Posted on 11/03/2016 7:04:30 PM PDT by workerbee
A San Antonio Thai restaurant, Yaya, will let you use its Wi-Fi for free-- but only if you can solve this insanely complex math equation.
Reddit user Joshua_Glock posted a picture of the restaurants handwritten Wi-Fi equation this past weekend, but no one has been able to connect to the network yet, First We Feast reported.
(Excerpt) Read more at foxnews.com ...
That's great. I have math degrees. I know what the equations means.
What's the password?
ML/NJ
P(m >= N/2) = (SUM from m=N/2 to N)[(N/m)(0.25)^m(0.75)^(N-m)]
That's no formula - that IS the WiFi password!
It’s not a complex number.
I don’t know what half of those signs are for, I was never a math whiz. Did pretty good at the regular stuff, addition and subtraction, fractions, but calculus is out of my league.
These days I’m surprised anyone can count change for a dollar. I had to do it in my head, they didn’t let us use calculators. of course, a calculator that would fit in your pocket didn’t exist until just before I got out of high school...
By my calculations, my meal was free.
Get Your Yayas Out?
Sittin in LaLa
Waitin for my Yaya
Uh huh
Uh huh
wait doesn’t the second half of the equation state that m = N/2 to N? Therefore the probability of m>=N/2 would be 100%?
I think it’s a trick question that they give you the answer built in.
“Looks like the upper half of a binomial distribution, in which case the answer is 1/2. Maybe the password is one-half, or some variant thereof. But, if thats the case, the term written as (m/n) after the summation sign should be the binomial coefficient, which is normally written without the division sign, just as the two numbers over each other inside the parentheses.
But, Im a bit rusty, to tell the truth.”
I hate you. :-)
42
“Its so like math people to screw up the spelling of customers.”
It was Halloween and he was referring to those who came in “costumes”.
Discrete Probability Distribution
The password is “1” or maybe “one”.
Alcohol and calculus don’t mix.
Never drink and derive.
p@ssw0rd
I don’t have a math degree, but given p_sub_success = 0.25, the likelihood of achieving at least N/2 successes is small:
NotLikely
ChancesAre
LongOdds
OddsAgainst
BadLuck
SuckerBet
Call me obsessed, but I actually plugged in the values for N=8, and M gte N/2 to N, incremented by .1. I then summed the P(M) values. The answer I came up with is 0.000483, which is meaningless.
I did note that as M and N got larger, the result approached zero. Once M and N went negative, that all changed.
It should also be noted that the answer cannot be a real number since not all Ns are evenly divisible by 2, making the quotient irrational and any product irrational also.
Someone plug this into Wolfram and see how it graphs out.
LOL, very good!
*you’re :-)
Since it was Halloween, “GhostOfaChance”?
In order to be "the" winner you must show your work, systematically until the results show 42 is the answer.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.