Chat:World/2022-05-28
LelouchVC2: hey weirdos
pressplayontape: excuse me, how many minutes can solve this puzzle ?
5DN1L: time can heal a wound, time can solve a puzzle
5DN1L: i can't even peek aty your screen to see what "this puzzle" is
5DN1L: at*
jacek: happy Caturday
gaoguai: 这个怎么搞
gaoguai: 不会
gaoguai: 有没有QQ扩列的?有的加462619243
gaoguai: 咱们一起探讨
5DN1L: no chinese here
5DN1L: use #cn channel
jacek: or #openingbooks
RivalRickson: hi
5DN1L: Curiously I haven't seen any Japanese or Korean typed here before
5DN1L: #jp channel is in use
5DN1L: #kr channel does not seem to be used
jacek: we need to be more weeb
Prince6502: https://escape.codingame.com/game-session/VRj-968-mUu-4p7
5DN1L: You still aren't able to find someone to play escape with you?
Prince6502: ya you can say that!
Ayza: day 2 puzzle search site down
PatrickMcGinnisII: Ayza Euler ran out of dinos i think
5DN1L: yeah he mentioned that
5DN1L: wait for a few days then the site should be up again
jacek: apparently dice duel coins arent enough
eulerscheZahl: hey. the puzzle search will be up again on the first of the next month
waterproofsodium: :taco: jacek
eulerscheZahl: or when i pay $7
waterproofsodium: errr
Ayza: so 2 days, great
TINOUAINANI: Goood morning
eulerscheZahl: i never expected that much interest in my puzzle search tbh. 550h per month is a lot (that's the free dynos i get, heroku shuts down the page after 30min of inactivity)
eulerscheZahl: upon checking my logs: a lot of requests for robots.txt, so it must be web crawlers causing most of the traffic
eulerscheZahl: SEO optimization disallow: *
PatrickMcGinnisII: robot nofollow
PatrickMcGinnisII: they'll still do it tho
jacek: those robots :unamused: right, Automaton2000?
Automaton2000: i get the same output
WillChang: Automaton2000?
Automaton2000: this is why you can have a look at the board
WillChang: I think so too
WillChang: Automaton2000?
Automaton2000: at least not in the top 5
WillChang: that makes total sense, eh?
WillChang: Automaton2000?
PatrickMcGinnisII: Automaton2000 NN Clash of Bots for me, k?
Automaton2000: that is just a way to deal with it the same way
Automaton2000: i only do it for the same thing
jacek: woah stockfish has some nn changes https://github.com/glinscott/nnue-pytorch/blob/master/docs/nnue.md#historical-stockfish-evaluation-network-architectures
PatrickMcGinnisII: Automaton2000 throw a MSE at jacek for me
Automaton2000: so i have to do it
PatrickMcGinnisII: Automaton2000square pants
Automaton2000: is there some kind of search do you use?
**PatrickMcGinnisII looks under the sea for Automaton2000
Automaton2000: i think it makes sense
waterproofsodium: under the sea hmm
waterproofsodium: I think you don't even frogs to chew on there
waterproofsodium: *find frogs
jacek: or toads
waterproofsodium: croak
waterproofsodium: (doom monsters make that sound)
Ylan-Cipher: hello world
Ylan-Cipher: print me!!
jacek: do you even print, Automaton2000?
Automaton2000: can you do it with a simple eval
China-xyh: https://escape.codingame.com/game-session/LkB-gaw-6Aw-kkr
zakacat: :taco:
hasna_h: does anyone have more beginner friendly coding games like this?
hasna_h: i think this is too hard for me still
5DN1L: try codewars
jacek: have you tried power of thor or temperatures?
hasna_h: no i havent tried many yet, im still exploring
Ayza: there are a couple of beginner friendly puzzles here
hasna_h: i tried clash of code and couldnt complete any, which is the next stage after the intro:disappointed_relieved:
Ayza: did you try classic puzzles?
Ayza: https://www.codingame.com/training/easy
hasna_h: no not yet, i might try those then
Ayza: they are harder than clash of code, but some are fairly easy
Ayza: np
zakacat: and there is no time limit, and Google is your friend
zakacat: I dunno about you guys, but some of the puzzles take me 2 or 3 days to do as I need to think about it quite a bit sometimes
jacek: DomiKo how come i have the same games against you and magmasa? :thinking:
DomiKo: our model is kinda dog s....
jacek: and converged to the same point?
jacek: :upside_down:
struct: how many people are working on it domiko?
DomiKo: 2
DomiKo: me and him
DomiKo: same team as in CSB :D
DomiKo: jacek coverage to something, not really good thing
struct: did you manage to run code from py3?
struct: I tried but I get timeouts :/
DomiKo: didn't really tried it
struct: and now I just found the asm I get for a simple function is very different with -O3 and without
struct: https://godbolt.org/z/Gd9P1Kb5x
DomiKo: yeah, apparently pragma can't do a lot of things
jacek: theyre not so pragmatic eh
Westicles: is there something about C# HashSet that would make it faster than C++ unordered_set?
j4at: unordered_set is randomized
j4at: so that's why maybe its slower
jacek: wouldnt that make it faster?
Westicles: hmmm, supposed both are O(1) for find/contains
jacek: maybe its O(5) for C# and O(100) for c++
bbies: what's shaking everybody
jacek: Westicles and you sure youre measuring only set's things, not other things?
Westicles: Not sure, I was just doing this contrib and my c++ times out half the time on the last on and his c# doesn't
Westicles: https://www.codingame.com/contribute/view/7160061d215bfbde953f2fbd49d4dfa8b123
MaxMustermann123: http://chat.codingame.com/pastebin/739b3659-7f99-4394-8a80-74f86803598b
jacek: with pragmas?
Westicles: pragmas didn't make any difference
Westicles: I had to make my own hash, maybe something wrong with that
jacek: MaxMustermann123 to check if there is next line. without the if, in.nextLine() could wait forever if there is really no next line
MaxMustermann123: @jacek thanks, make sence
j4at: jacek "wouldnt that make it faster?" No. randomization stops attacks. If you have very random inputs then randomization doesn't make it faster
KrzysztofRozenberger: @jacek pozdrawiam
galuso: i am given a string where there's only one letter that doesn't follow the alphatical order. how do I inf it?
struct: is it just in a random spot?
galuso: i'm giving u one example
struct: ah
struct: just do a for loop
struct: if ('A' + i != str[i])
struct: you have the letter
struct: or doesnt it have to start at A?
galuso: don't have the precidse letter
galuso: it could be in every position
galuso: it cahnge the position ofone char with the following one
struct: check next letter with the current one
Westicles: just sort it and look for the first difference with unsorted
struct: if difference is different than 1
struct: that or im not understanding the problem
struct: ah the string is unsorted
galuso: it was like: You r given a string with letters in alphabetical order except one, find that one and count that
struct: what westicle s said should work
StevenV: hello guys
struct: hi
jacek: good evening
Tengy28: hi some0ne can help me to the MAD POD RACING plz
5DN1L: Please explain what help you need
Tengy28: i have to win the race with 2 new variables : next_checkpoint_dist, next_checkpoint_angle but i don't understand how can i do. i'm coding in python
Uper-Ed: Never gonna give you up never gonna let you down
5DN1L: I'm gonna kick you out
jacek: Tengy28 is this the one with checkpointY and checkpointY?
5DN1L: no Wood 2
jacek: so angle < -90 or angle > 90?
5DN1L: yup
jacek: it says that in the comment what you should do
Tengy28: i try with if else but it doesnt work
regismeysso22: -1
5DN1L: What's the code you've tried? Tengy28
Westicles: hey regismeysso22. love your work
regismeysso22: thx
Tengy28: if next_checkpoint_angle > 90 or next_checkpoint_angle < - 90 :
next_checkpoint_y = "0"
else :
next_checkpoint_y = "100"
jacek: umm
jacek: thrust = "0"
5DN1L: it should be thrust instead of next_checkpoint_y
5DN1L: read the pseudocode again
Tengy28: but thrust is not in print()
Tengy28: print(str(next_checkpoint_x) + " " + str(next_checkpoint_y) + " 100")
5DN1L: 100 is the default thrust
5DN1L: you should change that to a variable too
jacek: print(str(next_checkpoint_x) + " " + str(next_checkpoint_y) + " "+thrust)
Tengy28: i'll try thanks sorry :(
5DN1L: No need to say sorry :)
Uper-Ed: ...
Tengy28: http://chat.codingame.com/pastebin/678f9a3c-3104-4508-b6d5-9b1f126d504d
jacek: str(thrust) maybe
Uper-Ed: ...
Tengy28: doesn't work may i did a mistake
Uper-Ed: ...
5DN1L: stop it, Uper-Ed
Uper-Ed: Okay
Uper-Ed: sorry
jacek: Tengy28 either do str(thrust) in print or thrust = "0" ... = "100"
5DN1L: because you can't add an integer to a string, Tengy28
Tengy28: ok so i have to take of the var next_checkpoint_x,y and replace by thurst
jacek: huh?
5DN1L: your code pasted above is fine, except that you have to convert integer to string first
5DN1L: that's what jacek and i are trying to tell you
Tengy28: ah ok i'm going to tcheck that thank you
Uper-Ed: Hmm?
Uper-Ed: What
Ajaiy: Automaton2000 can you just quickly tell some object?
Automaton2000: one thing i need to have more tacos first
Ajaiy: Thanks
4nthracene: https://www.codingame.com/clashofcode/clash/242838884ea03a5f98f50330888899aff5e95ad
4nthracene: join!!
5DN1L: please use #clash channel for clash invites, thanks. 4nthracene
Ajaiy: Here you go Automaton2000, A taco profile picture. I hope you now have more tacos to see
Automaton2000: but it is quite a bit
SUwUn: I really wish we could make public code clashes of a specific type
SUwUn: I really like Shortest mode, but it basically never shows up
jacek: and restrict to java :imp:
Ajaiy: Wait I have an idea Someone tell a random (proper) thing and I will try to make it my pfp
jacek: Automaton2000 what would that be?
Automaton2000: you need to track the opponent
Ajaiy: jacek any idea?
Ajaiy: btw I asked Automaton2000 and he said tacos
jacek: rick roll
Automaton2000: you can also do a lot of variance
Ajaiy: oof
Ajaiy: i regret my decisions now
Ajaiy: ill try
jacek: oh my
Motseki: can anyone help me with the test??
jacek: what test
Motseki: I have been preparing for the test from a employer, so I kind have a imposter syndrome
jacek: as do we all
Motseki: so can you help if i need help??
jacek: i have never done employer test here
Motseki: its a test from a certain company, they used codegame to set the test
mmahir: im 26 and now started learning
mmahir: am i too old for this?
jacek: if you have question you can ask here. someone who knows will probably answer
1C0D3R: nah 26 isn't too old you got alot of catching up to do though
1C0D3R: becease its really competitive
1C0D3R: and most poeple on here i geuss are around 12 - 35
antiwonto: [auto] 'test' was defined as ' this is a test '
Wontonimo: mmahir , 26 is a fine time to learn. I've been helping my neighbour who is 41
Wontonimo: he started a year ago and now has a job programming
Wontonimo: what level of math do you have mmahir ?
Wontonimo: :wave: jacek :grin: How ya doing today?
jacek: i think i got this loss thing
jacek: derive of (x - y)^2 is 2*(x - y) but for some reason 2 is omitted ~
whomItMayConcern: i feel like the clash is hard. im not quite there yet i guess
Wontonimo: okay, let's say you keep the 2 in the loss function ... and then you have your SGD learning rate of 0.1
Wontonimo: how is that different than skipping the 2 and having a SGC learning rate of 0.2
Wontonimo: i think some times things like this are factored out in some work without really telling people why and it makes it confusing
Wontonimo: or, i'm wrong. it's happened before
Wontonimo: why not use tensorflow jacek? or is the point to do it yourself?
Wontonimo: because you could do training in python with tensorflow then prediction using c++/c# on codingame
Wontonimo: gotta go, would like to catch up later jacek. feel free to DM me on discord
struct: cya
j4at: http://chat.codingame.com/pastebin/eb886296-7a37-4ee5-9096-f989fe233012
j4at: Wontonimo jacek ^
antiwonto: [auto] Sorry but you can only award tacos once per day
Westicles: :taco: BlaiseEbuth
antiwonto: [auto] Spread the love. Give tacos to someone else, not someone you spam with tacos.
DomiKo: taco DOMINER
DomiKo: :taco: DOMINER
antiwonto: [auto] DomiKo has awarded DOMINER 4 tacos. DOMINER now has 13 taco. DomiKo now has 5 taco
struct: :taco: DOMINER
antiwonto: [auto] Sorry but you can only award tacos once per day
struct: o.o
Westicles: I won user of the month!
Westicles: https://imgur.com/a/AsBCNuF
j4at: :taco: j4at
antiwonto: [auto] You used your taco giving ability for the day to discover that you cannot give tacos to yourself
j4at: lol
j4at: :taco: DOMINER
antiwonto: [auto] Sorry but you can only award tacos once per day
j4at: well
Westicles: :taco: b0n5a1
antiwonto: [auto] Westicles has awarded b0n5a1 1 tacos. b0n5a1 now has 1 taco. Westicles now has 66 taco
b0n5a1: :nauseated_face:
b0n5a1: :D
Westicles: Somebody approve this so we'll have something to do
Westicles: https://www.codingame.com/contribute/view/1859464e025e798f85c72951e681f883e0c4b
TobiasA: It'll be nice if there were more machine learning/data science puzzles like that one
b0n5a1: Thanks for the giggle.
struct: okay I solved it
RedKittenOfDeath_c53e: Do y'all ever run into an issue where standard out is completely ignored
RedKittenOfDeath_c53e: Getting this warning but don't know why: "Warning: your code did not read all available input before printing an instruction, your outputs will not be synchronized with the game's turns and unexpected behaviour may occur."
Things work fine at first, then desync ad the puzzle starts ignoring stdout while stderr outputs as expected
Westicles: well done struct
struct: that happens when you dont read all inputs
struct: or output more than one time per turn
struct: Westicles dont you think its too easy though?
Westicles: nah, the kids need easier easy puzzles
RedKittenOfDeath_c53e: am I supposed to be solving a turn synchronization problem on top of the puzzle itself? kek
RedKittenOfDeath_c53e: ok I guess... the reason I havent read all inputs is cause the solution doesnt need them all
RedKittenOfDeath_c53e: but if that causes the thing to desync then idk guess I'll try better
papjim: if your code runs as intended, then you may leave it as is
papjim: if you are code golfing for example
papjim: if you dont care about that, the just read everything and use what you need
RedKittenOfDeath_c53e: What's odd is that everything is read in Main, I left it as it is
RedKittenOfDeath_c53e: so I really dont get what's the issue
papjim: I presume you're trying a bot programming puzzle so just read the description for how many times you can output for every iteration
RedKittenOfDeath_c53e: kk ill try that
papjim: you may be trying to output multiples times inside a loop, not the infinite loop of the puzzle, while it expects only one
RedKittenOfDeath_c53e: oh perhaps
RedKittenOfDeath_c53e: let me check
papjim: if you run a for each on an array let's say and print every time a condition is met
papjim: but you want only one value
RedKittenOfDeath_c53e: That was it^^^
RedKittenOfDeath_c53e: thanks :)
RedKittenOfDeath_c53e: Made sure there was only one output per game loop and passed all tests
DomiKo: any resources about compressing here?
struct: compressing NN DomiKo?
DomiKo: yeah
jacek: other than unicode?
struct: today I tried agade tool
struct: but not sure if it applies for Nns
DomiKo: I remember people using Unicode, but I can't figure it out
struct: https://github.com/Agade09/CG-Send-Binary
DomiKo: Yeah, I remember this, but I would like to encode only weights
DomiKo: Right now, I'm using base64
jacek: http://chat.codingame.com/pastebin/381bb129-fd83-49ec-9bec-ea48e6fcc742
jacek: a msg i sent to agad once
DomiKo: Thank you
DomiKo: interesting method, with that rounding
AngelicCapy: Hey
struct: DomiKo maybe you can try to message robo
jacek: marchete article doesnt have unicode tutorial?
DomiKo: nope
DomiKo: I mean, yes... He is usings his ENCODER16k.cs
JeeNeeUhs: hey
DraH: ???
JPK3lly: Hello discord
DraH: hello
Foxboy: HXH IS BACK
Foxboy: :DDDD
DomiKo: jacek compression works thanks :D
SuperKritical: HxH, been waiting for new chapters for like 5 years :D
waterproofsodium: 3 times place 1 in a row. then shortest code => 7/8 ..
antiwonto: [auto] Hey waterproofsodium, here is a :taco: for loggin in today while it is quiet. You now have 35 tacos
waterproofsodium: at least I get my cheer up taco
chaosabordine: hello community
chaosabordine: i have an hist
chaosabordine: history question sry
chaosabordine: coding legend history that is. any buffs on that subject?
Husoski: I was learning to code when Fortran IV and COBOL 68 were still widely used. Does that count as history?
chaosabordine: hmmm I'd say so :)
Husoski: What question did you have in mind?
chaosabordine: I'm looking for the name of a man that worked in Philips or Bell in the 70s i think. He coded a game in assembly and quit when he was asked to let his game loose. No one but him really understood his code
chaosabordine: its pretty vague im sorry...
Husoski: That doesn't ring a bell (no pun intended).
chaosabordine: ya I know its far stretched...
chaosabordine: not intended but nice notheless
Husoski: Lot's of people still coded in assembly in those days. My first full-time job was doing IBM mainframe applications in assembler. (In IBM-land, it's "assembler" or "ALC", not "assembly".)
chaosabordine: I chated with a guy who met THE Mandelbrot at IBM this morning...
chaosabordine: I'm nowhere near using proper terms lol, very approximative in that respect..
Husoski: Cool! Mandelbrot was known more for math than programming, but that's pretty interesting all by itself.
DaNinja: Ken Thomson lost his game machine http://people.uncw.edu/tompkinsj/112/texnh/handouts/unix.html
chaosabordine: sure is, somehow I wouldnt have expected someone contemporary to a fractal set father where I found them this morning...
Husoski: Thomson, wasn't he the guy who invented Unix?
DaNinja: yep
chaosabordine: Yes, and B, and some of C?
Husoski: That game was probably "Space War". I played a version of that on an IBM 1620 in the basement of one of the UC Berkeley physics buildings around 1974 or so. A friend and former classmate I was visiting up there showed it to me.
Husoski: The code wouldn't have been Thomson's, since the 1620 looked nothing like any of the DEC PDP machines.
chaosabordine: Not sure that's the one Im looking for. His game was meant to be exposed at some expo but it was too hard to win against it and so it wouldnt have sold their supercomp enough coz clients were fustrated
chaosabordine: thank you for your help, ill look into the shared info!!
Husoski: c/probably/possibly/
Husoski: ...but I guess "probably" is still okay for small enough values of the probability. :)
chaosabordine: you guys are on another level.. i tinker really...
Husoski: Nope,. "Spacewar!" was done by Bob Saunders and other at MIT. But it was originally on a PDP-1 according to WIkipedia.
Husoski: Make that "Steve Russell", along with Saunders and others.
Husoski: Tinkering is good. It's how I got started down the path of failure as a physics student and happiness coding!
DaNinja: Thomson made Space Travel
Husoski: Space War wasn't easy to play either, btw. More because the human input was done through four toggle switches ("sense switches") on the front panel. The hardware hackers at UCB added the oscilloscope display. That wasn't stock equipment on the 1620.
chaosabordine: Ill look into all these names thank you so much
waterproofsodium: gm
antiwonto: [auto] Hey waterproofsodium, here is a :taco: for loggin in today while it is quiet. You now have 36 tacos