@AidenChow Maybe create a bot that creates math questions and answers while the other tries to solve it? Also the created math problems has to have sense. So you can;t have Question: "What is 1+1? ", Answer "3" or something

@AidenChow that is cool. It’s interesting that the auto completion is much more useful than the search result

but technically anything before "://" is likely a protocal

that might make an interesting challenge if specified clearly. the biggest problem i see at the moment is how you are going to rigorously differentiate sensical math problems from nonsensical ones. i feel like that is going be very subjective, and quite hard, if downright impossible, to define rigorously

for example, some ppl might find the questions outputted from a bot to be nonsense, but other ppl might say otherwise

for instance

"proof the famous-very-hard-to-solve-some-problem" and the answer just copy-pastes

well my point is that u need a set of rules to properly distinguish between the two

hmm

maybe greed control from aops?

artofproblemsolving.com

i think that will be good let me search dupes

> We couldn't find anything for greed control

NICE!!!

hmm ig greed control could be used to make a koth, idk how interesting it will be tho

what does ig mean

i'll draft the question... hmm

"ig" -> "i guess"

i'm surprised there's no dupes

is there a diff sandbox for koth?

@NumberBasher nope

u can post ur koth idea to the sandbox

ok

lemme draft q

u dont need controller script while its in sandbox im pretty sure, but make sure u have a controller script before posting to main

@UnrelatedString just ascii

i mean what if there's like way too many langs

@NumberBasher i think language specific koths are okay

OK I'll do python

Just to make it actually non-trivial you get to know what the bots chose last time but you don't know which which (so you might know two chose 25, but you don't know which 2) and i guess we will choose from 5 nums?

the problem is greed control takes about 10+ bots

i think a distribution of the guesses from the previous round would be a good idea

yah for me i would give all rounds

@NumberBasher ok that sounds good too

what if a bot raised an error

we kick that bot outta da game?

@NumberBasher take it out when running the code and put a comment under the post saying that theres an error

once its fixed u can put it back in

ok

maybe we can encourage previous entries by running one greed control every day, say?

ig u can do that to keep an updated leaderboard

like run 1000 rounds per day or smth

and see which bot gets the highest points total

i would like 10 round per day and we add all the records

@AidenChow i think i'll only give one round of info actually

i'll probably do it like i store last times records into NumberBasherInput.txt and substitute NumberBasher for your username and you will write to NumberBasherOutput.txt and you name your file NumberBasher.py?

i think there will be too few bots for 10-100 tho, maybe allow each bot to do 2 or 3 guesses ?

@NumberBasher koth submissions are usually a class iirc

and the controller script takes in a list of class objects i think?

ok I'll do classes then

@AidenChow wdym dont understand

i would preferrably start with this

users = [NumberBasher, YourUsername, BotAnother]
bots = [user() for user in users]

and you name your bots NumberBasher?

@NumberBasher r u saying ppl just all name their bots numberbasher ?

like your username without spaces

(Tho I hope there's noone named Num BerBas Her or something...

i think ppl like to have some freedom in what they name their bots

like RandomGuesserBot or smth

u could look at past koths to see how they do it

OK

I'm okay with that

@NumberBasher yeah this looks fine to me, tho maybe one of the attributes of each class could be the previous round distribution

like instead of user() it could be user(*dictionary of the distribution*)

and require that all submissions have to have an __init__() to handle that

so maybe smth like:

class ExampleBot:
    def __init__(self, distribution):
        self.distribution = distribution
    # rest of the class methods

and maybe have one of the methods return what the bot guesses

My current code

users = [NumberBasher, YourUsername, BotAnother]
user_names = ["NumberBasher", "YourUsername", "BotAnother"]

points = list(range(1,11))
user_num = len(users)
bots = [user() for user in users]
scores = [0.0 for i in users]
information = {i:0 for i in points}

for i in range(100):
    choices = [bot.choose(information, points) for bot in bots]
    information = {score:choices.count(score) for score in points}
    scores = [scores[index] + choices[index]/information[choices[index]] for index in range(user_num)]

As you can see, I am a fan of list comprehension

Bye for now!

It's not finished, I still have the final report and testing to do

there might be some obvious errors because I wrote this on a random text editor that doesn't even have a mono-spaced fnt

*font

i assume choose is the method to guess a number ?

cant really make sense of the code without the specifications

maybe write up the sandbox post first

Yah ok

so basically bot.choose takes in two input. one is a dictionary of frequencies and the other is the range to choose from (1 to 10 inc.)

i think that should be taken as class attributes instead, but its ur choice

like

class Example:
    def __init__(self, distribution, points):
        self.distribution = distribution
        self.points = points
    def choose(self):
        # code here

then, in ur controller script, u do bots = [user(information, points) for user in users] and move that under the initialization of scores and information

u can then call bot.choose() directly in choices, like choices = [bot.choose() for bot in bots]

But I have to create a new bot every one distribution

very time consuming and i think the code will look bad...

@PyGamer0 Jelly, 9 bytes (without padding)

ye true, but then there would be no point in making a class in that case. i think a list of named functions would work just fine then

i found an online c++ compiler

^ can anyone tell what library is that textarea using?

maybe the player should have access to more info

like how many points they have currently

or what round number it is

i... actually am not sure if i understand how base64 padding since the input i'm testing on is something i'm specifically testing against semantle and wikipedia's saying padding brings the length of the encoded text to a multiple of 4 but then what the fuck is Z2V0aGVy=

or maybe how many points other bots have total (but dont show their names, only have a list of nums)

@UnrelatedString testing what against semantle 0_0

specifically the url of semantle.novalis.org/nearby_1k/Z2V0aGVy=

@UnrelatedString maybe equal sign not included ¯\_(ツ)_/¯

the = is the padding

oh idk :\

just a guess lol

maybe it dont do base64 padding whatever that is

@AidenChow 'll go with func

i will go with only these info

@UnrelatedString there was this one website I found which explained the process well

base64.guru/learn/base64-algorithm/encode

that seems to agree with my padding-free jelly sol

My current code, completed:

from random import choice as r
number_basher = your_username = bot_another = lambda x,y: r(y)

# Above is test functions !!!

users = [number_basher, your_username, bot_another]
user_names = ["Number Basher", "Your Username", "Bot Another"]

rounds = 3
points = list(range(1,11))
user_num = len(users)
bots = [user for user in users]
scores = [0.0 for i in users]
information = {i:0 for i in points}
print('\n'*100)

for round_number in range(rounds):
    choices = [bot(information, points) for bot in bots]

I will add the description later.

Yay!!! Past 400 rep!!!

hmm i feel like u could have more strategies if u gave the bots a little more info, like how many points they had, a list of everyones points (but no bot names), round number, etc. but its really ur choice

u only going to simulate 3 rounds at once ? i suggest at least 500 with how simple the game is

@NumberBasher nice!!

#define A(t) \
    typedef struct t##arr { \
        t *p; \
        size_t len; \
    } t##arr;

A(int)

^ templates in C ??

lmao

@AidenChow It was a test, I was going to do 500 but I just wanted to know there was no bug

@UnrelatedString is works lol

@AidenChow will do that

i didn't expect it not to

i need to remake the C++ std::vector in C like this lol

and frankly it's not even that bad an abuse of the preprocessor

i know i've defined a macro that for some reason starts with a close paren and ends with an open paren

think it was something to do with casts

#define TEMPLATE_VEC(t) \
    typedef struct t##arr { \
        t *p; \
        size_t len; \
    } t##arr;

#define VEC(t) t##arr *

#define NEW_VEC(t) (t##arr *) malloc(sizeof(t##arr))

TEMPLATE_VEC(int)

int main(void) {
    VEC(int) v = NEW_VEC(int);
}

^ going somewhere lol

> welcome to this episode of ab using the c preprocessor

for some semblance of cleanliness maybe #define VEC_TYPE(t) t##arr at the start so you can change how you construct the name without having to update it everywhere

and adding onto that i'd also recommend changing how you construct the name :P

maybe t##__vector or something like that

@PyGamer0 Quick question: Why

Oooooh

For the templating stuff

My b

is there a way to make a python code runnable (as a leaderboard) on the top or only javascript/a specific lang?

only js

why oh no

because js runs natively in browsers

its the web...

they should copy paste from aops or something

they should...

We could probably use wasm and Cpython..?

how?

thnx

or use pyodide...

probably somehow yeah

typedef struct int__vector { int *p; size_t len; } int__vector;

int main(void) {
    int__vector * v = (int__vector *) malloc(sizeof(int__vector));
}

^ current preprecessor output

the only thing i know about js is console.log

hahaha

@NumberBasher do you know any c-like language?

no

only python

oh..

and a bunch of esolangs

and a tiny bit of everything

so a little bit of js, c++, c, julia, blablabla

but to little to do pretty much anything

Learning the C family is probably a good idea long-term.

so i know how to write a for or a while I think something like for(){}

int x=101022918812978

how do i run some python code in js?

is there a way to run python in js

Theoretically?

...

is there a way i could just copy paste it

@NumberBasher github.com/pyodide/pyodide

Let's X Y a little here

@NumberBasher how do you except to do that?

Why do you want to use Python in JS

@ATaco its for a koth

b/c i want to embed runnable code into cgse

Probably not worth it. The amount of Overhead python would require is huge.

^ example of python in js

I've never seen the "suggested edit queue is full" message before

@cairdcoinheringaahing I finally figured out my confusion: all this time I thought that request you bountied was on the moderators' private Q&A "team," and I couldn't figure out on what site you were a moderator! Turns out, I don't recognize Meta.SE's color scheme very well....

gg

Speaking of mother meta, still no update on what's popping with Q1 and Q2 2022

hiiiiiiiih

@ATaco actually, since it runs in WebAssembly it can be surprisingly fast

@PyGamer0 nope

onioni?

@lyxal ???

well there's hasn't been a Q1/Q2 2022 roadmap yet, and apparently we're actually in Q1 2023

and i want to know what's SE plans to do with SE for this quarter

I knew Australia was ahead but wow

acquisition?!

when the hell did that happen

what rock have you been living under

you didn't hear about the network being sold for roughly a billion dollars?

also, I don't trust a company with "sus" in its name

@lyxal nope! I've been under my rock writing Wii Remote libraries for the past month or two

what effects is this going to have?

prosus.com/companies

they've got so many the images don't load

@Ginger well it happened in June last year, so not much clearly

oh

that explains it

they bought the network for $1.8 billion

I wasn't here last year

> Member for 6 months

as someone who's seen the last 1000 days of SE and code-golf I definitely saw it

I am slightly amused by the giant spike in my rep around the time I posted the flag challenge

Over like 5 days I gained ~700 rep

now, back to working on lambda.chat

Jesus Christ that's an expensive domain

I can get https://λ.chat for the low, low price of $40/year

erm

@Ginger why :(

@PyGamer0 no time

then i must make one ... :p

hmm, i just did a toefl that I suck at

and i got 66% for the reading and 100% for the listening

just wanna ask if you'd rate how many stars out of 5...

feedback?

can't believe no +2s

and no +4 wilds either

at least make 2s +2 and 3s +3

Maybe you want Uno?

I've played crazy eights with +2s and +3s

._.

/srs please

/srs crazy eights with +2s and +3s

and stackable too

non

what is srs

serious

um

@Ginger then that's my feedback done

i would prefer uno

I'm also tempted to suggest jump-in and draw-to-play

oh oh and 7-0 rules

@lyxal no

I would suggest uno is better than crazy eights...

I would too, but it's more complicated than crazy eights

speaking of uno,

CMQ: +4s on +2s?

yes

there are different versions, some say you have to draw 2 immediately and some say you can +4 and the next person after that +6

more cards, the better

@NumberBasher I mean stacking

@mathcat how about +2s on +4s and +4s on +4s

ooh nice

I personally think +2 on +2, +4 on +2 and +4 on +4 but not +2 on +4

I once got +4 on +4 on +2

mmm juicy 10 stack

my record was 52 i remember

one version is like you can't stack and the other is you can stack for all

@lyxal never heard of that, tho i think it is the most "reasonable"

(luckily the 52 wasn't to me

> A player may play a Wild Draw Four card only if that player has no cards matching the current color. The player may have cards of a different color matching the current number or symbol or a Wild card and still play the Wild Draw Four card.[5] A player who plays a Wild Draw Four may be challenged by the next player in sequence (see Penalties) to prove that their hand meets this condition.

i don't obey that law

we can play it at all times

> A player who plays their penultimate card must call "uno" as a warning to the other players.[6]

We don't do that now

@lyxal Wikipedia said this

> Progressive or Stacking Uno: If a draw card is played, and the following player has the same card, they can play that card and "stack" the penalty, which adds to the current penalty and passes it to the following player[5] (although a +4 cannot be stacked on a +2, or vice versa).[7] This house rule is so commonly used that there was widespread Twitter surprise in 2019 when Mattel stated that stacking was not part of the standard rules of Uno.[7]

I know

But Mattel is wrong

They made the cards, we make the rules

so +2 or +2 or +4 on +4 is allowed??? Hmm...

Alright which one of you turned pxeger blue and put him on my keyboard?

??? wdym

See? He's there on my keyboard

um... idk

It's a joke

does anyone know how to access function name in python given the function?

@lyxal i certainly don't get it

Pxeger's profile picture is a light bulb, and the icon on my keyboard is also a light bulb

how silly of me, name...

erm, __name__

hahaha

rSNBATWPL in Mathematica?

rosin batwaffle

what's rsnbatwpl?

@Romanp can't believe radvylf plagiarised mathematica smh

at least have the decency to plagiarise a lesser-known symbolic algebra language

@NumberBasher A new programming language. Kind of unusual syntax.

This is my code for the aforementioned koth

from random import choice

def random_better(information, points, round_number, sum_of_all):
    return choice(points)

def greedy_better(information, points, round_number, sum_of_all):
    return max(points)

def calculator(information, points, round_number, sum_of_all):
    return sorted(points, key=lambda x:x/(information[x]+0.000000000001))[0]

def smarty(information, points, round_number, sum_of_all):
    return sorted(points, key=lambda x:x/(information[x]+0.000000000001))[choice(range(1,40))]

@NumberBasher radvylf should not be allowed to write programming languages

Guess who wins, random_better, greedy_better, calculator, smarty

@lyxal why

no no that's the language name

oh

...

it's literally called "radvylf should not be allowed to write programming languages"

rSNBATWPL for short

It's like setting the wifi password to "I'm not telling you"

why not create programming language named plnp

Every function is an infinite loop

@lyxal it;s like me creating an esolang "language not created by sunny"

whoops

shit i used my real name

but then idc cuz i don't have an actual english name

(actual = legal

@lyxal pronounced /ˈɹɒ.zənbætwɒfl/

in my country there was a curse word that shortened to md which occurs in my mind every time I create a markdown file

well at least that works :p

@lyxal no +2s only on +2s and similarly for +4s

redvylf should really not be allowed to write programming languages...

the syntax is cursed

it's hella great for golf tho

lol what

pretty sure it's a parsing bug, because it just hangs

it also makes my tabs unclosable

maybe i should try writing a practical language

cmp: ^^

Thanks, please help

@NumberBasher what if one user decides to make 2 submissions with 2 different bots?

@PyGamer0 Do we even have an objective definition of a practical language?

Like what is needed fot something to be objectively considered practical instead of esoteric

practial is not esoteric :p

And when does esoteric cross over into the realm of practical

@PyGamer0 depends on your definition of practical and esoteric

For example, one might consider APL to be esoteric when it's actually practical.

hmm true

@lyxal and when the person considers that, we should nuke their fridge :p

i want to make my practical language in c (or maybe c++)...

/a comment!
/very cool comment
#define MACRO "eee"
class a_class {
  /...
}
fn function(x, y) {
  r = [];
  for (i = x; i < y; i++) {
    r.push(i);
  return r;
}
fn main() {
  print(MACRO);
}

@lyxal my idea ^

oh and btw divide is %

gl pygamer with memory management

dont have to worry about memory management when you dont have any memory to manage :P

how

its still an idea :p

i haven't started implementing it yet

also

@lyxal We don't. For example, somehow Dennis was persuaded to put Zephyr and Appleseed in the "practical" category on TIO, despite the fact that neither of them is even close to usable for any practical purpose.

Zephyr seems like a "practical language," just not a good one for doing practical things.

Yeah. It could be a practical language if I put a lot more work into it.

It has the readable, Englishy syntax of QBasic and the expressive power of... a small subset of QBasic. Oh, and fractions.

Any feedback on Irradiated Polyglots? My biggest worries are related to the scoring/difficulty level of the challenge, since I don't know whether 20% leeway is good enough and if the scoring penalties are structured correctly.

Unfortunately, it seems that due to replit's limits I cannot have message history in lambda.chat

what about like heroku and glitch? maybe those would work better?

@emanresuA it seems to me that in rSN, strings that cannot be parsed as an actual number are equal to -2 if parsed as an int, but NaN if parsed as a float

but a string being parsed as a float does not necessarily make the whole expression a float

"5.5" / 5 is 1, not 1.1

but "5.5" / 1.0 is 1.1

damn did youtube re-design their interface? it looks like video descriptions are now above the bar with the like button and whatnot

this feels really wrong

Looks the same to me, might be A/B testing

@emanresuA This is why I don't use IDEs :P

Finished AP World History test, last one of the year

Best part is I'm halfway through Chemistry so I can just skip it and have the absence counted as excused

Looking forward to tomorrow lol

(Context: Got somewhat informally accused of plagiarism because I didn't show the "references" slide at the end of my presentation since the audience obviously can't read 8pt font from halfway across a room)

@emanresuA Yep, casting a string to an integer returns -2 if it's not stringable

-1 is used for something else I think

why would it be -2 and not say,, Null

Null isn't an int

@nitsua60 If you'd like to clear up the confusion, I'm happy for you to add me to the mod team :P

@DLosc They've been A/B testing this for months

Sometimes I'll watch videos in an incognito tab to avoid spamming my recommrndationd with like, doorknob repair, and sometimes it's looked like that

Chrome OS changed the boot splash thing to be dark theme

I love it, since now I can boot up my laptop in the middle of the night without blinding myself

@Romanp I am Radvylf and I support this message

@lyxal I turned pxeger blue

(I have a userstyle that randomly hue shifts their pfp)

(Since they're the only non-identicon user I could find who's active in chat and has a solid-colored pfp)

(And I don't like orange anyway)

ok, if i understand correctly, c++'s virtual means "overridable", right?

@Seggan I think it's one of those C++ keywords which means one of far too many different things depending on context

Ah no, stackoverflow.com/a/2392656 explains it well

it essentially means use dynamic dispatch on this superclass when this method is overriden in a subclass, instead of static dispatch

mhm. basically "overridable"