1000 math PhDs got this wrong - the Monty Hall problem

[Riddles]

it kinda is based on maths. like entirely. Just cos it's in a cool youtube video or written as a paragraph to make it easier to digest doesn't mean it's not mathematical.

[dickman69]

Dumb thread wub

U blew it

[kaili]

no why would I? If I get a goat I can milk it + just download the car

i wonder if this will be possible in the future with 3d printing
pirating cars loll

[ezza]

see man driving a pirated whip

had to download it cus i aint got ps n dat

[wub]

i wonder if this will be possible in the future with 3d printing
pirating cars loll

Think we should be more concerned with widespread deployment and the ramifications from the spam email industry.

[lovelydivot]

it kinda is based on maths. like entirely. Just cos it's in a cool youtube video or written as a paragraph to make it easier to digest doesn't mean it's not mathematical.

It's mathmatical - but only after the assumption is fixed...it's not a wholey random pattern.

[Riddles]

what assumption needed to be fixed?

[lovelydivot]

The assumption is - that if you picked a goat - they couldn't show you the car...

The game would be over.



They couldn't show you the car and be like - do you wanna change your answer?





Wait - That still doesn't make sense....

It makes sense only if you know you've picked a goat....

I don't know - I'm back to thinking it's 50/50.

[Riddles]

nope, it always makes sense because you don't know what you've picked.

The question isn't "what guarantees you the car every time?" it's "does swapping or not give you the best chance of getting a car?"

It's not a practical exercise its a theory, in practice you could do the experiment 1000 times and not get the car at all. It's unlikely, but it's definitely possible.

[lovelydivot]

I'm going to do this experiment right now...

I'm going to do 20 where I keep my original pick...
and 20 where I always switch...

I will report back in 10 minutes with my findings...

Ok
so my control group got 9 wins
my variable got 12

so 45% of the time - I won keeping my original pick...
and 60% of the time I won with the switch...

That supports their arguement...

I don't know if 20 rolls is enough though...

[hugh]

ALWAYS SWITCH

sorry if there was still a debate but I've arrived to clear things up now yw

[AxeD]

Statistically, not switching means going with your guess when you had a 1 / 3 chance,
switching means getting a 2 / 3 chance.

If I approach it logically, I think I can disregard the opened door with the goat and say: I have a
1 / 2 chance of getting the car now.
But it has been done already in practice, even with 100 doors etc.. and the result was
the chance does in fact increase when you switch.

[lovelydivot]

Ok I did a second round of 20...

my control got 4 wins at 20%
and my switch got 10 at 50%

...so the switch people are still in the lead....


I am still not completely convinced...
-why does the right compound to the larger group?


Ok - so why would you bet on the 1/3 group over the 2/3...like AxeD said
the switch just gives you the opportunity to opt back into the original 2/3 group...ok

[skimpi]

if he didnt know what was behind the doors then maybe its best to switch, but as he does, how you know he not trying to trick you out/double bluff you?

[lovelydivot]

It's actually really simple, once you work it out...

If someone said to you - do you want 1 chance to win a car or 2 chances...
You're gonna go with 2...for a reason

- and the switch is you opting into the group with 2 chances.

What is confusing is that they don't offer it straight away...
Your first pick just defines which group has 1 chance and which group has 2...

He's always going to show you the goat - so that changes nothing...

but the odds are always in favor of the group with 2 chances.


I am in no way endosing two timing
- though it may increase your chances of finding a mate...
It can also sabotage the whole thing, as people are sentient.

[rickyarbino]

Didn't get it until lovely divot's post.
I think the way they asked it is mildly misleading though.

Asking someone, when they have a 1÷2 chance of getting it, if changing their choice of one of the two will give them a better chance of getting it doesn't make a difference because if you have two options, either one you pick is equally as likely as the other.
But fair enough.

[Muncey]

Fuck statistics, if you change your mind from your initial decision you always feel like more of a tnuc if its wrong.

Never switch.

[rickyarbino]

If it's wrong, you'll feel like way more of a tnuc than you would feel if you would've if you had done.

[lovelydivot]

If it's wrong, you'll feel like way more of a tnuc than you would feel if you would've if you had done.

ha - the ethics of it...

...I maybe crying in the hood....
...but at least I understood....

[AxeD]

Didn't get it until lovely divot's post.
I think the way they asked it is mildly misleading though.

Asking someone, when they have a 1÷2 chance of getting it, if changing their choice of one of the two will give them a better chance of getting it doesn't make a difference because if you have two options, either one you pick is equally as likely as the other.
But fair enough.

Not sure you got it and lovelydivot doesn't seem to be completely convinced either

Let's approach it from the 'goat perspective':
We can all agree the chance of picking a goat at first is greater than picking the car.*
When this happens (2/3 chance) the host is forced to open the other door with a goat behind it.
Now think about this for a minute: switching now actually means you get the car!

"Yes, but that's assuming you picked a goat the first time and not the car."
The thing is... we should assume this, because the odds of this scenario happening are twice as big as picking the winning door before the host gives you any information.

(* this becomes vital as soon as the host open the other door with a goat behind it)