Dear Adam, can we truly compare the way of calculating the probability of someone having the same birthday in the same room with someone one having a disease? A birthday is something objective but being in the room despite being ill is in itself another equation to add as it will depend of other factors... Sorry if I am missing something obvious and THANK YOU for your so comprehensible articles, a pleasure to read them.
That’s a fair point. Would probably have been clearer to say ‘person meets an objective pre-defined criteria for a flu infection’ I.e. has a clear yes/no answer
This is cool but it seems the "Secretary Problem" solution here is very limited vs real world situations. Seems it relies on a few assumptions:
1) you are not looking to get as good of a candidate (as high of a number) as reasonably possible, you are trying to maximize your odds of getting the single highest number. In real life there's a benefit to ending up with the 2nd highest number over the lowest number, but here we make a sharp line between the absolute best and everything else
2) the distribution is strictly random. If we find the underlying distribution is say normal, then once we have enough experience to calculate a mean and std deviation, we should be able to recognize true outlier results and calculate the odds of finding something higher in the remaining observations
Dear Adam, can we truly compare the way of calculating the probability of someone having the same birthday in the same room with someone one having a disease? A birthday is something objective but being in the room despite being ill is in itself another equation to add as it will depend of other factors... Sorry if I am missing something obvious and THANK YOU for your so comprehensible articles, a pleasure to read them.
That’s a fair point. Would probably have been clearer to say ‘person meets an objective pre-defined criteria for a flu infection’ I.e. has a clear yes/no answer
This is cool but it seems the "Secretary Problem" solution here is very limited vs real world situations. Seems it relies on a few assumptions:
1) you are not looking to get as good of a candidate (as high of a number) as reasonably possible, you are trying to maximize your odds of getting the single highest number. In real life there's a benefit to ending up with the 2nd highest number over the lowest number, but here we make a sharp line between the absolute best and everything else
2) the distribution is strictly random. If we find the underlying distribution is say normal, then once we have enough experience to calculate a mean and std deviation, we should be able to recognize true outlier results and calculate the odds of finding something higher in the remaining observations
Math can sure be very interesting! Your students are lucky to have you Adam!