 = .306 = %30.6
3        1 - .833^3 = .422 = %42.2
4        1 - .833^4 = .519 = %51.9
etc
etc

If you plot it out to say 20 rolls, you can see it is converging to 100%, but it will never get there.

(https://7782181789889801422-a-1802744773732722657-s-sites.googlegroups.com/site/mittelstadtc/Home/file2/1sidediceodds2.jpg?attachauth=ANoY7cp1Woqgw3zRQSfSKpfLqqiYvveNRIEBdzfMEFuHrceWTLVEvn1lSuguxzLvc7f7z8DKHIrrLNhLlD59p6Bm3vbAzk4nMDTri_anx-2B2bd6VCr-SFZ_lLtpWA91n8SNaKxMP97VBHRZ7w3DRW0LZMng1RuWLrki7WebQRghxEsIsKSvzuS9xF5IQMXUezSaZVaas_bBlEzMno14x43tUqkf90AQMCMej8jtcIhOKL4-dLE-DEI%3D&attredirects=0)
Title: Re: Math Question
Post by: SacDuc on December 21, 2010, 08:11:21 PM
Quote from: mitt on December 21, 2010, 04:49:24 PM
So, I am pretty sure your answer is 299 times.

n = ln(.95)/ln(1-.01) = 289.07


mitt


Two things, 1) I know nothing about natural logarithms  2)  289 trials to be 95% sure that something is a 99:1 shot seems really really low.

Please show your work.

sac
Title: Re: Math Question
Post by: Rameses on December 21, 2010, 08:46:15 PM
Quote from: Sắc Dục on December 21, 2010, 08:11:21 PM

Two things, 1) I know nothing about natural logarithms  2)  289 trials to be 95% sure that something is a 99:1 shot seems really really low.

Please show your work.

sac


You would be correct.

Title: Re: Math Question
Post by: Adamm0621 on December 22, 2010, 12:02:50 AM
Why does this triangle look like that triangle?







BECAUSE IT make the beast with two backsING DOES!!!! >:(
Title: Re: Math Question
Post by: RAT900 on December 22, 2010, 06:21:00 AM
I did this shit for a living way back when....probability statistics for sizing networks

with a make the beast with two backsing calculator...figuring blocking probabilities based on offered peak volume attempts and carried volume and blocking percentages

make the beast with two backs you for reminding me of this torment, next time rub one out...or 2 or 3 if necessary and go to sleep ...here chew on this

http://en.wikipedia.org/wiki/Poisson_distribution (http://en.wikipedia.org/wiki/Poisson_distribution)

or  http://en.wikipedia.org/wiki/Erlang_distribution (http://en.wikipedia.org/wiki/Erlang_distribution)
Title: Re: Math Question
Post by: ungeheuer on December 22, 2010, 06:39:04 AM
if they all build arks, wont they be over capacity by 50%??
Title: Re: Math Question
Post by: mitt on December 22, 2010, 07:15:48 AM
Quote from: Sắc Dục on December 21, 2010, 08:11:21 PM

Two things, 1) I know nothing about natural logarithms  2)  289 trials to be 95% sure that something is a 99:1 shot seems really really low.

Please show your work.

sac

The wording is like this: In 298 attempts you will find at least one 1% event 95% of the time.  


I might have had a typo above, I need to double check, but here is how to get 298:

n = ln(1-.95)/ln(1-.01) = -3.00/-0.0101 = 298


lets try it with dice.  95% of the time to find a 1/6 (18%) event (getting a 5 side)

n = ln(1-.95)/ln(1-.18) = -3.00/-0.182 = 16.4 = 17 times - that matches the graph I posted. 

In words:  You need to roll a dice 17 times to guarantee at 95% of the time you will find at least 1 five side. 



mitt
Title: Re: Math Question
Post by: SacDuc on December 22, 2010, 08:42:25 AM

mitt,

I've probably stated the question poorly. Let me try again.

Let's say someone tells you that an independent has is a 99:1 shot.  You know that is in the ballpark, but you want to verify that claim and be 95% sure that your finding it is correct. How many trials do you need to run to be 95% certain that this event does in fact happen 1% of the time?


sac
Title: Re: Math Question
Post by: Drunken Monkey on December 22, 2010, 09:29:17 AM
As someone else has already said, this is what binomial distributions are for.

There are some nice prepackaged formulas built into excel for this.

I used these in a whitepaper once to show that a VP was a complete twat.

I wish the prof in my statistic class had told me I could do that with these formulas as I might have paid more attention at the time.
Title: Re: Math Question
Post by: derby on December 22, 2010, 09:52:45 AM
Quote from: Drunken Monkey on December 22, 2010, 09:29:17 AM

I used these in a whitepaper once to show that a VP was a complete twat.


i'd love to read that. haha.
Title: Re: Math Question
Post by: SacDuc on December 22, 2010, 10:09:38 AM
Quote from: Drunken Monkey on December 22, 2010, 09:29:17 AM
As someone else has already said, this is what binomial distributions are for.

There are some nice prepackaged formulas built into excel for this.

I used these in a whitepaper once to show that a VP was a complete twat.

I wish the prof in my statistic class had told me I could do that with these formulas as I might have paid more attention at the time.

Okay. I don't know anything about binomial distributions. How does one apply them to this question:

Quote from: Sắc Dục on December 22, 2010, 08:42:25 AM

Let's say someone tells you that an independent has is a 99:1 shot.  You know that is in the ballpark, but you want to verify that claim and be 95% sure that your finding it is correct. How many trials do you need to run to be 95% certain that this event does in fact happen 1% of the time?



Please show your work.

Thanks.


sac
Title: Re: Math Question
Post by: RAT900 on December 22, 2010, 12:04:34 PM
Sac...this behavior is all very reminiscent of how John Forbes Nash began his dreadful  slide into the abyss
Title: Re: Math Question
Post by: sno_duc on December 22, 2010, 01:23:13 PM
Quote from: Drunken Monkey on December 22, 2010, 09:29:17 AM
I used these in a whitepaper once to show that a VP was a complete twat.

Are they flexible enough to be usable on any co-worker???  [evil]

if so, don't tease us.  [bang]

Please show your work.  ;D 
Title: Re: Math Question
Post by: Ddan on December 22, 2010, 01:43:24 PM
Anal
Title: Re: Math Question
Post by: GAAN on December 22, 2010, 02:08:16 PM
Quote from: Dan on December 22, 2010, 01:43:24 PM
Anal

Quote from: sno_duc on December 22, 2010, 01:23:13 PM
Are they flexible enough to be usable on any co-worker???  [evil]

if so, don't tease us.  [bang]

Please show your work.  ;D 

:-\
Title: Re: Math Question
Post by: Ddan on December 22, 2010, 02:09:19 PM
Quote from: Mother on December 22, 2010, 02:08:16 PM
:-\
Someone had to say it 
Title: Re: Math Question
Post by: mitt on December 22, 2010, 02:46:12 PM
Short and final answer I think.  You can never be sure that the odds are a finite number.  Like was said on the first page, you need a range, or reword the odds like at least 1:99 or less than 1:99, but to come up with a single finite number, is like saying something is exactly 1 inch long, but it really always has a tolerance.

(unless it is a dice or something with a finite number of sides and outcomes)


here are some excel functions to use if you want to play with it:

normdist
binomdist

Excel Magic Trick #22: NORMDIST function for Probability (http://www.youtube.com/watch?v=qgHekFh6hX8#normal)



mitt
Title: Re: Math Question
Post by: GAAN on December 22, 2010, 03:07:34 PM
wow

ive never wanted to punch a guy in the throat nearly as bad as I do Mr Excellfun
Title: Re: Math Question
Post by: SacDuc on December 22, 2010, 03:23:38 PM
It is finite. Either A happens or A doesn't happen. Lets call that one B. So the claim is that A happens 99% of the time and B happens 1% of the time. How do I figure out how many trials I have to run in order to verify this to a degree of 95% accuracy. 0 trials gives you 0% accuracy on your verification. Infinite trials gives you 100% accuracy. How do I determine how accurately I have verified the 99:1 claim with a given number of trials somewhere between 0 and infinity.

sac
Title: Re: Math Question
Post by: derby on December 22, 2010, 03:39:12 PM
Quote from: Sắc Dục on December 22, 2010, 03:23:38 PM
Infinite trials gives you 100% accuracy.

i thought the more trials you ran the closer it would get to 100% but never actually be 100%.
Title: Re: Math Question
Post by: sno_duc on December 22, 2010, 07:47:28 PM
A little off topic, but along the same lines.
           ____
Does 1.99999 = 2


   ____
1.9999 x 10 = 2 x 10      (multple both sides by 10)
     _____
19.99999  = 20
    _____      ______
19.99999 - 1.999999 = 20 - 2  (subtract orginal number from both sides)

18 = 18

18/9 = 18/9        (divide by nine)

2 = 2


Try it with any number....multiple by 10, subtract orginal number, divide by nine.
Can't remember the professors name, he was one of the people who wrote modern math, fun guy always little tidbits to keep the lectures entertaining.
Title: Re: Math Question
Post by: Randimus Maximus on December 22, 2010, 07:57:54 PM
Quote from: Mother on December 22, 2010, 03:07:34 PM
wow

ive never wanted to punch a guy in the throat nearly as bad as I do Mr Excellfun

+ 11ty billion
Title: Re: Math Question
Post by: Rameses on December 22, 2010, 07:59:09 PM
Quote from: derby on December 22, 2010, 03:39:12 PM
i thought the more trials you ran the closer it would get to 100% but never actually be 100%.


That's true.

But it's also true that no matter how many trials you do, you haven't done an infinite number.

Title: Re: Math Question
Post by: Randimus Maximus on December 22, 2010, 08:00:27 PM
Quote from: Rameses on December 22, 2010, 07:59:09 PM

That's true.

But it's also true that no matter how many trials you do, you haven't done an infinite number.



Are you still in school?
Title: Re: Math Question
Post by: Rameses on December 22, 2010, 08:31:40 PM
Quote from: Randimus Maximus on December 22, 2010, 08:00:27 PM
Are you still in school?


[laugh] [laugh] [laugh]

I think you already know the answer to that, but yes, I am.

I'm no longer a stat major though.  I always did well in stat classes without really putting much effort into them, but one day realized I didn't really enjoy it and didn't have any interest in doing it for the rest of my life.




And sac, 9,604.

That's your answer.

I'm too lazy to look up the equations you need, but you can use this link (http://www.surveysystem.com/sscalc.htm).

Even though the population is infinite, the percentage of the population represented by a given sample size changes less and less as your population increases.  Therefor, you can chose some arbitrary extremely large population size and use that to represent infinity fairly closely.  If you notice, there's no difference in the sample size needed for a 95% confidence level and a confidence interval of 1 with population sizes of 200 million and 900 trillion.

Title: Re: Math Question
Post by: akmnstr on December 23, 2010, 05:58:42 AM
I can't believe that this thread has now gone on 5 pages when the answer was given on page 1.  It has been stated and re-stated.  Someone seems to be a little slow on the pickup here.  No disrespect intended, just a morning observation.  I'm sure there is something other than statistics that Sac is good at. [coffee]
Title: Re: Math Question
Post by: The Bacon Junkie on December 23, 2010, 06:41:54 AM
Quote from: akmnstr on December 23, 2010, 05:58:42 AM

---snip---
 
I'm sure there is something other than statistics that Sac is good at. [coffee]

He's pretty good at spatchcocking a chicken...  ;)

I've seen it in person, the man's got mad skills!



P.S.  It's not what you think...
 


[bacon]
Title: Re: Math Question
Post by: Bick on December 23, 2010, 07:01:22 AM
Quote from: The Bacon Junkie on December 23, 2010, 06:41:54 AM
He's pretty good at spatchcocking a chicken...  ;)

I've seen it in person, the man's got mad skills!



P.S.  It's not what you think...
 


[bacon]

???

You mean it is not a cooking term?

:-X
Title: Re: Math Question
Post by: RAT900 on December 23, 2010, 07:21:26 AM
Quote from: The Bacon Junkie on December 23, 2010, 06:41:54 AM
He's pretty good at spatchcocking a chicken...  ;)

I've seen it in person, the man's got mad skills!



P.S.  It's not what you think...
 


[bacon]



I think I saw a spatchcocking in Tijuana once...many years ago...I was very drunk on Mexican Worm Whisky
Title: Re: Math Question
Post by: The Bacon Junkie on December 23, 2010, 07:56:03 AM
Quote from: Bick on December 23, 2010, 07:01:22 AM
???

You mean it is not a cooking term?

:-X

Hush, you!   ;)

I was going to make a reference to eating the chicken afterwards, but now it doesn't sound as dirty...   :-\


[laugh]



[bacon]
Title: Re: Math Question
Post by: akmnstr on December 23, 2010, 08:07:54 AM
Spatchcocking  defined  http://www.nakedwhiz.com/spatchdef.htm (http://www.nakedwhiz.com/spatchdef.htm)
Title: Re: Math Question
Post by: SacDuc on December 23, 2010, 09:04:30 AM
Quote from: akmnstr on December 23, 2010, 05:58:42 AM
I can't believe that this thread has now gone on 5 pages when the answer was given on page 1.  It has been stated and re-stated.  Someone seems to be a little slow on the pickup here.  No disrespect intended, just a morning observation.  I'm sure there is something other than statistics that Sac is good at. [coffee]

Um no. No it is not.

But since I'm such a retard and you are so smart I'm certain your next post will contain an equation that answers this question:

Quote from: Sắc Dục on December 22, 2010, 03:23:38 PM
It is finite. Either A happens or A doesn't happen. Lets call that one B. So the claim is that A happens 99% of the time and B happens 1% of the time. How do I figure out how many trials I have to run in order to verify this to a degree of 95% accuracy. 0 trials gives you 0% accuracy on your verification. Infinite trials gives you 100% accuracy. How do I determine how accurately I have verified the 99:1 claim with a given number of trials somewhere between 0 and infinity.

sac

Some how I think it will not actually contain a solvable equation but rather just rambling about how certain mathematical concepts need to be applied.

Thanks for the derision though that always brightens my day. Especially when its sprinkled with arrogance.

sac
Title: Re: Math Question
Post by: The Bacon Junkie on December 23, 2010, 10:08:17 AM
Spatchcock























In my infinite lack of maturity, I just wanted to say it again!   ;D




[bacon]
Title: Re: Math Question
Post by: akmnstr on December 23, 2010, 10:20:47 AM
QuoteSome how I think it will not actually contain a solvable equation but rather just rambling about how certain mathematical concepts need to be applied.

Thanks for the derision though that always brightens my day. Especially when its sprinkled with arrogance.

sac

Oh Sac,  don't be so sensitive.  This is all in fun. 
Title: Re: Math Question
Post by: ducpainter on December 23, 2010, 10:47:30 AM
Quote from: akmnstr on December 23, 2010, 10:20:47 AM
Oh Sac,  don't be so sensitive.  This is all in fun. 
It's only good when he's serving. ;)
Title: Re: Math Question
Post by: SacDuc on December 23, 2010, 11:02:19 AM
Quote from: humorless dp on December 23, 2010, 10:47:30 AM
It's only good when he's serving. ;)



And its only thread jacking when other people do it. Go be a hypocrite in some other thread.


sac
Title: Re: Math Question
Post by: ducpainter on December 23, 2010, 11:13:20 AM
Quote from: Sắc Dục on December 23, 2010, 11:02:19 AM


And its only thread jacking when other people do it. Go be a hypocrite in some other thread.


sac
Like I said...
Title: Re: Math Question
Post by: il d00d on December 23, 2010, 11:30:40 AM
I don't want to live in a world where a thread about *math* on a motorcycle forum gets locked.  Let's all get in the secular Winter holiday celebration spirit.  I'm saying let's all and drink and wallow in existential anguish.  About other things.
Title: Re: Math Question
Post by: Speedbag on December 23, 2010, 12:46:33 PM
Spatchcock?!?

Sounds like a great innerweb forum name.....