User Controls

Math question

  1. #1
    Because I am curious. Say someone has a certain quantity of candy, 1 lb. The candy pieces all have different amounts of sugar, and the distribution of their sweetness follows a Gaussian distribution. What would be the minimum amount of candy needed, so that if you were to take that amount of candy and blend it together and made new candies, to get new candies that are about as sweet as the total average of sweetness in the whole batch of candy?
  2. #2
    *bup*
  3. #3
    @sploo @lanny
  4. #4
    mmQ Lisa Turtle
    I'm just gonna say probably about 62°-68% and see if I'm close. Thanms
    The following users say it would be alright if the author of this post didn't die in a fire!
  5. #5
    mashlehash victim of incest [my perspicuously dependant flavourlessness]
    I would still end up Diabetic.
  6. #6
    Originally posted by mmQ I'm just gonna say probably about 62°-68% and see if I'm close. Thanms

    lol did you intentionally make one of those degrees and one percent?
  7. #7
    mmQ Lisa Turtle
    Originally posted by greenplastic lol did you intentionally make one of those degrees and one percent?

    nyes
  8. #8
    fanglekai Yung Blood
    5
  9. #9
    Originally posted by greenplastic Because I am curious. Say someone has a certain quantity of candy, 1 lb. The candy pieces all have different amounts of sugar, and the distribution of their sweetness follows a Gaussian distribution. What would be the minimum amount of candy needed, so that if you were to take that amount of candy and blend it together and made new candies, to get new candies that are about as sweet as the total average of sweetness in the whole batch of candy?

    this doesnt have an answer because "about as sweet" is arbitrary

    when i was doing correlation work the data started to average out around n=10

    or you could just crush all of the candy together then stir it, then every portion will be about equal sweetness
  10. #10
    Originally posted by roglahonz blaj this doesnt have an answer because "about as sweet" is arbitrary

    when i was doing correlation work the data started to average out around n=10

    or you could just crush all of the candy together then stir it, then every portion will be about equal sweetness

    I just didn't specify exactly how close it had to be because I don't have a real reason to, but that doesn't mean nothing can be said about how many you would need to get an accurate result.

    And I mean yea but I didn't even mention any number of candies there were, I just meant by weight.
  11. #11
    even then its all up to probability, meaning the results cannot be accurately predicted

    https://en.wikipedia.org/wiki/Law_of_large_numbers

    basically you're asking which number is properly "large", and the answer is theres no way to determine this, but trends tend to even out over time, and there's probably a logarithmic function to determine accuracy based on a bell curve, but i cant say for sure
  12. #12
    mashlehash victim of incest [my perspicuously dependant flavourlessness]
    Originally posted by roglahonz blaj even then its all up to probability, meaning the results cannot be accurately predicted

    https://en.wikipedia.org/wiki/Law_of_large_numbers

    basically you're asking which number is properly "large", and the answer is theres no way to determine this, but trends tend to even out over time, and there's probably a logarithmic function to determine accuracy based on a bell curve, but i cant say for sure

    sources, kick the stool.
  13. #13
    Originally posted by mashlehash sources, kick the stool.

    Yo want a source for some simple logic?
  14. #14
    Lanny Bird of Courage
    An actual number depends on the parameters of the original distribution and what "about as sweet" means. There's always a possibility that your sample (candy picked to be combined) mean deviates from population mean but it goes down as sample size increases. If you take "about as sweet" to mean "within one standard deviation of the original mean" you would expect about 1-.32^n of samples of would be "not about as sweet" as the population mean.
  15. #15
    Originally posted by Lanny An actual number depends on the parameters of the original distribution and what "about as sweet" means. There's always a possibility that your sample (candy picked to be combined) mean deviates from population mean but it goes down as sample size increases. If you take "about as sweet" to mean "within one standard deviation of the original mean" you would expect about 1-.32^n of samples of would be "not about as sweet" as the population mean.

    shut the fuck up retard
  16. #16
    Daily an(nu)ally [dissolutely whisk the pantheon]
    Originally posted by Lanny An actual number depends on the parameters of the original distribution and what "about as sweet" means. There's always a possibility that your sample (candy picked to be combined) mean deviates from population mean but it goes down as sample size increases. If you take "about as sweet" to mean "within one standard deviation of the original mean" you would expect about 1-.32^n of samples of would be "not about as sweet" as the population mean.

    2 plus 2 is 4 minus 1 that's 3 quick mafs
  17. #17
    Lanny Bird of Courage
    Originally posted by roglahonz blaj shut the fuck up retard

    You don't get to call people retards when you wrote this unironically:

    Originally posted by roglahonz blaj even then its all up to probability, meaning the results cannot be accurately predicted

    https://en.wikipedia.org/wiki/Law_of_large_numbers

    basically you're asking which number is properly "large", and the answer is theres no way to determine this, but trends tend to even out over time, and there's probably a logarithmic function to determine accuracy based on a bell curve, but i cant say for sure
    The following users say it would be alright if the author of this post didn't die in a fire!
  18. #18
    Originally posted by Lanny If you take "about as sweet" to mean "within one standard deviation of the original mean" you would expect about 1-.32^n of samples of would be "not about as sweet" as the population mean.

    thanks for the useless and probably incorrect info
  19. #19
    Lanny Bird of Courage
    "even then its all up to probability, meaning the results cannot be accurately predicted"

    --sploo, 2018
  20. #20
    >probability
    >accurate
Jump to Top