fiddy/fiddy

>>17046455
Undefined because either child may choose to present themselves as female

The traditional state-space expansion for the "Tuesday boy" riddle yields a filtered sample space of 27 valid microstates out of 196 total day-gender combinations, with 13 target microstates where both children are boys, resulting in the classic probability: P=13/27 or .48148
Pay Me

About 48.15%. May I know the reason why you ask?

Basic Statistics

>>17046455
I've never understood this idea. "quants" aren't a resourced defined variable under the Chicago school.

Essentially there's infinity Indians

>>17046455
50.000%

I ran a monte carlo simulation with 1000000 iterations and I got a syntax error

$5.00

>>17046478
lol
but allow me to offer some nuance: a quant is a specific type of indian

isn't it the same trick as the monty hall problem?

>>17046455
50%
either one is a boy or both are boys

1/3
Just ran all possibilities for having 2 kids
girl boy
girl girl
boy girl
boy boy
This is assorted by birth order.
Three pairs have boys, only one of the three has 2 boys
hence 1/3

Assuming the fact is genuinely conditioned as stated, the answer is:

[eqn]\frac{13}{27}\approx 0.481481 = 48.15%.[/eqn]

There are [math]14\times14=196[/math] equally likely ordered two-child families (sex plus weekday). The condition “at least one Tuesday-born boy” leaves

[eqn]14+14-1=27[/eqn]

possibilities. Of those, both are boys in

[eqn]7^2-6^2=13[/eqn]

cases.

I have now.

From 1,000,000 simulated two-child families:

[eqn]
\text{eligible families}=138{,}367
[/eqn]

[eqn]
\text{eligible families with two boys}=66{,}680
[/eqn]

[eqn]
\hat p=\frac{66{,}680}{138{,}367}=0.4819068=48.1907%.
[/eqn]

That is close to the exact result:

[eqn]
\frac{13}{27}=48.1481%.
[/eqn]

The small difference is ordinary sampling noise.


The interview trap: if the father volunteered the sentence selectively, or you happened to meet one child who is a Tuesday-born boy, the answer can instead be [math]1/2[/math]. The information protocol has to be specified.

>>17046455
This is a 50/50 joke question ain’t it

Inb4 “BUT WHAT IF HE TRANS??”

To which I tell you: “meds”

Tuesday you say?

>>17046455
I would tell the proctor:

1. Only a sucker bets on an event that already occurred. Applying mathematical analysis to an event that already occurred is a misuse of statistics. Bringing up the Monty Hall problem is applicable, because it too is a contrived game show, where Monty Hall KNOWS the outcome before you do, and never shows you the goat. In other words, a fixed game. If they fix that part of the game, what is stopping them from fixing the entire thing? No amount of statistics will help you against cheaters.

2. No couple has an even odds of the sex of their child. Therefore, using the constraint that a child is either a boy or a girl as the foundation for the probability of the sex of the unknown child would be actuarial malpractice.

3. Using clearly independent events to try and “trick” the test taker into treating the day of birth as irrelevant, when you want them to treat the problem as a joint probability state space is just a dick move, and tells me the proctor is an asshole. No sane person reads “boy born on Tuesday” as the joint probability of “boy-born-on-Tuesday” instead of the independent events of “boy, born on a Tuesday.” Ask a fucking real joint probability question if you want to see if they can count a population properly.

At this point I would realize that working for these people was going to be the most tiresome and pathetic waste of my time, where my bosses would be more concerned with showing how smart they are at the expense of everyone else, and no one would be doing any actual work, that I would simply get up and walk out. I would then notify the SEC that, if these morons were picking winning stocks, it would be almost 100% sure that they were engaging in insider trading.

File: __mare_bello_fiore_overlo(...).jpg (1.22 MB, 1200x1600)

1.22 MB JPG

>>17046467
Wrong, males presenting themselves as females doesn't imply they identify as females.

>>17046455
That problem requires only two steps: Give good names and do extremely simple calculations. Is the first step that filters out 97% of applicants.

>>17046470
There are lots of "people" chimping out in r*ddit because the answer is not 50% so they/them would fail the interview.

>>17046493
No, the Monty Hall Problem makes sense. This question asks you to consider a probability space solely consisting of sex and birth day-of-the-week which is why it's so counterintuitive because it's a seemingly familiar scenario except it isnt

>>17046587
If you want to be cheeky you can throw in the trivial fact that births are not actually spaced evenly throughout the week and are significantly rarer on weekends

File: pepe-smile-thumbs-up.jpg (49 KB, 750x699)

49 KB JPG

I like this variant
The reason the original seems like a paradox is the reasoning
>if the younger child is a boy, there's a 50-50 chance the older one is; if the older child is a boy, there's a 50-50 chance the younger one is. So either way it's 50-50
which fails because "younger child is a boy" and "older child is a boy" aren't mutually exclusive.
>if the younger child is a boy born on tuesday, there's a 50-50 chance the older one is a boy; if the older child is a boy born on tuesday, there's a 50-50 chance the younger one is a boy. So either way it's 50-50
is still invalid, but since the possibility that both kids are boys born on tuesday is so much smaller, it has less of an effect on the final probability, and so the answer is much closer to 50%
is this valid intuition?

Everyone who is not saying 50% has AI psychosis and needs meds immediately

/thread

there are external factors that influence the sex of a child, so you can't know the probability.

https://en.wikipedia.org/wiki/Boy_or_girl_paradox

Finance really isnt an issue of probabilities. It needs statistics about industry and good models of the economy, nothing depends on chances. In economics chances are guessing and not something you calculate.
When movie studios approve a movie and calculate the expected revenue, they dont calculate probabilities, they poll the population and check historical statistics hoping that the same product will make a similar revenue.

>>17046455
75%

No, you guys are retards and super autists.

>>17046798
>statistics are just a model
yeah and it turns out models are pretty useful, what's your point

>>17046812
Thares no probabilities involved. You calculate expected revenue with statistics, not probabilities like its a game of dice.
Consider insurance, companies have massive statistics about how many car crashes there will be based on location, age and economic profile of the driver, theres millions of data points. This isnt guessing the future, you simply extrapolate it from the past.
Doesnt work so well for black swans and rare events like earthquakes, although you can predict them to a point. Everyoe knows there will be earthquakes in the ring of fire.

>>17046455
depends. what percentage of births are mixed XX and XY? in what country? is mr. smith an alias?

>>17046812
13% of the population commits 50% of murder. This is a history, not a calculation.

>>17046817
there is always a little margin of error within statistics, and if you understand that it's quite a simple debunk of your whole paragraph. Extrapolation from data is at its core imprecise and inexact because as time goes on things change that can cause relationships between two variables to have lower correlation or a different correlation. statistics makes an educated guess so it isn't as random as a group of dice but every educated guess based off of statistics goes within a margin of error that in and of itself can be wrong if based off of outdated or unreliable information. the sheer amount of unreliable information out there is crazy, but I will grant you that out of everyone, finance and insurance corporations have the best datasets. it is true that finance and insurance have very reliable statistics to go off of but if it was perfect they wouldn't need to have policies to fight claims. you say there isn't any probability but yes, there will always be variability that can fuck things up for finance. nobody extrapolated the housing market crash, and all the retard analysts of today are using datasets about AI and extrapolating it to a point in the future where it might be profitable after the resources of the country are sucked dry ten times over. Real life often provides the "probability" that you claim doesn't exist

It's clearly 50/50. He is either a boy, or he isn't.

/s

>>17046817
>you simply extrapolate it from the past.
using techniques derived from probability theory