|
|
SegaRob 02-21-2008, 10:02 PM If the test returns the wrong result at random over the entire population, then those rates are the same.
But yeah, attack well known statisticians now.
This is the same type of backpedaling we saw in the PoaT threads. "I didn't understand the clear question!"Test results are hardly ever random (else we'd get 50% accuracy.. durr). Most of these tests are actually on a scale where you can "trade" false positives for false negatives simply by changing the cutoff. In other words, in general you can arbitrarily choose the false positive rate.
It's a false assumption he used to teach the AVERAGE public. I'm sure he would not teach it that way in a real statistics class. And if he did, there would hopefully be a smartass in the audience to punk him like I just did to you.
SegaRob 02-21-2008, 10:04 PM TI seems to think that if he posts enough, he'll be right.
This thread is fully of poorly-defined trick problems.
OH this is Tribal Imperialist's smurf?
-raven- 02-21-2008, 10:06 PM Test results are hardly ever random (else we'd get 50% accuracy.. durr). Most of these tests are actually on a scale where you can "trade" false positives for false negatives simply by changing the cutoff. In other words, in general you can arbitrarily choose the false positive rate.
It's a false assumption he used to teach the AVERAGE public. I'm sure he would not teach it that way in a real statistics class. And if he did, there would hopefully be a smartass in the audience to punk him like I just did to you.
Nobody said anything about the test results being random. Are we at the stage in the argument where we make things up now? Did I say the test is a coin flip?
The only person making assumptions in this thread is you and Darkstrand. You can't seem to understand a simple problem... so how can you have any hope in solving it?
opsayo 02-21-2008, 10:06 PM edit: oops MORMONS LOL
Darkstrand 02-21-2008, 10:09 PM So, now ur position is that you don't know how to state all the information for a statistics question?
You can't just change the question afterwards when you make yourself look like an idiot.
-raven- 02-21-2008, 10:11 PM So, now ur position is that you don't know how to state all the information for a statistics question?
You can't just change the question afterwards when you make yourself look like an idiot.
You remind me a lot of Dumpy Dooby. Keep attacking when shown wrong.
I didn't change the question at all. Go back and read it. Go back and read my solution. Go back and watch the video. For once in your life, get a clue.
PS: "ur" :lol:
Teratos 02-21-2008, 10:11 PM TI changing the question = everyone else backpedaling about it being poorly-defined. duh
Darkstrand joins the ranks of the retarded.
By the way, I minored in math.
PS This problem and the coin problem are both discussed in:
TED | Talks | Peter Donnelly: How juries are fooled by statistics (video) (http://www.ted.com/index.php/talks/view/id/67)
Go to 11:15.
oh wow you MINORED in math :rofl:
-raven- 02-21-2008, 10:26 PM Major was computer science, in which I hold a BS and an MS.
They aren't exactly completely distinct fields.
None of that has any relevance to the actual problem. Just a response to a personal attack. From what I gather, Darkstrand has no degree period.
SegaRob 02-21-2008, 10:28 PM Nobody said anything about the test results being random. Are we at the stage in the argument where we make things up now? Did I say the test is a coin flip?
The only person making assumptions in this thread is you and Darkstrand. You can't seem to understand a simple problem... so how can you have any hope in solving it?
You assumed FPR=FNR, which we had no reason to believe that was true. Can you admit that, yes or no?
There are about 300,000,000 Americans and 5.5 million of them are mormons.
That means there are 5,500,000 mormons and 294,500,000 non-mormons.
Since our test is 95% accurate, it identifiers 5,225,000 of the mormons as mormons. Since our test is 5% inaccurate, it identifies 14,725,000 non-mormons as mormons as well.
So out of all Americans that the test marks as mormon, only 26.2% are actually mormon.
hey man thanks for posing a question and leaving out critical information
like im going to fucking look up the census numbers to solve it
although to be fair you couldn't really answer the original question without knowing the probability that someone is actually mormon, but usually when you ask a brain teaser you should provide all the information
hey guys whats the probability that if i add two numbers it's 10 or more!??!?!
p.s. oh i forgot to mention 4 pages back that the two numbersa each have to be between one and ten
opsayo 02-21-2008, 10:29 PM ICS is one of the easiest majors\
Major was computer science, in which I hold a BS and an MS.
They aren't exactly completely distinct fields.
None of that has any relevance to the actual problem. Just a response to a personal attack. From what I gather, Darkstrand has no degree period.
cool, i was a biomedical engineer and was pretty much 1 course short of a math major (some fucking abstract high level calc class that was all proofs) due to my engineering requirements. doesn't mean i remember jack shit about statistics and am 100% credible in whatever the fuck i say
-raven- 02-21-2008, 10:31 PM hey man thanks for posing a question and leaving out critical information
like im going to fucking look up the census numbers to solve it
although to be fair you couldn't really answer the original question without knowing the probability that someone is actually mormon, but usually when you ask a brain teaser you should provide all the information
hey guys whats the probability that if i add two numbers it's 10 or more!??!?!
p.s. oh i forgot to mention 4 pages back that the two numbersa each have to be between one and ten
The point is that you have to realize you are missing information, and you need to figure out what that information is. If I told you that about 2% of the population was mormon, it would have been a pretty big hint.
SegaRob 02-21-2008, 10:32 PM hey man thanks for posing a question and leaving out critical information
like im going to fucking look up the census numbers to solve it
although to be fair you couldn't really answer the original question without knowing the probability that someone is actually mormon, but usually when you ask a brain teaser you should provide all the information
hey guys whats the probability that if i add two numbers it's 10 or more!??!?!
p.s. oh i forgot to mention 4 pages back that the two numbersa each have to be between one and ten
Well when someone does that you have three options:
1) Make up a wrong answer.
2) State what information you would need to solve the problem.
3) Look up the information from 2) and actually solve the problem.
I took option 2. Just because he left out critical information doesn't justify people taking option 1.
SegaRob 02-21-2008, 10:33 PM cool, i was a biomedical engineer and was pretty much 1 course short of a math major (some fucking abstract high level calc class that was all proofs) due to my engineering requirements. doesn't mean i remember jack shit about statistics and am 100% credible in whatever the fuck i say
haha, props to being honest
-raven- 02-21-2008, 10:35 PM cool, i was a biomedical engineer and was pretty much 1 course short of a math major (some fucking abstract high level calc class that was all proofs) due to my engineering requirements. doesn't mean i remember jack shit about statistics and am 100% credible in whatever the fuck i say
Of course it doesn't. But if Darkstrand wants to claim that it does, then why not take that point of debate out of the conversation?
I'm not sure exactly how close to a math major I was. I looked at switching at the end of my third year when I was enjoying math courses more than CS courses and it wasn't going to change my graduation date. My school doesn't do double majors.
opsayo 02-21-2008, 10:36 PM ayz,pretty much all higher level math classes are all proofs
was minoring at your college pretty easy?
Soldevi 02-21-2008, 10:41 PM dont argue you with rob about smart shit, you will lose
|
|