« first day (1214 days earlier)      last day (2428 days later) » 
00:00 - 10:0010:00 - 00:00

10:03 AM
@ffao sorry! But I still don't see the problem
 
@boboquack X2 is the event "the dice takes 2 rolls to get the first 6, given that the rolls are even". You then claim P(X2) = P(A2 | B2), which is not true since in the case where the first roll is a 6 the second roll is not guaranteed to be even
So B2 is actually not the condition you want
 
Okay, is this about right?
GImme a sec
 
The condition you want is "all rolls up to the first 6 are even"
 
Chance that all rolls up to the first six are even is 1/3
The expected number of rolls to get a six after a bunch of evens is 1/2, where getting an odd before a six counts as a 0
 
@Wen1now is it? I get 1/4
 
10:10 AM
Therefore the expected value given that all rolls up to the six is 3*1/2 = 3/2
The chance is 1/6+1/3*1/6+(1/3)^2*1/6+(1/3)^3*1/6+...
Wait a second
 
@ffao fixed and got the correct answer, yay!
 
Second chance gets me 3/8
So that's 3/8*4=3/2
 
wait
half of my answer got redeleted...
whoops, caching
 
Yeah, my solution was pretty much oray's
If I understand his solution correctly
 
I love Bayesian statistics, when I get it correct :D :P
 
10:20 AM
But I'll see if my third solution works
Just give me a second
 
To make more explicit what was wrong with your earlier answer @bobo: assume that we continue rolling the dice after we get a 6, but now we don't care whether we get odds or evens. Clearly nothing changes about the expected value the problem asks about, that is, we don't actually care what happens after we roll the 6. But by saying that "we rolled two evens" was a given, you were caring about what happened after the 6, since a sequence such as 6 1 would have been ruled out.
 
^I don't follow that, sorry
 
I'm not sure if you got it still, certainly took me a long time to get it
 
@ffao yeah, once you laid it out on the table, I got it
 
Oh, nice, I was running out of explanations :P
 
10:26 AM
@ffao you've still got to explain it to Wen
 
@ffao link to question pls?
 
5
Q: Elchanan Mossel’s dice puzzle

GamowHere is an amazing probability puzzle that is currently going viral in the discrete mathematics community: You throw a (single) die until you get $6$. What is the expected number of throws (including the throw giving $6$) conditioned on the event that all throws gave even numbers?

 
@boboquack Now I can outsource that job to you :P
 
@boboquack so average throws until a 6 appears given that all throws are even?
 
yep
 
10:29 AM
Don't worry, I got the question
I just never worked out why bbqs previous answer was incorrect
 
@Wen1now since 622462462... was getting put in B_n for all positive integral n, not only B_1
 
Give me a second
 
Time's up!
 
tries to solve
I messed something up :/
 
Welcome to the club
 
10:34 AM
I cbs getting your original solution
 
cbs?
 
@ffao can't be stuffed
 
^
@ffao Are you ready for a 4c yet?
 
That an Aussie thing?
 
I guess then? I've got no idea
 
10:37 AM
OK, now that we have finally fixed bobo's solution I can think about a 4c
 
I don't think that's common at all though, I've only ever seen cbf (you can guess what that's short for)
 
I've only ever seen cba
 
Haha @ffao:
 
So the PPCG one is different nvm the difference in conditioning does not appear to change anything
 
programming solutions are always much easier
For practically any question like this, I'd program a solution
 
10:41 AM
eight lines of Ruby :-)
 
@JohnDvorak I can simulate an infinite number of dice rolls in 3 characters in Python 2/3 REPL - 1.5 :P
 
That's not a simulation
 
@JohnDvorak how do you know what interpreter I use?
 
As long as it follows the spec, it doesn't matter
I take it yours doesn't?
 
@JohnDvorak you'll never know >:)
 
10:44 AM
@JohnDvorak Unfortunately it has a terrible bug :P
 
Either the code is wrong, or the interpreter as well.
 
I have a question, I can't seem to work out why this doesn't work
Let the expected probability be E
Then we have: E = 1/6+2/6(E+1)
 
E+1?
Also why 1/6 and 2/6
 
@Wen1now did you mean E = 1/6+2/6(E+1)+E/2?
 
Since from the beginning we have an expected value of E
 
10:47 AM
@Wen1now Yes, but why E+1
 
How come + E/2?
 
@Wen1now if you get an odd number?
 
@boboquack But you never get an odd number
 
If you get an odd number it is disregarded
 
@Wen1now So it should be 1/3 and 2/3
 
10:49 AM
@Wen1now well then you're missing the retrial
 
Oh right
Okay, that makes sense
 
but that doesn't work either >:D
since E+1 is bad
 
I thought it did
 
'cause you might get a retrial after another roll
or another roll after that etc.
 
Okay don't worry
I can't multiply
 
10:50 AM
well the easy explanation is: chance is geometric with r = 2/3
s_n = (2/3) / (1 - 2/3) = (2/3) / (1/3) = 2
chance with starting value of 1/3 = 2 * 1/3 = 2/3
rolls = 1/(2/3) = 3/2
yay done
 
Why is the chance geometric with r=2/3...?
 
1/3 + 2/3 * 1/3 + 2/3 * 2/3 * 1/3 etc
 
GameN all
 
Looks like it's explained in the latter half of Oray's answer
 
10:51 AM
I don't get the latter half of Oray's answer
I don't see 5/9 pop up anywhere
 
@ASCII-only , that is wonderfully intuitive. My take on what you said: If it is geometric, the factor must be 2/3 because that is the ongoing probability of continuing the streak.
 
probably need more explanation
 
Woah, humn's answer is elegant
 
@humn :| no i meant it is geometric
 
Can't be sure on its face, but it almost must be.
 
10:54 AM
chance of getting a 6 is 1/3; chance of continuing (the next term in the geometric series) is 2/3
a_n = chance it takes n rolls
 
The only issue I have with humn's answer is that I'm wary of believing claims of the form "is equivalent to"
 
Guess that could be explained more . . .
 
@ffao Sorry to bug you, but when is the 4c coming?
 
When it's done
And I would appreciate it if you were slightly less insistent. :)
 
10:58 AM
Okay, sorry
 
He did apologize
 
Has everyone heard the sleeping beauty problem?
Okay, whatever. I'll paraphrase:
Someone (willingly) takes part in an experiment. They are told the rules of the experiment:
They will be put to sleep. A coin will be flipped. If heads, they will be woken up on Monday and Tuesday, for a total of two wake-ups. If tails, they will be woken up on Monday only.
Now, each time they get woken up, they will be asked if they believe the coin was heads/tails and then they will be given something to make them forget that they had woken up. Then they will be put to sleep again.
 
D: Oray uses \inf not \infty
 
On Wednesday they will be allowed to go free
 
11:14 AM
So what's the problem
 
The question is what probability the subject of the experiment should assign to the proposition "The coin came up heads". [EDITED to fix a tense error]
 
Oh, the problem is that you're the someone in the experiment
You get woken up - what should you assume about the coin flip?
 
Note that in many probability questions you can translate that question into something like "What odds should you be prepared to bet at, so as to end up not losing money whichever side of the bet you take?" but it's not so clear that that translation is the right one here.
 
squints
There must be something suspicious
 
If you do the experiment many many times, then the coin comes up heads on half the coin-flips, but two thirds of the times you get asked the question (because you get asked it twice when it's heads)
 
11:18 AM
wait is that it :/
 
If I say heads, I will be correct once per experiment on average, if I say tails, it will be once per two experiments.
 
It may be worth saying that unlike the Monty Hall problem, where there is a single right answer that everyone competent ultimately agrees on but it's easy to get it wrong at first, here you can find highly competent people arguing for (1) p=1/2, (2) p=2/3, and (3) there is no fact of the matter as to what p is. It's more a philosophy-of-probability question than a probability question.
Anyway, I need to be AFK for a bit now, sorry.
 
I call those "unclear" :P
 
Okay, now suppose you wake up accidentally in the middle of the night
What are you inclined to believe now?
 
Probably missing something really obvious but I'd say both are equally likely
Also just realized for the dice question humn did the self-referential solution
 
11:31 AM
Couldn't resist after ditching my first approach by empirical induction.
I hope the equivalence is more convincing now, @ffao, but it's too much past zzzzleepy-time to 🖉 more . . .
 
@ASCII-only let's consider another experiment: if the tails land, you get woken up on Monday. If the heads land, you don't get woken up at all. Now, clearly both outcomes are equally likely. But, what should you bet on when you get woken up?
 
I was going to do it that way at first but couldn't figure out how to do it :P
@JohnDvorak Tails :P
 
Correct.
 
Wait. So for the second part of wen's question is it still heads then? Because 3/5 > 2/5 (not sure if I can assume the probabilities like that though)
 
11:59 AM
Do you only have to guess correctly once? Or for heads must you guess correctly both times?
 
12:15 PM
@Apep Once
 
So, if your strategy is to choose one option with certainty, it doesn't matter how many times you wake up. If you chose the wrong answer the first time, you would choose the wrong answer again the second time.
However, if when waking up in the morning, you flip your own coin and claimed its result, you would have a 5 in 8 chance of being correct.
 
12:28 PM
CCCC: Member with unworldly enthusiasm earlier harbored primarily certain feeling (10)
(I wanted to say "enthusiasm for the unworldly", but then "for the" would have been unused. But I think this still works, @Sp3000 can opine after it's solved)
 
That's a very suspicious after-note :P
 
^_^
 
I would bet that the article considerably overstates things.
 
Sid
I guess.
 
1:41 PM
For c4: unwordly enthusiasm = enthusiams -earth = nusims?
 
no r
 
yeah thats what I was wondering about...
 
It could work if the rest of the clue adds an R and unworldly is meant to be applied at the end
 
2:24 PM
what is a nusim?
nvm, that's not the end
 
 
2 hours later…
4:19 PM
The only scenario I can think of in which this comment makes any sense is where "unworldly enthusiasm" and "enthusiasm for the unworldly" are not equivalent - that is, not used for their mutual meaning, but rather where (at least) one of the two words is an indicator.
The likely indicator is then 'unworldly', with a subtraction sense, as noted by others. But if that's what it is, "world" cannot be subtracted literally from surrounding tex, making it indirect either in that something synonymous with "world" is subtracted instead, or in that "world" is subtracted from something synonymous wi
 
It could just be that the order of the two matter. I think the only thing it excludes is that they are part of the definition.
 
not even that is sure
could be the unworldly type of enthusiasm
 
4:40 PM
(Literal meaning of "enthusiasm", by the way, is "having a god in you". Unworldly indeed. I make no claim that this has anything to do with the current clue.
)
 
TIL, nice
 
 
1 hour later…
6:04 PM
@Rubio Youu're reading too much. You weren't happy about an extraneous "are"; I didn't think you would be happy with an extraneous "for" and an extraneous "the"
But I wanted them for surface reading
 
Fair enough :)
 
Damned if you do, damned if you don't. The only winning move is not to win.
 
 
4 hours later…
9:44 PM
0
Q: Possible number based cipher… but which one?

Jason ButlerHeres a sequence of numbers: 133, 173, 30, 133, 176, 160, 171, 63, 137, 176, 160, 60, 177 I'm sure they are an important part of a puzzle i'm trying to solve which i found hidden in a image - but i cant identify the cipher used. Does anyone here have any ideas?

 
10:36 PM
o/
 
11:14 PM
Why will no-one try the number parts of my puzzle again
 
I tried it more than once and failed, why would I have reason to believe that I'll succeed now?
 
@Rubio Why doesn't your reply display as <@User and instead display as <:messageID?
Is this some mod thing?
@ffao because there's a hint dangling out there
I tried the dice problem more than once and failed, and I succeeded later when you gave me a hint
 
Well, Rubio's "level 0" hints are notoriously unhelpful (on purpose), so was yours :P
 
@ffao yes, but there's a level 1 hint now (which is probably higher than level 1 but anyway)
 
Also note that most people are not MCN and will only post answers if they actually succeed
Doesn't mean they are not trying
 
11:19 PM
@ffao I take the silent chat as an indicator :P
And hehe:
Hints
helpfulness level 0
^ go look at that
Also if you're scrolling through the transcript Rubio:
^do you know why that happens?
 
00:00 - 10:0010:00 - 00:00

« first day (1214 days earlier)      last day (2428 days later) »