Jhereg
Elite
- Joined
- Dec 13, 2005
- Posts
- 2,657
- Society
- Rangers
- Avatar Name
- Feng Huan SecretAznMan Zho
Just putting this here while it is on my mind.
If (based on anecdotal evidence) we assign a probability to the chance of getting a 5k per hour, we can use a binomial distribution in order to calculate the approximate chance of looting a particular number of 5ks per 15 hour run.
Let's assume that, for a given kill rate that's close to optimal for a category, you end up with a 2.5% chance of looting a 5k point per hour.
Using the binomial distribution (https://en.wikipedia.org/wiki/Binomial_distribution) and excel!, you can calculate the following:
Chance of looting exactly 0 x 5k within a 15 hour time frame: 68.4% (1 in 1.462)
Chance of looting exactly 1 x 5k within a 15 hour time frame: 26.31% (1 in 3.8)
Chance of looting exactly 2 x 5k within a 15 hour time frame: 4.72% (1 in 21.2)
Chance of looting exactly 3 x 5k within a 15 hour time frame: 0.525% (1 in 191)
Chance of looting exactly 4 x 5k within a 15 hour time frame: 0.04% (1 in 2500)
Cumulatively, the odds of looting at least 1 x 5k = 31.6%, or 1 in 3.2.
Just putting this out there. The exact numbers are for sure not right, but it gives you a sense of what the possibilities are.
Now you might say, how can anyone get 3x 5ks, that's 1 in 191! Should be super duper rare.
That brings us to something akin to the birthday paradox (https://en.wikipedia.org/wiki/Birthday_problem). Basically, as the number of things sampled gets larger, what seems rare is suddenly not so improbable.
https://www.scientificamerican.com/...-long-shots-miracles-and-winning-the-lottery/
Because there are many players, probably more than 10 per category, and each player might do multiple runs, you are bound to see one more more who get insane scores.
Hope this helps some people, sometimes I like being a nerd and sounding smart lolol
If (based on anecdotal evidence) we assign a probability to the chance of getting a 5k per hour, we can use a binomial distribution in order to calculate the approximate chance of looting a particular number of 5ks per 15 hour run.
Let's assume that, for a given kill rate that's close to optimal for a category, you end up with a 2.5% chance of looting a 5k point per hour.
Using the binomial distribution (https://en.wikipedia.org/wiki/Binomial_distribution) and excel!, you can calculate the following:
Chance of looting exactly 0 x 5k within a 15 hour time frame: 68.4% (1 in 1.462)
Chance of looting exactly 1 x 5k within a 15 hour time frame: 26.31% (1 in 3.8)
Chance of looting exactly 2 x 5k within a 15 hour time frame: 4.72% (1 in 21.2)
Chance of looting exactly 3 x 5k within a 15 hour time frame: 0.525% (1 in 191)
Chance of looting exactly 4 x 5k within a 15 hour time frame: 0.04% (1 in 2500)
Cumulatively, the odds of looting at least 1 x 5k = 31.6%, or 1 in 3.2.
Just putting this out there. The exact numbers are for sure not right, but it gives you a sense of what the possibilities are.
Now you might say, how can anyone get 3x 5ks, that's 1 in 191! Should be super duper rare.
That brings us to something akin to the birthday paradox (https://en.wikipedia.org/wiki/Birthday_problem). Basically, as the number of things sampled gets larger, what seems rare is suddenly not so improbable.
https://www.scientificamerican.com/...-long-shots-miracles-and-winning-the-lottery/
Because there are many players, probably more than 10 per category, and each player might do multiple runs, you are bound to see one more more who get insane scores.
Hope this helps some people, sometimes I like being a nerd and sounding smart lolol