RE: Unfair Coin Flip
September 30, 2017 at 5:20 am
(This post was last modified: September 30, 2017 at 5:23 am by Edwardo Piet.)
So... I reread the OP.
"simulate a fair coin flip" was indeed mentioned from the very beginning.
But this doesn't seem like a simulation of a normal fair coin flip. As a normal coin flip isn't flipping the coin multiple times and calling that a 'simulated flip'. A simulation of a flip isn't grouping multiple flips together and calling that a simulated flip.
I don't know how it can be claimed that a redefinition wasn't used.
It's not really a single flip that's being simulated here. It's the very purpose of what a coin flip is for that's being simulated. i.e. you want to use a normal coin to achieve 50/50 fairness and you can do that by redefining what a coin flip is. Flipping multiple times the normal way, ignoring H/H and T/T results and adding those results together and calling that a 'simulated flip'. i.e. : When it's NOT a single normal coin flip being simulated. You're flipping several times the normal way without simulating anything. What you're doing is achieving the same purpose as a normal fair coin flip has (achieving 50/50 fairness) by redefining several coin flips with T/T and H/H results discounted as a 'simulated flip'. A truly simulated flip would be an accurate enough simulation to work exactly the same way as a normal flip does. i.e. you'd still flip the coin the same way.
I didn't think discounting certain results would be allowed because we're supposed to be flipping a coin here to achieve 50/50 fairness... and to discount some of those results is OBVIOUSLY gong to change the unfairness if it isn't 50/50 fairness.
If we're told to achieve 50/50 fairness with a coin that isn't 50/50 fairness but it isn't mentioned that we're totally allowed to ignore some results.... then suddenly it feels like a trick question again.
"simulate a fair coin flip" was indeed mentioned from the very beginning.
But this doesn't seem like a simulation of a normal fair coin flip. As a normal coin flip isn't flipping the coin multiple times and calling that a 'simulated flip'. A simulation of a flip isn't grouping multiple flips together and calling that a simulated flip.
I don't know how it can be claimed that a redefinition wasn't used.
It's not really a single flip that's being simulated here. It's the very purpose of what a coin flip is for that's being simulated. i.e. you want to use a normal coin to achieve 50/50 fairness and you can do that by redefining what a coin flip is. Flipping multiple times the normal way, ignoring H/H and T/T results and adding those results together and calling that a 'simulated flip'. i.e. : When it's NOT a single normal coin flip being simulated. You're flipping several times the normal way without simulating anything. What you're doing is achieving the same purpose as a normal fair coin flip has (achieving 50/50 fairness) by redefining several coin flips with T/T and H/H results discounted as a 'simulated flip'. A truly simulated flip would be an accurate enough simulation to work exactly the same way as a normal flip does. i.e. you'd still flip the coin the same way.
I didn't think discounting certain results would be allowed because we're supposed to be flipping a coin here to achieve 50/50 fairness... and to discount some of those results is OBVIOUSLY gong to change the unfairness if it isn't 50/50 fairness.
If we're told to achieve 50/50 fairness with a coin that isn't 50/50 fairness but it isn't mentioned that we're totally allowed to ignore some results.... then suddenly it feels like a trick question again.
