« first day (1270 days earlier)      last day (3594 days later) » 
00:00 - 21:0021:00 - 00:00

12:58 AM
@Doorknob I posted my attempt at an entry to your challenge
if you see this, let me know if it looks more or less okay
if it doesn't I'll fix it tomorrow
 
1:09 AM
also, feel free to edit my answer if there are obvious, important yet rather small mistakes
I'll go to be dnow
bed now*
 
1:22 AM
@overactor Great! I'll test it with three other simplebot.rbs. :P
 
@Doorknob apparently I'm not sleeping yet, so let me know how it's going.
 
 
6 hours later…
7:01 AM
@MartinBüttner plus there is the extra limitation that a tag needs at least 100 questions to be eligible for tag badge (see meta.stackexchange.com/a/68395/200089) which currently apply only to 6 of our tags...
 
7:13 AM
I just noticed that the Lambda Calculus examples have been updated. codegolf.stackexchange.com/questions/284/…
We got tromped.
I mean trumped, by Tromp.
Mine fails the new test.
 
@plannapus Huh, I didn't know that. But so far I'm only waiting for and badges anyway. For the latter I've got almost 400 upvotes (which would be enough for a silver badge) but only 10 answers (which isn't even enough for bronze).
 
7:32 AM
That's quite an impressive ratio.
 
7:53 AM
@overactor I'm getting a bunch of exceptions for your submission.
Exception in thread "main" java.lang.StringIndexOutOfBoundsException: String index out of range: 340
	at java.lang.String.charAt(String.java:658)
	at ShallowBlue.getCharAt(ShallowBlue.java:244)
	at ShallowBlue.calculateMove(ShallowBlue.java:181)
	at ShallowBlue.getBestMove(ShallowBlue.java:46)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
@plannapus Yours is also giving me errors:
Error in area[1, 1]:area[2, 1] : result would be too long a vector
Calls: %in% -> match
In addition: Warning messages:
1: In min(x) : no non-missing arguments to min; returning Inf
2: In max(x) : no non-missing arguments to max; returning -Inf
3: In min(x) : no non-missing arguments to min; returning Inf
4: In max(x) : no non-missing arguments to max; returning -Inf
Execution halted
Problem in submission #2 (getMove): no output received
Error in area[1, 1]:area[2, 1] : result would be too long a vector
 
@Doorknob I might have solved it, can you see if it works now? I can't access an IDE now unfortunately...
 
Sorry, have to eat. WIll be back soon
 
8:14 AM
@Doorknob Are you in a different timezone than Texas?
It just seemed strange to me that you would eat at 3:00 AM. :P
 
@Doorknob This error basically mean that I am not on the board... Do you know what was the string passed as argument when it threw that error?
 
@ProgramFOX Yeah, I'm in Korea right now. :P
@plannapus Ok, I'll check...
 
Ah, right, I saw you posting something about that in Charcoal.
 
@plannapus 3,11222222,11222222,11222444,11144444,11222444,22224444,22244444,2..44444
 
Indeed
 
8:21 AM
Looks like you don't check to see whether all of your slimes have died.
 
ok i ll add an IF statement for that. Thanks
 
Alright, thanks. Now I'll try running overactor's again
 
fingers crossed
 
I guess i didn't expect my bot to be THAT bad :)
 
i don't get that, your tactic seemed pretty good planna
 
8:24 AM
@overactor I'm still getting exceptions. :/
 
what was his bot up against?
darn it
where?
 
Every move -_-
 
same stack trace every time?
 
Exception in thread "main" java.lang.StringIndexOutOfBoundsException: String index out of range: 71
	at java.lang.String.charAt(String.java:658)
	at ShallowBlue.getCharAt(ShallowBlue.java:244)
	at ShallowBlue.getNeighboringMatches(ShallowBlue.java:207)
	at ShallowBlue.getBestMove(ShallowBlue.java:43)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.getBestMove(ShallowBlue.java:30)
	at ShallowBlue.main(ShallowBlue.java:15)
 
or several?
 
8:25 AM
^ same one every time
Except that 71 changes to -10 halfway through for some reason O_o
Oh, it changes to -10 when your only slime dies (because it errors on every move)
@plannapus Um. I think there's a slight problem with your bot. It never moves :/
Let me check the output...
result for submission #2: 7 0 8 1 7 0 6 1
@plannapus It seems to be outputting too many numbers.
 
ah ok
 
^ Go in there so the main room isn't all clogged up with slime. :P
 
9:22 AM
You know what, I think I'm just going to post the caveman KotH in the sandbox anyway because why not.
 
if anyone has time, could they please check out this sandboxed challenge? it's not mine, it's by Nate Kerkhofs
2
A: Proposed Question Sandbox - Mark XIV

Nate KerkhofsIBAN validator Write a program (can be either a function or a code snippet) that accepts a string of alphanumeric characters and validates it according to ISO 13616:2007. The algorithm for validation is (source: wikipedia article on IBAN http://en.wikipedia.org/wiki/International_Bank_Account_Nu...

 
0
A: Proposed Question Sandbox - Mark XIV

Doorknobking-of-the-hill Author note: Another KotH challenge, aimed at being slightly easier than my previous one (because I just realized how complicated it was). Caveman Duels (or: Me poke you with sharp stick) Caveman mad. Other caveman take stick but stick was for me. Caveman fight! Descripti...

 
0
Q: Changing winning criterion in a question

Cort3zI recently posted a question to codegolf which did not have a well formulated win criterion. However, I did receive some answers to the question. Because the question is on hold due to "off-topic", and further "Questions without an objective primary winning criterion..", I guess I should try to ...

 
here a onebox there a onebox everywhere a onebox onebox
 
 
1 hour later…
10:52 AM
@Doorknob Why are you still asking him for a move if he doesn't have any legal moves? I'd call that a bug in the controller.
 
@Doorknob I can't comment because of rep, when I poke and the opponent blocks, does my sharpness go down?
 
11:08 AM
@PeterTaylor Detecting whether there is a legal move is too hard. For example, in 122\n222\n222, 1 has no legal moves, but testing for this for every player for every turn would take too long.
@overactor Yes, I'll include clarification about that in the edit I'm about to make
 
It's a linear scan, which should be faster than the I/O and process overhead, so unless all of the matches go down to the wire it seems reasonable to hope it would be a net speed-up.
 
12:05 PM
@plannapus It's mostly due to a single answer (the Vigil one on "Code that only executes once")
 
1:04 PM
Don't you love it when people endorse you for skills you actually have?
Three people now have endorsed me for PHP
 
@Rainbolt by that, do you mean linkedin-endorse?
 
Maybe I know it but I just don't know it yet?
 
ugh, I don't even care about it any more
 
@MartinBüttner Yup
 
the endorsement systems seems to pointless
 
1:05 PM
How on earth would I have learned PHP?
 
It has such a very specific use like Javascript
Except that I know Javascript
Lol our employees are arguing over bagels about what is better: Marmite or Philadelphia Cream Cheese
 
@Rainbolt Out-by-not error?
 
Googled and I have no idea what that means :-/
 
The kind of error where you forget a ! or add one which shouldn't be there.
Your previous comment makes a lot more sense with the addition of the word "don't" before "actually have" and a sarcastic tone of voice.
 
1:17 PM
More sense than what?
Don't you love eating apple pie?
It's pretty common, at least in the US, for people to say, "I love this. Don't you?"
Oh I get it now
 
"Don't you love it when [the world makes sense]?" sounds semi-sarcastic.
 
My very next comment has something to do with PHP, which is a skill I don't have.
 
I'm not sure whether to read it "straight" or as sarcastic.
 
But the original comment was with reference to a skill I DO have, and you couldn't have known that because my next comment didn't follow
I was endorsed for Visual Basic (a skill I actually have)
Am I always this hard to follow? I generally don't go back and study what I said like this
It looks like I have two options: speak more to explain myself, or speak less to avoid needing any explanation.
 
Not always, no. Although I'm not sure how much that's because I'm getting used to the way you express yourself. I remember that when you first showed up we had a lot of mutual misunderstandings.
 
1:55 PM
I haz scores! Check it out: docs.google.com/spreadsheets/d/…
Going to award the bounty in 15 minutes to the submission with the highest median if nobody proves that my math is wrong.
 
Your math is wrong. Proof: Because
 
Haha I was hoping you would miss it
 
CG.SE has exploded in popularity since I was last on here.
 
I need to get better at ninjedits
 
I almost didn't see it before my two-minute mark :)
 
2:04 PM
@Rainbolt can you see what sort of results taking the geometric mean of the middle 3 results yields?
So chop of the best and worst result
then multiply the 3 middle ones
and take the cube root of that
no personal interest since I'm not in the competition
 
@Rainbolt so i m guessing that if a bot is not on the list, it means it timed out?
 
Sure
 
did the scientist and meta scientist time out?
 
That, or they threw an exception. Unfortunately, I left my list at home. I have records of exactly what round they timed out or threw an exception.
Also, I gave every submissions two chances before disqualifying them
Some of them I even fixed if I was able to
Also, I recorded my CPU activity. It was never over 30% except for when my antivirus scanner was running, then it maxed out
When my antivirus ran, nearly 20 entries failed lol
 
The scores seem to be wildly variable. If possible, I would conduct more than 5 rounds. How long did it take to run a single round?
 
2:08 PM
8.5 hours approximately
And one of the trials always fails to output meaningful results, so I get about two trials per day
Wait I'm sorry I lied. It takes 3 hours to run a trial
8.5 to run three trials
 
Instead of 5 trials of 1000 rounds, I think it would have been better to run more trials of fewer rounds each.
The fact that the scores are really variable (some really high) shows that the scores depends a lot on the mixture of contestants in the round.
So you would want to take more trials to get a better sampling of all the possible mixtures of contestants.
 
That's why I went for a median instead of an average. A single score of 450000 has no effect on a median, but can wildly sway an average
But I do agree that more trials would be better
It just takes so long to collect the results, and Doorknob has already extended the bounty on it. It was due last Saturday
 
@RainBolt are you trying the geometric mean thing?
 
Yea but I'm trying to post scores at the same time
 
multithreading ^^
 
2:16 PM
How do I write a Google spreadsheet formula to disregard the min and max of 5 items?
Also, why are we disregarding the min and max? That isn't a true geometric mean...
 
Since you have a column for min/max already, just add the items up and subtract the max/min value straight from those columns.
 
Oh that makes a lot of sense lol
 
well, divide them
 
Multiply all 5, divide by min and max, then take the 3rd root
 
yeah
 
2:18 PM
Or that.
 
it's not a true geometric mean
it's a combination of median and geometric mean
I think dropping the two extremes is a rather common way of calculating fairer mean value witha small-ish sample size
 
I'm plotting histrograms of the median and mean.
 
especially fitting when a top score is orders of magnitudes larger than the same bot's median
a few top socres even
 
Couldn't we apply the same logic and drop the two highest and the two lowest?
And then we just end up with a median anyway
 
Median histogram: i.imgur.com/MyMrqTV.jpg
 
2:22 PM
Why does =C3*D3 not work as a formula...
 
yeah, but then it's basically up to chance again
 
That's not quite a bell curve, but it's not that far off
Ugh
I even tried =MULTIPLY(C3,D3)
What's going on?
 
what does it say?
 
It literally says =MULTIPLY(C3,D3)
In the cell
I also tried =C3*D3 and it literally says =C3*D3
It won't eval
 
2:26 PM
wth
 
Here's another idea, though it won't work the same way: We can determine the order of finish for each group, and use Condorcet voting.
 
does the cell have textformat maybe?
 
I fixed it somehow.
 
well done
 
Done. Geom. Mean column is added
 
2:31 PM
Is the type set to something else?
If there's string data in the same column or row, it could force the cell to be string type.
 
you should make clear the two extremes were dropped though, otherwise people might complain
how does it compare to the median column?
 
Geometric mean histogram: i.imgur.com/bLAij8T.jpg
 
interesting
as expected, median is the choppiest
average looks nice and bellcurvey safe for the ridiculous leaders sticking out
 
It was the median which was bell-curvy
 
and geometric mean is somewhere in between
median is choppy in between
as in it has two bumps
 
2:38 PM
Is there a Stack Exchange where I can ask for the best way to interpret the data?
Or would it just be closed as "We don't care about your data"
 
mathemathics?
 
You think they would be open to the question?
 
Well there's cross-validated for statistics and data analysis, but I don't know if it's on-topic. I've never spent any time there.
 
I feel like it's too narrow
I'll see if the mathematics chat room is open
 
I think this is statistics.
 
2:40 PM
You could ask: How to get a meaningful mean value from data from small sample size with high peaks.
is there a statistics stack exchange?
 
stats.SE
 
#TheMoreYouKnow
 
The Math folks pretty much shut up when I joined and asked my question
 
I plotted the round 1 scores versus the round 2 scores.
 
at stats or at maths?
what does that look like?
 
2:47 PM
Stats is pretty barren
I asked on Math
 
X = round 1, Y = round 2: i.imgur.com/N7bZX0Y.jpg
That plot doesn't look good.
 
Yea, the point potential is the same every round, but the actual points are not
I have a team meeting to go to shortly. I guess I'll just hold off on the bounty
 
3:22 PM
Even after removing some of the outliers, the correlation between round 1 scores and round 2 scores is only about 0.09.
 
4:15 PM
There are two types of people in the world; those who crave closure and.
 
 
1 hour later…
5:15 PM
Oh nooooo... two of the participants are claiming that the raw data doesn't match up with their bot's expected behavior!
 
@Rainbolt you did it!
wat?
well
that's technically there fault for their bots having a bot, right?
 
Haha
 
Where was the question that you posted on math.se?
 
I asked in chat
One of them gave a complex formula for calculating a meaningful value
I decided to keep it simple
 
5:30 PM
I want to try doing Condorcet voting
In the "Trial 1 Pod 1" tab, are they organized by score?
It appears not.
Then Condorcet voting would be too time-consuming to do.
 
No I organized them in the order Java output the results, to avoid messing with the raw data too much
Yet I still somehow messed that up...
I'm at a loss. I don't know what happened with the results. Anyone have an idea?
Downloading Ruby to verify the results
Do I still have an Internet connection?
 
5:55 PM
No...?
 
compileonline can't even put hello world without crashing... repl.it can't accept command args
Guess I'm waiting for this 30 minute, 3MB download to finish
 
That's rather slow. Are you downloading it via the internet or using carrier pigeons, each carrying a chunk of the whole?
 
It's more like someone is trying desperately to send me chunks through the tiniest hole imaginable. I disabled something and it was fixed
 
6:23 PM
Check... my PPCG account caught up with my SO account in terms of bronze badges... now I just need 4 more silver badges
 
I'm four away for bronze. Silver... not so much. I have more silver on meta.se than on PPCG.
 
Nice answer on B&CG for me. This is hard to do because so many people there hate me
 
@Rainbolt how come? :D
 
Including the mods :-/
 
Did you start shit in chat and on meta....? ;)
 
6:26 PM
Guilty and guilty
 
I can't believe I guessed that!
2
 
Ok I need to focus on work for a few hours. The problems with the challenge are really distracting me
Later guys
 
see you
 
6:49 PM
woof
pops in
looks around
 
woof @TheDoctor!
 
HIIIIIIIII
hi @PhiNotPi
hi martin, hi rainbolt
hi what does the progrmfox say?
 
0
Q: Closing old question as duplicate of a new one

Martin BüttnerWhat is our policy regarding closing an old challenge as a duplicate of a new one? I know this happens on other sites (at least SO), if the newer question is generally better or usually if it gathers better answers (such that the old one acts as a "redirect" to the canonical answer). Now this do...

 
goes back to work
syncs dogecoin wallet
checks balance
votes on new questions
 
6:53 PM
checks rep
:D
 
codes
 
looks at his screen full of italics :P
 
looks at spreadsheet coworker emailed me
screams
feels comforted by my 162,157.00 dogecoins
wonders what @rainbolt means
 
@TheDoctor by what?
 
6:57 PM
?
 
or do you mean "What 'Rainbolt' means"?
 
yes i was wondering that
 
it's his actual name
 
It's his last name.
 
(last name)
 
6:58 PM
:o
 
And... I'm ninja'd! :D
 
we ninja'd each other
 
looks around office
 
Doesn't "ninja'd" mean that you post something, but someone else has already posted the same thing? In that case, you didn't ninja me.
 
well I ninja'd you in saying that it's his name, but you ninja'd me in clearly saying it's his last name ^^
 
7:01 PM
ah, I see
 
0
A: Are two questions duplicates if they have different scoring rules?

Martin BüttnerI think it's a bit unfortunate that this question was asked in the context of two rather similar scoring systems and is now used as a reference on community consensus. I agree with the existing answers, that a question is a duplicate if answers from an old one can be reused on the new one. Howev...

o.O
you can edit deleted chat messages, but you can't undelete them afterwards
 
yeah, I see
 
you cant edit!??
 
Recently I found something else interesting: if you onebox a deleted chat message, the content of that message is shown (if you are the owner of that deleted message).
@TheDoctor Press the up arrow, then you can edit.
1 min ago, by ProgramFOX
test
^ That's my deleted message.
If someone else would try this with my message, only a link would show up.
What would happen if I edit a deleted message, and then onebox it?
 
weird
 
7:06 PM
do you guys have a suggestion for a catchier title for codegolf.stackexchange.com/questions/34875/… ?
originally it was just "Domino Day" which was nice, but too short
 
21 secs ago, by ProgramFOX
This is my edited message.
 
now "Golfing for Domino Day" is kinda weird, especially since domino day (if it was a thing) is in November...
 
/me wonders why this isnt IRC
 
(@PeterTaylor? ^ you were the one who basically came up with Vectory last time :D)
 
/me thinks that deserves a star
 
7:08 PM
what does?
 
my first /me
 
@ProgramFOX .
huh?
 
Why "huh?"?
 
stuff
testing a theory
i wonder if my new job will pay in dogecoins?
 
7:12 PM
ah, were you testing what would happen when you reply with an empty message?
 
mm hmm
secretly i hope so
 
@TheDoctor ​
 
yes?
testing?
 
@TheDoctor yep
 
@MartinBüttner I think that the pun in "Master of the Domino" is probably a bit too high-brow.
 
7:16 PM
twhat kind of chat is this?
 
@TheDoctor A system that's made by the Stack Exchange team.
 
thmm
ti have a doge
 
I have to go now, see you later!
 
tbye
@PhiNotPi do you program propeller chips?
 
7:38 PM
Do you get notified of comments on meta answers?
 
Yes
But not of votes, because they don't affect your rep
 
Ok thanks
 
7:57 PM
Ok, I have convinced myself that median was the right choice to use. Can someone tell me if this makes sense?
If you take a median, you are the most resistant to change when a new value is added
If you have 5 trials and no more than 2 huge outliers per submission per trial, then the median cannot be affected by outliers
Wikipedia appears to back this up: so long as no more than half the data is contaminated, the median will not give an arbitrarily large result
 
The problem is that with only 5 trials, the median can still vary wildly and rather randomly
but it might be the lesser of all possible evils
 
Of course, if you get a lucky (or unlucky) pairing three times in a row, then that is unfortunate
But it's the only statistic suggested so far that isn't affected at all by TWO unlucky pairings
 
can you show me your excel spreadsheet again?
just out of interest
 
Arithmetic mean - affected by any outliers
Min - affected by low outliers
Max - affected by high outliers
Geometric mean with two outliers removed - affected when there are two outliers on the high side or two outliers on the low side
Median - affected when there are three two outliers on the high side or three outliers on the low side
 
8:03 PM
Are three outliers on the high side out of five actually outliers?
 
but the median gives a more randomly distribbuted outcome among the unaffected scores
 
Actually, your chart was scaled differently. I didn't notice it earlier
One of them had a spread of nearly 5000. The one that appeared more random was actually much more tightly grouped
A simple average varied from 1800 to 100000. A geometric mean using only the three middle scores ranged from 1700 to 5000. The median ranged from 1700 to 2900.
When you zoom in on the graph (because the scale shrunk), it looks more random
 
i think median should do btw
 
Yea. It's just a bunch of imaginary points anyway right? And I posted the geometric mean so people can say "Hey look I won that category"
 
@PeterTaylor hmmm...
 
8:13 PM
@MartinBüttner Add a rule that says "Program must complete in 30 minutes."
I guess only Americans would understand that rule
 
@Rainbolt Ummmm?
 
Oh wait... Domino's is international
 
ah
haha
well, yes Domino exists here, too
 
Yea I just Googled "Domino's London"
 
They have them in Tokyo, too.
 
8:15 PM
regarding that meta answer... was that downvote yours (in conjunction with the comment)? if so, was it only for the "delete vote" part or for the rest of it, too?
 
It was for the deletevote part, and because the other part was sort of a copy of the already topvoted answer
 
not at all
in fact Peter Taylor used the top-voted answer today to argue exactly against the point I'm making in my own answer
 
Let me reread
 
which is why I felt the necessity to post it
 
Here was the caveat that you said all other answers failed to cover: "I think for the new challenge to qualify as a duplicate, those old answers also have to be competitive."
Oh, I get what you were asking now
 
8:21 PM
any suggestions for expressing myself more clearly?
 
huehuehue
 
@MartinBüttner Sorry I'll look again in a sec
work stuff
@MartinBüttner Ok, maybe I focused too much on the last paragraph, where it looks like you are proposing that we punish answers that are copied and pasted. And even on a second read, that does seem reasonable (unless it was the same author)
 
8:39 PM
ah I'll remove that part anyway
apparently it distracts from what I really want to say
 
We are slowly but surely moving blowjob snowman off of the transcript.
 
ah after only 4 days that's not too bad
I've seen things there for 10 days
 
Well, mods sometimes pin things
 
no, I mean due to stars only
 
I was on a roll for about three weeks where I got at least 1 star per day. I think it was last month
Apparently PPCG has become desensitized to my type of humor.
I've seen some 9 or 10 star items stick around for that long. You think stars are correlated to longevity?
 
8:47 PM
yes, definitely... the moment you star something it's at the top... but I've seen newer 1-star messages drop below older 3-star messages after a while
 
Oops... our customer is asking us why their configuration log has BLAH FOO BAR BAZ TAY RAY written to it.
 
whose fault could that be?
 
I'm not allowed to install the product on my own machine, because it breaks what we call a "clean dev environment". So when I debug the installer, I just put Foos and Bars everywhere
The sad part is, I'm not allowed to check in code until I get a code review
And they have to review every file I touch
 
@TheDoctor No, I don't program propeller chips.
 
Luckily this isn't a released product. We just handed it to the customer for internal testing. Hurray for internal testing?
 
00:00 - 21:0021:00 - 00:00

« first day (1270 days earlier)      last day (3594 days later) »