Millionaire, should he switch?

Riskymove wrote: »

TBH, while I understand the concept I'd be happy to stick with my random selection of 1 from 3 and risk losing

It will kill me if I had chosen the car and then handed it back based on the idea that I had improved odds

they ran an experiment using this problem and pigeons and the pigeons learned that switching was the best option. Humans seem to be a lot more obtuse when it comes to a problem that contradicts what they think is obvious.

In terms of how random the selection of wrong answers that the 'computer' removes, it strikes me that this could be predetermined by the question creators. The questions become progressively harder, so the effectiveness of 50:50 might be designed to scale too.

For example, the first question is always extremely easy, but I'd imagine the writers would ensure that if somebody did have to use 50:50 at this stage, they would keep the most implausible wrong answer alongside the right answer. For the final question they could leave the most plausible wrong answer to make it harder for the contestant.

pleas advice wrote: »

A boat's a boat, but the Mystery Box could be anything! It could even be a boat!

or a goat.

Python code if anyone wants to experiment:

import random
import sys
import io
from contextlib import redirect_stdout

# WWTBAM - 1 answer of 4 is correct. 
# Contestant  mentally selects one of the four. 
# Computer removes two random incorrect answers. 
# Should contestant now switch?

######################################################
# One iteration - prints STICK if intial guess 
# was correct, SWITCH if it remains and was incorrect, 
# and OOPS  if it does not remain  after the removal.

def one_go():
    global stick
    global switch
    global oops
    options=['A','B','C','D']
    correct = random.choice(options)
    print('The correct answer is ', correct)

    guess = random.choice(options)
    print('The initial guess is ', guess)

    incorrect = list.copy(options)
    incorrect.remove(correct)
    print('The incorrect options are ', incorrect)

    for i in range(2):
        removed = incorrect.pop()
        options.remove(removed)
    print('The remaining options are ', options)

    if guess in options:
        if correct == guess:
            stick=stick+1
            print('STICK')
        else:
            print('SWITCH')
            switch=switch+1
    else:
        print('OOPS')
        oops=oops+1

#######################################################
# Prompt for number of iterations
# Returns counts for 
# stick (where you should have stuck)
# switch (where you should haves switched)
# oops (where your guess was removed at random)

stick = 0
switch = 0
oops = 0

trials=input('How many trials would you like?')

junk='Invalid'
while junk not in ('y', 'n'):
    junk=input('Do you want to see the output junk? (y/n)')

for i in range(int(trials)):
    if junk == 'y':
        one_go()
    else:
        trap=io.StringIO()
        with redirect_stdout(trap):
            one_go()

print('STICK=', stick)
print('SWITCH=', switch)
print('OOPS=', oops)
#######################################################

Output from 100000000 trials:
STICK (better off sticking): 24997043

SWITCH (better off switching): 24990445

OOPS (your guess was removed during the 50:50): 50012512







Somebody please hire me

abff wrote: »

If he doesn't know the answer and tries to figure it out, I think most people would refer to it as a guess. But why use a common word when presented with an opportunity to split hairs?

The point is that it's not a random selection process anymore.

This nothing like the Monty Hall example at all. /thread