Why would Dennis need to bet on it if he could just split my winnings?

Probably like four pairs of socks

We'll say anything Dennis gains from this goes to getting a mathematica license so we can get mthmtca on TIO.

And then everyone wins

He'd probably need a commercial license so good luck

mthmtca is different tho, right? In the same way MATL can exist m

MATL uses Octave on TIO.

and MATL Online

codegolf.ninja is $7.59

@quartata you mean OCTL

@MartinEnder the commentator is good

but I never played the game

is it impressive what they're doing?

it doesn't look too hard (15 mins in)

Ok. Mathematica enterprise is 7.5k

We'll make a Kickstarter

@orlp What game?

Dota

CMC: sequence of OEIS sequences that contain themselves

What does it mean for a sequence to contain itself?

@quartata shovel man

@Rainbolt *to contain its index

Using the index in the URL?

As in, the A-number. E.g., A000055 would have to contain "55" as a member

Oh, I was thinking more along the lines of "the sequence contains itself as a proper subsequence".

Sufficient to check the terms listed, or do we need to enumerate the whole sequence?

Terms listed. The latter would be impossible without some scary data analysis methinks

unless if you did it in mathematica and harvested the solution posted there

@ConorO'Brien even then it might not be possible to determine whether a certain number appears in a sequence.

@flawr this is true

It would be decently easy to loop and screen scrape with a ?|{} clause in PowerShell, but I'm on a phone.

Golfing a dozen byte golflang answer on a phone, sure. Golfing a several hundred byte PS answer, no thanks.

I'd like to mention that recently I've seen a worrying trend on PPCG

what would that be?

if any challenge is somewhat hard and doesn't include a direct formula to calculate what is asked there often trickle in a couple close votes as unclear what you're asking

or comments demanding a formula

this often happens to my challenges, where I intentionally leave any particular direct formula out

another example is this: codegolf.stackexchange.com/questions/106260/…

IMO code golf is also about understanding what you're doing, and learning about many different things, more than just "take formula, apply, and golf"

now my wizard has two pets!

@DestructibleWatermelon diablo 3?

nethack

on the page for homunculus, my new pet

> Because of their intrinsic sleep property of their melee attack they also make great pets.

@orlp I agree. I can see how one can argue that it's unclear (because ideally challenges should be self contained), but if it's something obscure, then I'd think a formula might be needed, esp. if ambiguous.

I'm doing well now

EMACS IS BETTER THAN VIM! grabs popcorn, ducks out of room

@noɥʇʎԀʎzɐɹƆ Didn't I tell you yesterday to stop posting random messages?

> Avoid posting messages that add nothing to the conversation (^, ಠ_ಠ, rofl, etc.).

see sidebar, read context

reads sidebar I still have no context. Though I did find this:

@noɥʇʎԀʎzɐɹƆ Still totally random from what I can tell.

@El'endiaStarman I'm wondering if we have enough data to do supervised training for topic detection

I can definitely say that this was random. It's not justification, but sometimes I believe noise is being nuked on the spot too often.

In the little preview that I got while in another room, I saw "3 hours ago, by Dennis Sheesh, [...]" and I came to figure out who Dennis Sheesh was.

@quartata Hmm. Perhaps. Remember that idea I had about looking for messages close in time that shared rare words?

Yeah

aw yes three pets now

@ConorO'Brien definitely not, if a correct solution can be unambiguously verified, and an incorrect one rejected, there is no formula needed to calculate it

@noɥʇʎԀʎzɐɹƆ I'm still working out why some random messages are bothersome whereas others aren't. That said, yours was intended to incite drama whereas quartata's was sharing a humorous life experience. The latter contributes more to this community than the former.

The problem is that one word might not be enough. I was thinking about if we should segment it into n grams when I realized that a convolutional neural network would be much easier for at least getting the topic from a whole conversation

@orlp I think I know what you're saying and likewise agree, but what do you mean by "no formula needed to calculate it"? I assume you mean that said formula if existent doesn't need to be spoon-fed to the golfer

I think somethng simpler can be employed for whether a message is in a conversation

@El'endiaStarman It was supposed to be a sarcastic joke

A lot of easy to define classifiers for that -- pings, word frequency, time, etc

@noɥʇʎԀʎzɐɹƆ Nothing in the message I replied to indicates that it was meant to be a joke or sarcastic.

@noɥʇʎԀʎzɐɹƆ In response to what

@ConorO'Brien for example, 'given a polynomial, return (one of) the extreme values'

this is easy to verify

@El'endiaStarman sigh (I have never seen a joke that ends with /s in the sidebar)

and unambiguous

Another great example right here: noise is the topic but not rare enough

but I didn't give a formula as to how to calculate that

Extracting the topic is better suited to full blown ML

@quartata huh?

how about a ML algorithm to decide which messages should get moved to trash?

That'd be certainly easier to gather data for.

twitch.tv/gamesdonequick

super monkey ball 2 being run right now

looks really difficult damn

@orlp ahhh okay. yeah, total agreement

@noɥʇʎԀʎzɐɹƆ ...

Huh?

@quartata Hmm. I think "noise" is a rare-enough word because it's not in the top 1000, say. This site has it at rank 2030.

same applies to "..."

Hmm. That's anoher good question actually -- where should we get our frequencies?

Something more suited to programming would be nice

@quartata Perhaps from the corpus of the transcript itself?

I don't know if that's enough data

Could we get volunteers to sift through and categorize the transcript maybe? I don't think unsupervised will work here

Perhaps we can use some freely-available English word frequency list and then modify it with transcript data?

Like put together the conversations and agree on topics

how about a ML algorithm to predict which messages will get moved to trash?

@noɥʇʎԀʎzɐɹƆ Quit trying to instigate drama.

I think we can all agree repeating noise is worse than posting it once

@El'endiaStarman I'm not. See the above permalink.

died because I didn't have enough food...

What happened to assume good faith?

I thought putting it in present tense would be better

It could be a tool that people could use to see how noisy their messages are

@noɥʇʎԀʎzɐɹƆ I'm not confident that you're asking stuff like this in good faith.

@DestructibleWatermelon Didn't have your prayer ready?

@quartata um...

> Sometimes X is hard to perceive across the 26 symbols of our alphabet.

well, I guess I forgot that you had to wait until you were weak and not hungry

@noɥʇʎԀʎzɐɹƆ Yes, because that's a great idea. Encourage users to post messages just a little less noisy than the threshold.

*could

@noɥʇʎԀʎzɐɹƆ You say that but you're not doing anything about it.

That would be much easier to categorize, message and binary decision variable

@DestructibleWatermelon haha it's ok I do that a lot

I intentionally make it very obvious when I'm being sarcastic, and when I write something that appears rude, I either apologize for it and/or correct my statement, or I actually am being rude.

I had five pets

two floating eyes, a homunculus, a lichen, and a kitten

I try not to be rude, though, and I try to avoid writing stuff that seems to be rude, so that doesn't happen often.

@DestructibleWatermelon Didn't lose much then

homunculus is good pet

Not especially

Not sure why you charmed the lichen :P

The last time I got to the mines it isn't even saved on this computer now

@quartata IDK either

@quartata How much would we have to do?

probably should have killed and eaten it

@El'endiaStarman which is why you and I produce most of the messages (ending) with :P

I had both create and charm monster spells

@flawr If you could, I'd appreciate it if you could validate the answers posted to your challenge, since I'll be awarding a bounty soon. I'd like to know for sure that they're valid. (I suppose it's possible they take too long to test, but hopefully not.)

but, yknow, too hungry for it to help when I died

@ConorO'Brien Yeah, :P is very useful for showing that I don't intend to be rude/mean/whatever.

@Dennis You're making me hungry.

@El'endiaStarman yeah, it's really quite a fantastic face

@flawr Just gotta wait for that special time when the period is hit.

@El'endiaStarman The whole transcript would be preferable. A conservative estimate of 20 messages per conversation leaves 40k for the training set and 10k for the test set which is a low for a deep network but passable in a pinch

@orlp I played the game last year (I can highly recommend it... it has everything that was great about the old SNES platformers but really benefits from more modern game design... also free DLC that is basically as big as the main game). some of the stuff later on in the speedrun is super impressive (there are some tricky skips and quick kills). it's also a very entertaining run, a) because of the commentator and b) because of how close the result is.

@quartata Um......what? You want people (including you and me) to annotate conversations over the entire transcript?

@MartinEnder btw did you just see the monkey ball run?

damn that looks so hard

@El'endiaStarman I didn't say it was feasible....

nope, don't really know monkey ball

@MartinEnder I didn't know it before this either

the next things on the schedule for me are SMB and Ori

it's 100% instant intuitive and super entertaining

@MartinEnder Pokemon Emerald???

But an unsupervised model would probably be more of a pain

haven't played pokemon since Gold

Pokemon Emerald is my favourite out of all

@DestructibleWatermelon shouldn't eat pets anyways

@quartata I meant instead of charming it

Oh yeah absolutely

@El'endiaStarman I could see it working with a shallower network - in that case we could probably just work with one or two years worth

I kind of don't know for sure

Maybe start with a small set and see what the evaluation error looks like

The message in conversation detection would need far less. Only one layer

With a good chunk of volunteers I could see us processing a year pretty quickly actually

@El'endiaStarman Is it worth it though? :P It's not like we could run the bot in TNB

@MartinEnder twitch.tv/gamesdonequick

a game very similar to monkey ball is being ran now

actually

it IS a monkeyball run

why was i suspended earlier

(I read the runner name, pokehero, and somehow thought it was the game)

oh, this kind of game

so close

@MartinEnder do you see what I mean?

yep

@MartinEnder the previous runner was better though (less mistakes)

or maybe this one is significantly harder

@xnor I like your meta post on combining answers

@El'endiaStarman what about this: chat.stackexchange.com/transcript/message/34663878#34663878

@MartinEnder this version and/or runner is significantly less entertaining than the other one by 'Peanut'

so if you thought this was somewhat entertaining I can highly suggest checking that one out later on youtube (was streamed like 15 mins ago)

like this guy misses every other trick =/

I'll check that out tomorrow

thanks for the recommendation

HOLY CRAP ITS POKEMON TIME

\o/

grab your popcorn

have fun :) I really need some sleep

@noɥʇʎԀʎzɐɹƆ that was sarcasm >_<

@orlp what? pokemon? Did I miss something?

@Pavel no! starting now!

twitch.tv/gamesdonequick

any% pokemon emerald GLITCHLESS!

3 hours of glorious pokemon!

"any%"?

also, emerald!!

@Maltysen basically get to the end game scream ASAP

ye

no extra stuff

100% is doing all content, side quests, items, collectibles, 151 pokemon etc, w/e is relevant for the game

I'll watch the VOD later

emerald best pokemon version period

@Pavel blasphemy!

[Insert obligatory pokemon uranium best pokemon]

@Maltysen end game screen* not end game scream

I love it so much

that whatever agdq is streaming

is #1 on twitch

ONE stream has as many viewers as all league of legends streams combined

@orlp how long till it starts btw?

meanwhile the announcer is streamed through a mcdrive system

@Maltysen < 10 mins

no time is shown on stream

but it's next up and no setup block in between

k

heyyy they have undertale on saturday

isn't undertale the closer?

ye

really big of them to do that

@orlp ?

@EᴀsᴛᴇʀʟʏIʀᴋ well the closer is the most important run of the event

ah

it's the prime slot

nice then, undertale is epic

@Maltysen starting!

ayy it started

ninja'd

So, I was playing progress quest, right? 101.25600000028498 seconds remaining

Updating every ~3 seconds

@orlp did he just waste a minute to go back and increase the text rate?

think so

why did he save?

ohhhh smaaaaaaart

@Maltysen no

he chose 'A' as trainer name

out of habit

but the donation incentive was 'Daisy', so he had to remake

IMO ehs hould've reset time but w/e

oic

damn mudkip got pounded

HAHA this is going to be a reset

Does anyone here have experience with Pyinstaller?

the IRC channel apparently doesn't want to answer e.e

BARELY

@TrojanByAccident no, but ive used cx_freeze

@Maltysen link?

cx-freeze.readthedocs.io/en/latest

why's he saving so much?