Chat:World/2020-12-24

From CG community
Revision as of 11:57, 15 June 2021 by Chat Log (talk | contribs) (Created page with "<img src=/a/35448723594340> jacek: he can fit entire universe in bitboard File:Default_avatar.png REAL_TYPHOON: halo File:Default_avatar.png REAL_TYPHOON: I am real...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

jacek: he can fit entire universe in bitboard

Default avatar.png REAL_TYPHOON: halo

Default avatar.png REAL_TYPHOON: I am real typhoon yo

Default avatar.png REAL_TYPHOON: and you are

Default avatar.png REAL_TYPHOON: :rage:

Default avatar.png REAL_TYPHOON: :rage:

Default avatar.png REAL_TYPHOON: :point_right_tone3:

eulerscheZahl: stop spamming emojis

Default avatar.png REAL_TYPHOON: y6ou sayin' to real typhoon

eulerscheZahl: you risk a kick right now

Default avatar.png REAL_TYPHOON: do you have any idea who are saying to

Default avatar.png REAL_TYPHOON: i am an 11 year old kid

Default avatar.png REAL_TYPHOON: hi

Default avatar.png REAL_TYPHOON: bye

jacek: oO

Zorg1: yet another genious that will write high quality CoC puzzles ?

Or64bit: Hey guys

Or64bit: Can I ask a quick c++ question?

geppoz: no harm in try ;)

Or64bit: So this is only theory for now, but I'd like to make a class that collects other classes which can be uninitialized if the user needs them to be

Or64bit: Basically I'd make a scene class that would collect other objects of classes that need to be in a gamescene at once

Default avatar.png Piokemon: hello

BucurEva87: Where was the question in there, @Or64bit ?

eulerscheZahl: 404 not found

Skynse: I wonder what language CodinGame's backend was written in

Skynse: anyone know?

Megha.Akash: Just use What runs chrome extension and buitwith extension and check it out @Skynse

BucurEva87: I never tried that, but I highly doubt it works :))

aboqasem: HELLO

aboqasem: MY NAME IS SNIKE

Default avatar.png i33mr: HEY PALLLLLLLLLL

aboqasem: WASSSSSSSSSSSSSUP

aboqasem: :poultry_leg:

Default avatar.png i33mr: 7YOOOOOOOOOO ༼ つ ◕_◕ ༽つ

Astrobytes: bye

Default avatar.png MartinEduardo: Hello guys how are you?

artmyros: good

artmyros: im good

aamsii: merry christmas

TyLuu: me ri chịt mợt nha mấy ba mấy má

GamingWarrior: hi, I would like to join! (Also nice Chrismas eve)

MSmits: hi guys

struct: hi

MSmits: finished solver in the middle of the night, looking for other options now

MSmits: shared 2nd place basically, with jrke

Default avatar.png 7amasa: hello guys

Default avatar.png 7amasa: I am here to learn to be a software developer, what should I do now?

MSmits: mmh not sure if thats exactly what we do here

MSmits: we use algorithms to solve puzzles and play games

MSmits: not the same as writing an app, designing a database or maintaining a website

MSmits: you'll get faster at problem solving using code though

Default avatar.png 7amasa: Then what do you suggest I do if I want to find a job as a software developer :D?

ZarthaxX: pick a project, and you will learn while doing it

ZarthaxX: later you have that it show off in interviews

ZarthaxX: to*

Default avatar.png 7amasa: thats cool thanks for the suggestion, where should I pick a project?

ZarthaxX: that's smth you have to come up with

ZarthaxX: think of smth you would like to make

Default avatar.png 7amasa: aw makes sense :D

Default avatar.png 7amasa: Alright will start with that right away, thanks again

ZarthaxX: like a web app, or a desktop app

ZarthaxX: etc

ZarthaxX: you will start googling how to make your project and you will learn lots of stuff

ZarthaxX: better than just learning X thing without a purpose

Default avatar.png 7amasa: That makes perfect sense, thank you for your help bro

ZarthaxX: np :)

Default avatar.png zhoubou: I'm doing the puzzle of the week (2048). I'm starting by evaluating how the board looks after a move (Up, down, left or right). How do I ensure I don't merge the same tile twice?

MSmits: maybe check the referee?

MSmits: see how it's coded

MSmits: obviously the referee needs to check this as well

MSmits: (I mean the github link)

MSmits: usually i dont need to, the description is clear enough in most cases

Default avatar.png zhoubou: Where can I find this link?

MSmits: in the description I guess? Thats where it is usually add

MSmits: at

MSmits: scroll through it

Default avatar.png zhoubou: Ahh, I see it. It's hidden in the text for "Expert rules". Thanks!

MSmits: np

Gorter: Dumb question maybe, but why is there no factor for shortest in clash of code. Code size really depends on the programming language used...

Jasperr: Why would there be

Jasperr: Imagine winning shortest with 200 lines of Java vs 50 lines of Ruby

Jasperr: Kinda weird right?

Gorter: Yes and no

Gorter: I think that data type definition should not be counted as characters at least.

Default avatar.png zhoubou: I think it's unsolvable that way. Where do you stop with rules like that? Unless it was made so only people with the same programming language can join...

Default avatar.png zhoubou: Or if it was mandatory to choose a certain language

Jasperr: I disagree (with the arbitrary ignore-rules), shortest is as much a challenge of solving the problem as it is of picking the right language and figuring out nifty language-specific tricks

Gorter: Yea you are probably right. Wouldn't mind something like a ranked mode, where there are only 1 or few options. Which varies..

Jasperr: A language-specific clash mode would be fun, but would also be pretty much impossible to fill with non-bots as everyone has their own preferred language

Gorter: Depends on the language I think :P

Jasperr: Haha true

Gorter: Python would be popular

Jasperr: Most people seem to use Python or JS in non-shortest and Ruby in shortest

MSmits: people also use python in shortest because they just dont know ruby

Default avatar.png zhoubou: Anyone did 2048 puzzle recently? I'm having problems with it not "listening" to my output.

jrke: i last saw you when you were going to sleep MS now i am going to sleep lol gn

Astrobytes: gn jrke

MSmits: hey

MSmits: gn jrke

jrke: b-bye

jrke: btw before sleeping i wanna say that yes MS you play first 50-55 moves good to win in D&B

struct: MSmits SupSon asked if you used bitboards on D&B

jrke: have to play *

k4ng0u: zhoubou what's your problem exactly? I basically copy pasted (with small translation from java to c++) the referee and didn't not encounter major issues

MSmits: oh

MSmits: yeah i do use bitboards on D&B

MSmits: but i convert them into a graph for the solver, so its only temporary

struct: oh well, I told him you used them

Default avatar.png zhoubou: k4ng0u I'm sorry. My mistake. "Right" in my language starts with "D". That's where the confusion comes from.

struct: and linked him chat log

MSmits: well i do

MSmits: two uint64_t, one for hor and one for vertical

MSmits: then convert to a different system with boxes instead of lines

MSmits: a single uint64_t for that. Thats for the graph

Astrobytes: makes sense

k4ng0u: zhoubou by language you mean your code right? not sure how the language can influence on what "starts" by what => it seems to be very implementation related

Default avatar.png zhoubou: My native natural language. English is not my native language.

k4ng0u: oh ok ok I see

Default avatar.png zhoubou: Thanks for chiming in :)

Astrobytes: desni/desno

Astrobytes: ?

Default avatar.png zhoubou: Yes ^

Astrobytes: just googled it :)

Salted: Hello World!

Astrobytes: hello

MSmits: hi

Default avatar.png zhoubou: Astrobytes nice googling skills there

Astrobytes: thanks, I was hoping I'd chosen the correct ones :)

Default avatar.png zhoubou: Haha :P

TobiasA: in this defibrijksjdfjsjftor thing i done everything correctly but it keeps choosing the wrong one

Mohamed_Kharrat: guys where do we find puzzle of the week ?

eulerscheZahl: https://www.codingame.com/multiplayer/optimization/2048

eulerscheZahl: while the default code is enough to get 100%, that's not what you are supposed to do here

eulerscheZahl: improve the algo to get a higher score, as this puzzle has a leaderboard

Mohamed_Kharrat: thank you :D

jacek: or maybe this puzzle should reject invalid moves

eulerscheZahl: i allow them to avoid being too mean (users predicting the next spawn wrong)

eulerscheZahl: if not for the possibility to print multiple actions in a single frame, it would be different

eulerscheZahl: sneaking around CG limitations and trying to minimize the unwanted side effects

icecream17: I love this hack: ["2", "3"].map(Number)

BrunoFelthes: just dont do it

icecream17: generally it would be ["2", "3"].map(a=>+a)

BrunoFelthes: it is just a bad practice, because the map will pass more arguments.. try ["2", "3", "4"].map(parseInt)

BrunoFelthes: ["2", "3", "4"].map(n => parseInt(n)) works and ["2", "3", "4"].map(parseInt) dont works

icecream17: I used Number instead but your point about bad practice is valid

jacek: is code gold ever good practice? :thinking:

icecream17: no i'm using it in regular code b/c it looks better than nested for loop for input

Default avatar.png NeoXX: hi

Default avatar.png NeoXX: anyone have good resources to learn c++ ??

Default avatar.png Wings501: http://cplusplus.com/doc/tutorial/

VizGhar: yup cplusplus rules

VizGhar: any recommendations for scoring function fo 2048 puzzle?

icecream17: 2048 strategy is to put the biggest one in the corner i think

VizGhar: this is insufficient :) In puzzle on this site we have much more knowledge to base our evaluation on

icecream17: https://puzzling.stackexchange.com/a/2753/51442 can help maybe?

VizGhar: Hmm might help...

icecream17: you can play an action that doesn't move anything (invalid but doesn't make you lose), to try to manipulate the rng, but that's only possible if there's an invalid action in the first place

struct: These were linked here before, I never read them though

struct: https://www.researchgate.net/publication/311555004_Systematic_Selection_of_N-Tuple_Networks_for_2048

struct: https://www.researchgate.net/publication/327912401_Comparison_of_Expectimax_and_Monte_Carlo_algorithms_in_Solving_the_online_2048_game

VizGhar: :thumbsup:

Astrobytes: with a common them of a 'snake' of increasing numbers culminating in the largest in a corner

Astrobytes: *theme

VizGhar: Seems like best evaluation from all articles I read so far as well...

Astrobytes: Also, you can get to ~77th with sheer laziness and a hardcoded repeating sequence of moves :P

Astrobytes: (I haven't written a proper bot yet)

Astrobytes: I think the n-tuples approach shows a lot of promise.

VizGhar: i'm 119 with ulululul / dlululul :D

VizGhar: a lot of reading tho

Astrobytes: do you print those separately (by letter) or just as is?

Astrobytes: And yes.

VizGhar: when {

           board == currentBoard -> println("DULULU")
           else -> println("LULULU")
       }

VizGhar: Reading the article, and I don't like this part:

Each of them requires a dedicated LEARN EVALUATION algorithm, which adjusts the evaluation function on the basis of the observed experience represented by a tuple (s, a, r, s0, s00).

VizGhar: I'll go simpler most probably

Astrobytes: jacek's your source for n-tuples stuff, he has some recent experience with them on here

Astrobytes: *go-to source

VizGhar: I'm missing too much theory as well. not even sure what n-tupes means :P

Astrobytes: just a tuple of n-size

VizGhar: 10 years writing android apps. I didn't need this stuff

icecream17: i think a tuple is like a list or array

Astrobytes: representing positions (patterns) on the board

icecream17: oh

Astrobytes: but with weights assigned to them

icecream17: so an n-dimensional board where each cell has a value?

Astrobytes: hey, don't worry, I have trouble with some theory at times too. I'm not even a developer

Astrobytes: not quite icecream17, just patterns from the board

icecream17: ah. i see that i've never seen this before. hmm

icecream17: i'll search it up

VizGhar: We have a lot of work to do icecream17 :P

Astrobytes: It's tricky to get your head around, especially from the papers

VizGhar: And what you do Astrobytes if I can ask?

Astrobytes: It's worth asking jacek when you see him online

icecream17: https://www.google.com/search?q=pattern+programming https://www.google.com/search?q=board+pattern+programming Is this close

Astrobytes: Currently I do pharmaceutical regulatory affairs, previously (pre COVID and a back problem) I was also assistant manager at my local wildlife conservation centre (I'm a biologist)

Default avatar.png JBM: cool story bro

Astrobytes: oioi JBM

Default avatar.png JBM: ahoy there

VizGhar: Wow, quite impressive :) I can't imagine how you "end up" here

VizGhar: I know some biologists, pharmacists myself... none of them ever touched code

Astrobytes: I've been into it as a hobby for a long time, I've done a few courses here and there, and I was once a junior webdev (2000/2001)

Default avatar.png JBM: leave the code alone you perv

Astrobytes: heh heh

Astrobytes: how's it going JBM

Default avatar.png JBM: more and more alcohol as the evening goes

Default avatar.png JBM: you?

Astrobytes: Working towards the same goal

Default avatar.png JBM: yeah xmas w/o family is weird

VizGhar: :D :D man of culture JBM?

VizGhar: ah I see

Default avatar.png JBM: culture... cool word, so many interpretations

icecream17: I think maybe I found an error in the second reasearchgate source?

Default avatar.png JBM: whew

Astrobytes: Family or not, most festive times I try to maximise my alcohol consumption as evenings progress even if simply to deal with family members

Default avatar.png JBM: for a sec i thought you found a bug in temperatures

Astrobytes: :D

Astrobytes: Everyone knows it's bugged already. No news there.

Default avatar.png JBM: hey, old classics

Default avatar.png JBM: yeah right

Astrobytes: The latest bug is that 2048 just accepts random admissible letters and you get 100% :o

Astrobytes: *latest 'bug'

VizGhar: Yup... Original 2048 is solved after you get 2048... This should have been minimal requirement. But it's optimization puzzle, so who cares

Astrobytes: ah, no, I was just being facetious to be honest, sorry :)

icecream17: interesting, the paper says that there's an 80% chance to get a 2, but in here there's a 50% chance

icecream17: wait no

Astrobytes: Also remember that in this version (CG version) you're given the RNG and the seed, so you know the location of the next spawn

VizGhar: Hmm it is 50% isn't it? just checking second bit of seed

icecream17: oooohh i was acutally correct. i got confused about the bitwise & operator

VizGhar: But i'm not sure whether that <seed * seed%something> isn't favoring some values more

VizGhar: not important

Astrobytes: ask euler when you see him next about it, it's a conversation

VizGhar: have you seen his 3d puzzle? I fear we will see mars lander 3D in a year :D

Astrobytes: Yes, Blockout, looking forward to trying it out! Hahaha, now that would be something

VizGhar: I'll most probably transfer to contributor rather than problem solver. I loved my first experience with SDK. Hope I can finish the game successfuly

Astrobytes: I look forward to trying it out soon!

struct: too many games

struct: so* many

Astrobytes: Hey, you still have to have a crack at some games once in a while too

VizGhar: You can try it right away :P https://www.codingame.com/contribute/view/59519a876ee17f2efd6a5e0813e7a8fdd0c9

Astrobytes: I know, I know. I mean I don't have the time right now :D

VizGhar: Know that feel bro

Astrobytes: Anyway, time for some brandy and maybe a few pálinkas

Astrobytes: goodnight all

VizGhar: :wave:

struct: gn

Ortieez: I finally finished my first clash of code :D I am so happy I finished it and I also got score of 100%

Default avatar.png Mirras: nj man

shilezhi: :grinning:

Default avatar.png CRH17: i want to learn code but dont even know where to start