3 ms·

Accommodating everyone except those taller than four standard deviations above mean would (assuming gaussian distribution) mean accommodating 99.997% of your pa

Accommodating everyone except those taller than four standard deviations above mean would (assuming gaussian distribution) mean accommodating 99.997% of your passengers.

nightski 9y ago
This completely misses the point. First of all, you are only correct if assuming a gaussian distribution of heights (for which I have seen no evidence). Who cares if it accommodates 3SD, 4SD or 5SD? You would never engineer a system around standard deviations of a distribution of heights. Instead it's about marginal cost. Ideally you would accommodate 100% of your passengers unless their marginal cost is just too high. You are also assuming the distribution of passengers is the same as the distribution of the population. Or the distribution of customers willing to pay certain prices is the same as the general population. Airline seating is very complex as it's not a fixed price, changes constantly, and even varies from seat to seat in a single flight. Prices aren't fixed so just because you increase marginal cost does not mean reduce profit. Lastly it also ignores completely the fact that even a shorter person would enjoy the increased room.
- totalZero 9y ago
  You responded to a comment about how a case is extremely uncommon, with a suggestion about half the population. All I pointed out is that your argument is less than perfectly applicable to the comment to which you're responding, by stating a mathematical fact. I didn't assume anything, ignore anything, assert anything, or even draw a conclusion about what should be done. Just spreading facts. Call me Johnny Factseed.
  - nightski 9y ago
    Again you assumed a Gaussian distribution. There is nothing "factual" about that. Not to mention I was pointing out that your facts were irrelevant because the miss they point (Using population statistics to guide the design/engineering of an experience is doomed)
    - totalZero 9y ago
      I'm not designing the experience. Nor does telling you what a particular distribution would indicate if assumed, imply that I have assumed the distribution. Mathematically we would expect height to be distributed in a somewhat gaussian fashion due to the central limit theorem. Therein lies its relevance. However, nobody is 70 feet tall (yet!), so it would be fair to argue that the tails are characterized differently by a gaussian distribution than by the "true" distribution for height (which would probably divert some of that probability mass from the far tails to the near tails). Still, it is quite misplaced to talk about 50% of the passengers when the discussion is about substantial outliers. Just giving you some information that will help you use more applicable axia in the future. Signed, Johnny Factseed
      - nightski 9y ago
        That's not how the CLT works. It does not mean every random variable is Gaussian.

Accommodating everyone except those taller than four standard deviations above mean would (assuming gaussian distribution) mean accommodating 99.997% of your passengers.

nightski 9y ago

This completely misses the point. First of all, you are only correct if assuming a gaussian distribution of heights (for which I have seen no evidence). Who cares if it accommodates 3SD, 4SD or 5SD? You would never engineer a system around standard deviations of a distribution of heights. Instead it's about marginal cost. Ideally you would accommodate 100% of your passengers unless their marginal cost is just too high. You are also assuming the distribution of passengers is the same as the distribution of the population. Or the distribution of customers willing to pay certain prices is the same as the general population. Airline seating is very complex as it's not a fixed price, changes constantly, and even varies from seat to seat in a single flight. Prices aren't fixed so just because you increase marginal cost does not mean reduce profit. Lastly it also ignores completely the fact that even a shorter person would enjoy the increased room.

totalZero 9y ago
You responded to a comment about how a case is extremely uncommon, with a suggestion about half the population. All I pointed out is that your argument is less than perfectly applicable to the comment to which you're responding, by stating a mathematical fact. I didn't assume anything, ignore anything, assert anything, or even draw a conclusion about what should be done. Just spreading facts. Call me Johnny Factseed.
- nightski 9y ago
  Again you assumed a Gaussian distribution. There is nothing "factual" about that. Not to mention I was pointing out that your facts were irrelevant because the miss they point (Using population statistics to guide the design/engineering of an experience is doomed)
  - totalZero 9y ago
    I'm not designing the experience. Nor does telling you what a particular distribution would indicate if assumed, imply that I have assumed the distribution. Mathematically we would expect height to be distributed in a somewhat gaussian fashion due to the central limit theorem. Therein lies its relevance. However, nobody is 70 feet tall (yet!), so it would be fair to argue that the tails are characterized differently by a gaussian distribution than by the "true" distribution for height (which would probably divert some of that probability mass from the far tails to the near tails). Still, it is quite misplaced to talk about 50% of the passengers when the discussion is about substantial outliers. Just giving you some information that will help you use more applicable axia in the future. Signed, Johnny Factseed
    - nightski 9y ago
      That's not how the CLT works. It does not mean every random variable is Gaussian.