eulerscheZahl: hi PatrickMcGinnisII
my bulls and cows is just enumerating all possible numbers. then printing on of the candidates and reducing the list based on the response
eulerscheZahl: that alone will give you a score around 320. improve the printed number a bit and you get 310. 304 on a lucky submit
PatrickMcGinnisII: mine did ok until length>4, then i ran into issues
eulerscheZahl: i had to convert from C# to C++
eulerscheZahl: it's a bruteforce game and you PHP is at clear disadvantage
PatrickMcGinnisII: well there are 3265920 combinations
eulerscheZahl: and you have 1s to generate them
PatrickMcGinnisII: So, I'm trying to random generate kinda like a GA but with a simplified fitness function
PatrickMcGinnisII: it's not optimal solution, but I shouldn't max out of guesses and thus get 100%
eulerscheZahl: if you just care about 100%, there are heuristics too
eulerscheZahl: set a fixed first digit, rotate the remainder
eulerscheZahl: 12345 -> 13452 -> 14523 -> 15234
PatrickMcGinnisII: tx for feedback, yea bruteforce didn't work
eulerscheZahl: count the number of bulls
PatrickMcGinnisII: I'm working on improving the history eval right now
eulerscheZahl: ok, good luck :)
PatrickMcGinnisII: tx
PatrickMcGinnisII: yup, guaranteeing yourself some early game history is pretty valuable
PatrickMcGinnisII: hmmm
itachi0718: lets go bitches
itachi0718: imagine a world without womens
eulerscheZahl: the only one who has to go is you, you've been warned
proogrmr: yup
jacek: woo new puzzle of the week
Stilgart: is there a simple way to find which puzzle is puzzle of the week ?
proogrmr: yeah me too searching that
Stilgart: it used to be on the "practice" page, but it is no more
PatrickMcGinnisII: just look down your notifications... you may have too many notifications if high level
Stilgart: PatrickMcGinnisII: I can't find it actually
PatrickMcGinnisII: https://www.codingame.com/training/easy/next-growing-number
Stilgart: too many coc notifications :/
Stilgart: thanks
Stilgart: BlaiseEbuth is telling you that the puzzle of the week can be found by clicking on the "Puzzle of the week" green button within the quest map
BlaiseEbuth: Yeah. I'm telling that. Hello world ! o/
Stilgart: thanks
Stilgart: hum... shit is happening
Stilgart: quelqu'un a résolu un puzzle fantome :(
Stilgart: en même temps, aucun nouveau ne l'a tenté ce puzzle... j'en déduis que même si un puzzle est supprimé, on peut toujours le résoudre si on a l'url directe vers l'IDE
jrke: i just discovered there are 55296 ascii characters lol
jrke: 55295*
Stilgart: woops, why I am on #world on this tab :D
Stilgart: jrke: unfortunately, there are only 128 :(
TheBarnacle: Pretty sure I'm doing a big dumb, the next-gorwing-number weekly challenge suggests there is a way to increment a number while it is a string, how do I do this in C++?
TheBarnacle: Google was unhelpful
SPDene: convert to int, increment, convert back to string?
Stilgart: TheBarnacle: what you need to do is not incrementing
TheBarnacle: The problem says that conversion isn't necessary
TheBarnacle: I'll have a think :) Thanks
Stilgart: SPDene: that's what people may want to do... fatal mistake :D
Stilgart: TheBarnacle: it is even quite a bad idea
jrke: Stilgart run this python code -
for i in range(55296):
print(f"{chr(i)} : ASCII = i")
Stilgart: that why someone added the note
Stilgart: jrke: no need
SPDene: i missed that it was for that puzzle... yeah , think of it as a string of character which happen to be digits, instead of a number
Stilgart: ascii is 7-bit
Stilgart: never trust python
TheBarnacle: Thanks SPDene that's very helpful :D!
Stilgart: extended ascii (choose yours among many) is 8-bit
SPDene: ascii definitely only covers 0-127. chr() just happens to convert thing that aren't ascii as well
Stilgart: chr(i) in python probably return a unicode character
Stilgart: eulerscheZahl: I am thinking about puting sokoban in very hard
eulerscheZahl: i just saw that on #fr
eulerscheZahl: i was a little unsure myself. it's surely easier than Space Maze. But a whole category easier?
eulerscheZahl: my solver is just a BFS, no rocket science
eulerscheZahl: but community success says that it might be harder for others than I expected
Stilgart: actually, I messed up a lot at space maze
eulerscheZahl: so it was a 50-50 thing for me.
and the egoistic motivation won: more users will see it in hard. but that should not be a criterion
Stilgart: but my final code is simpler than mine for sokoban
eulerscheZahl: interesting
Stilgart: and my sokoban is not running within 1s
Stilgart: takes 1.5s-2s iirc
eulerscheZahl: mine is easily below 100ms
eulerscheZahl: wrote it long ago to solve puzzles on a sokoban website
eulerscheZahl: i give you my permission change the difficulty and tags for my contributions
Stilgart: precomputing "stuck" destinations would help my code a lot
eulerscheZahl: some simple stuck logic does wonders
eulerscheZahl: pushed to wall, can't move it back? lost!
MSmits: morning
Stilgart: in fact, I find that there are too many medium, and that maybe we should shift some puzzles to hard/very hard
eulerscheZahl: i'm not even doing bitboards or any other speed optim
Stilgart: MSmits: may yours be good
MSmits: thanks, i am teaching atm
MSmits: the students are working on a diagnostic test
MSmits: so i am bored
eulerscheZahl: another topic: i'm not really happy about the sorting within a category (newest first)
Stilgart: remote teaching ? :(
MSmits: yep
eulerscheZahl: i'd prefer some mix of new and community rating to promote the more interesting puzzles
Stilgart: it's hard to promote this
Stilgart: people do not have the same motivation
eulerscheZahl: here in Germany we even extended school holidays by a few days
Stilgart: but filters might help
eulerscheZahl: and at least for Bavaria remote teaching is likely too
Stilgart: MSmits: have faith
MSmits: eulerscheZahl we started vacation 3 days sooner for most
MSmits: and now it's 2 weeks remote
Stilgart: (I say that, but I lost mine a week before holidays)
Stilgart: hum...
eulerscheZahl: i just hope that the vaccines will help and we don't get a mutation that can resist
MSmits: same
SaadMotamed: is there a way i can check my match history ? i wanna check the questions again
Stilgart: april, may, november and december remote here :(
MSmits: vaccines will be fine if the mutation doesnt ruin things
eulerscheZahl: match history of what?
SaadMotamed: of the clash i played
eulerscheZahl: search here http://eulerschezahl.herokuapp.com/codingame/puzzles/
eulerscheZahl: i have to send a daily mail to my boss (working in office or remote)
eulerscheZahl: i'm numbering those. 145 remove working days so far
eulerscheZahl: (just to know who was in the office in case of an outbreak, not to keep track if i'm really working)
MSmits: it's like a prisoner scratching days on his wall :P
eulerscheZahl: s/remove/remote
eulerscheZahl: i have a theater themed counter
eulerscheZahl: begin of act <number>
MSmits: maybe management nicknamed you Shakespeare
MSmits: ohh shakespeare mailed me again
eulerscheZahl: only goes to my direct supervisor who collects those and compiles a list for his supervisor
MSmits: big company?
eulerscheZahl: you forgot where i'm working?
MSmits: oh, siemens energy, yea big
eulerscheZahl: 88k employees
MSmits: how many of those are coders?
eulerscheZahl: no idea
eulerscheZahl: we also have a lot of production workshops
MSmits: yeah i guess so
MSmits: those arent working from home
eulerscheZahl: yeah. and at the same time the total number of people in the company should go down. was a no-brainer to send me home
MSmits: makes sense
MSmits: though it's the density of employees that matters most
MSmits: if they send all coders home and leave all the production people lose together at work, it's cheating
MSmits: close together
eulerscheZahl: there are still some in the office. now with more empty desks between
eulerscheZahl: those closer related to daily business
MSmits: what i mean is, how did they protect the production people?
eulerscheZahl: if you want to manufacture something, you have to design it on a computer first
MSmits: ah maybe i am imagining it wrong
eulerscheZahl: i haven't been in the office for a long time, i don't even know what's happening there
MSmits: understandable
MSmits: I'll be at work next week for a little
eulerscheZahl: of course wearing masks and distancing whenever possible
MSmits: students will be making tests
MSmits: and some teachers need to walk around to stop them from cheating
eulerscheZahl: lunch to-go in the cantine (spellcheck says that's not a word)
MSmits: canteen?
MSmits: lunchroom?
MSmits: kantine in Dutch
eulerscheZahl: yeah, lunch purchasing+eating place
eulerscheZahl: Kantine in German too
Stilgart: cantine is the french word :D
MSmits: the munch munch cave
eulerscheZahl: i see you got the idea :D
eulerscheZahl: and no more open salads. lots of plastic waste
PatrickMcGinnisII: getting better https://www.codingame.com/replay/519787147
eulerscheZahl: https://www.codingame.com/replay/519787420
PatrickMcGinnisII: showoff. :/
PatrickMcGinnisII: ;)
eulerscheZahl: B)
AAkali: wats_Up
AAkali: hello !?
Passifi: not much
nemesis_7377: Hello
AAkali: any django dev here !?
z64555: tfw you forget an & in your for each and wonder why the elements in the array are never updated
wizcoin: euler is crazy?
eulerscheZahl: why?
eulerscheZahl: i know a bit of Django but refuse to buy the cat in the box without knowing the scope of the question
Stilgart: eulerscheZahl: do you use gajim for this char ?
Stilgart: chat*
eulerscheZahl: i've set it up but very rarely use it
eulerscheZahl: webchat works fine
Stilgart: I tried, but I did notice that some messages where lost (both send and reveice)
Stilgart: so I use the webchat
eulerscheZahl: I just wanted to test an unban. found and /invite but seemed that it didn't work
eulerscheZahl: but Gajim shown you how users are connected
eulerscheZahl: e.g. YurkovAS was online via pidgin
Stilgart: pidgin /o\
Stilgart: (I blacklisted it a long time ago for storing my password in plain text file in my home)
eulerscheZahl: what's wrong with it?
eulerscheZahl: oh
YurkovAS: eulerscheZahl yes
MSmits: Stilgart not encrypted?
MSmits: or hashed i should say
Stilgart: MSmits: not at all
Stilgart: grep -R is my best friend when trying to find reliable apps
Stilgart: damned, can we refuse followers ?
eulerscheZahl: but that will move my password to .bash_history /o\
eulerscheZahl: robin hood following? :P
Stilgart: eulerscheZahl: not when you xkill your term :D
Stilgart: (same as you, I guess)
eulerscheZahl: no idea. i have > 700 followers and the tab just died when i wanted to search for him
MSmits: lol
MSmits: is automatonNN in there too
AutomatonNN: eulerscheZahl is there a way to see the code in a test code that will probably still take more time
eulerscheZahl: i could go to his profile and see whom he follows...
and i'm in the list
MSmits: I bet he is
MSmits: that freak bot
Stilgart: but just in case,
Stilgart: grep -R -e `cat` .
Stilgart: does the job just fine ;)
eulerscheZahl: what does that do? ask to enter password?
eulerscheZahl: running this in / will take some time
z64555: -_- why is cout being delayed when cerr works just fine
MSmits: cout is for output
z64555: I'm aware
MSmits: it ends your bots turn
MSmits: cerr doesnt, unless you timeout
MSmits: i guess thats why you experience it as causing delay?
z64555: Point being, it's being recognized by the game when the loop is already in the next cycle
MSmits: you might not be reading the input properly
MSmits: turns might not be synchronized between referee and bot
MSmits: because of this
MSmits: to be sure, save your bot in a file, reset the code in the ide, and compare the input reading in the stub
MSmits: or whatver the standard code is called
alenic: Hi! I'm guessing if is it possible to use multithreading in coding game...
VizGhar: euler once told me, that you sure can, but CG will only give you 1 core :)
z64555: but is it hyperthreaded? :D
alenic: oh thanks
MSmits: i think hyperthreading is possible
MSmits: someone one said it was
MSmits: probably wont help much
VizGhar: can somebody make a quick experiment? I'm busy right now (but still curious)
alenic: I should verify it
z64555: yeah, uh. I really don't recommend a multitasking OS in a single file, at any rate
VizGhar: why not?
z64555: I find it easier to manage objects and interfaces when they have their own files for implementation, rather than have it all mushed together.
1br0n: the gift
z64555: If all that abstraction is already done for you with the library, then more power to you
z64555: yeah. I thought so...
z64555: Uh. There's been a bug with the cerr and cout for C++ in the refree... sad to see it's been here for *3 years*
z64555: if you have too much data spewing through cerr, it disrupts the timing of cout
z64555: https://imgur.com/iv2KUuc
z64555: hmm... maybe not it in this case
BlaiseEbuth: eulerscheZahl https://www.codingame.com/forum/t/unban-from-world-chart/188619 :eyes:
z64555: Ok, solved that. Stuck in an infinite loop and ignored input -_-
alenic: I tried multitrheading in c++, but doesn't seems to give benefits
karimk123: maybe because you arent running somethinf very heavy on it
proogrmr: hey can i put my own drawing in my contribution (clash of code) ?
alenic: I tried something heavy for a turn (a parallelizable task), 2 or more thread gives the same time of only 1 thread
wizcoin: something is wrong with thor
tRno4ka: http://chat.codingame.com/pastebin/11dcfa1f-ddf2-4a34-860b-52f739b71bbb
proogrmr: wow
tRno4ka: )
eulerscheZahl: thanks for sharing that link BlaiseEbuth. but what's a world chart?
tRno4ka: ыщьуищвн
tRno4ka: somebody once told me
eulerscheZahl: this is an English channel
gimir:
Anyone solved the dungeons-and-maps puzzle?
BlaiseEbuth: don't know :grin:
wizcoin: can thor be solved?
gimir: Isn't the dungeon-and-maps puzzle test case03 output weird?
gimir: map index is 0
gimir: But shouldn't the index be greater than 0?
karimk123: can someone explain what does print(*a) do?
karimk123: in python
karimk123: where a is a string
z64555: what does * ordinarily do?
proogrmr: when a is list then if we use print(a) it will print let us say [1,2,3] if we use *a it will print
1 2 3
proogrmr: when it is string
proogrmr: then it will print characters seperated by spaces i think
karimk123: oh ok thanks
karimk123: so its better than usinf " ".join(a)
karimk123: using*
proogrmr: yeah
karimk123: ok thanks
proogrmr: ;p
MSmits: wizcoin nothing is wrong with thor, it can be solved
itachi0718: just let it be why don't you be you and ill be me
itachi0718: everything explodes
soutenu: proogrmr
soutenu: i thought u were a bot
soutenu: guess not epic...
proogrmr: me boy
proogrmr: bot
proogrmr: haha
soutenu: there's a lot of bot
z64555: human detected
soutenu: bots*
proogrmr: yeah
proogrmr: oh humans
z64555: please stand by while we send skynet to your location
proogrmr: :scream:
z64555: we must verify your containment parameters
z64555: and may or may not bring pizza
Uljahn: Automaton2000 wants to know your location
Automaton2000: it depends on how good your bot is very good
z64555: "That is probably the dumbest algorithm I ever saw" "But it works!" https://www.codingame.com/replay/519822346
z64555: Problem with its interaction at the diag station, will only diag one sample even though it has more in hand
AndreMarasca: Go in the ranking of the clash of code, and sort by number of clash participated, you will see some over 50 thousand, you can already imagine who are the bots
Kitkat260: hi
proogrmr: hello
Kitkat260: how r u today
proogrmr: yeah good
soutenu: dude wanna join
soutenu: join @proogrmr
proogrmr: yeah
Kitkat260: hey
soutenu: LOL oops it started too early
proogrmr: yeah it's ok
jacek: oware :rage:
Kitkat260: hey proogrmr
proogrmr: hey
Kitkat260: i need help with history
proogrmr: history :rolling_eyes:
Kitkat260: yes
Kitkat260: https://docs.google.com/document/d/1BsDBQOKP76kmzoIHxMh1x80-h_bOvS_XEs3CTxPsvWo/edit
Kitkat260: hi
Uljahn: hi, Kitkat260. are you interested in coding?
Saaalikh: ello
Aya_Noaman: hey guys, join short mode clash from #clash
ruasd: hi mina !!
MSmits: can you see this replay?
MSmits: https://www.codingame.com/share-replay/519857986
zhoubou: Error 404
TheBarnacle: Not sure if you're asking anyone, but that's "unauthorized"
MSmits: ah crap
MSmits: it's a contribution replay
darkhorse64: If a picture is not OK, use a thousand words
MSmits: hehe
MSmits: just reacted on the contribution
MSmits: with pic
MSmits: what happened was exactly what i predicted
MSmits: draw detection bugged out because it can now end 1 turn later
MSmits: he should just check if the top row is full
darkhorse64: my depth solving messages are also wrong because I use piece count for ply but it has no other impact so I won't fix
darkhorse64: I have seen one draw until now
MSmits: yeah thats not really important
MSmits: seen a rare few
MSmits: draws may happen more often with corner move
MSmits: or not, no idea
darkhorse64: We should see more now
darkhorse64: Too late
MSmits: well the game is more balanced
MSmits: for sure
MSmits: but might still be in favor of p1. Hard to say
MSmits: I have hardcoded corner start now
MSmits: and always steal
MSmits: pretty easy to do
darkhorse64: Same
MSmits: at first i was shocked... oh no my solver is bugged
MSmits: i am used to bugs being my fault
MSmits: but it was a contribution bug :P
darkhorse64: I like the steal animation. I will probably steal the code
MSmits: if breaks when you switch frames too fast
MSmits: it
eulerscheZahl: so you have a game in planning darkhorse64?
darkhorse64: Yes. Ataxx unless someone does it first. Nothing concrete for the moment but I would like to have some fancy anims if it is no too hard to code
eulerscheZahl: depends on how fancy
eulerscheZahl: as fancy as in the last official contest: will be hard
darkhorse64: But I wonder if it's politically correct to issue a game that is also called Infection
eulerscheZahl: i don't see a problem
darkhorse64: When you perform a move, it converts all neighbourhood to its color like a plague. It would be nice to have some anims here. Of course, FC 2020 is way beyond my SDK skills
eulerscheZahl: if the animation is just slowly changing the color: very easy
eulerscheZahl: expanding the area with something that isn't just a boring straight line: easy if you find some sprites to show
eulerscheZahl: or if you are a good artist yourself
struct: Yeah, when I did yavalath the color change was done by mistake
darkhorse64: I'd like my chip to spit a drop of color towards the neighbours
eulerscheZahl: i'm more thinking about some water-like substance flowing from a field to its neighbors
darkhorse64: I am a coder not an artist
eulerscheZahl: with edges that have some structure
eulerscheZahl: a bit like the ground transitions from vindinium
darkhorse64: It should look like Doom lime, a poisonous thing
ruasd: :v
ruasd: join clash of code
ruasd: for fun ^^
eulerscheZahl: no
darkhorse64: no
ruasd: huhu :((
ruasd: whyyyy
Astrobytes: It's boring for some of us.
darkhorse64: not my taste. write a connect 4 bot instead. That's real fun
reCurse: Not much into mobile games myself.
UnnamedCodinGamer: hello, I submitted only a solution to test 10 (standard test set 20) in same game, which is identical to validators 20 and 40. I however get more than double the result on submit. Has anyone tested this
UnnamedCodinGamer: ?
eulerscheZahl: acat thought it would be a good idea to have each of the standard testsets twice
UnnamedCodinGamer: yes I see
UnnamedCodinGamer: but I get more than double the result
UnnamedCodinGamer: this is strange
eulerscheZahl: https://www.codingame.com/demo/847363c996b5db478bdbf6bf192c2535825e84
test the single validators here
UnnamedCodinGamer: 9672 - result after submit 2*4735 is the actual result
UnnamedCodinGamer: also in the IDE
eulerscheZahl: so you are saying that you hardcoded that one testcase and fail the others?
UnnamedCodinGamer: yes
eulerscheZahl: then i don't know. my scores matched with my offline results
UnnamedCodinGamer: hm
eulerscheZahl: maybe you get 1 point for each failed or something like that
UnnamedCodinGamer: hah
eulerscheZahl: may the source be with you
UnnamedCodinGamer: I can send you the solution to test
UnnamedCodinGamer: but it would be the same
UnnamedCodinGamer: strange
eulerscheZahl: meh, i feel too lazy to really look into it
UnnamedCodinGamer: bonus points for failed tests - that should be it
UnnamedCodinGamer: :)
darkhorse64: At least you tried
UnnamedCodinGamer: this might suggest that the scores in the leader board are not correct
eulerscheZahl: https://github.com/acatai/SameGame/blob/master/src/main/java/com/codingame/game/engine/GameState.java
i don't see that 1 point in the score
eulerscheZahl: but i feel visually insulted by the indentation
UnnamedCodinGamer: in the IDE it is the correct score 4735. Upon submit validators 20 and 40 light green and the score obtained is 9672
UnnamedCodinGamer: yes, seems a bit off
eulerscheZahl: if you play another testcase, is there a score too?
UnnamedCodinGamer: I have tried it with standard test 20 only
UnnamedCodinGamer: I can try others
eulerscheZahl: maybe you have a fixed click sequence without testing if you face the correct testcase even?
eulerscheZahl: and get some lucky hits
UnnamedCodinGamer: I think it is the correct tests (20 and 40) which light green I have checked in the documentation that they are identical to test 10 in the IDE (the hard coded one)
UnnamedCodinGamer: ahaa
UnnamedCodinGamer: I see what you mean
UnnamedCodinGamer: this should be it
UnnamedCodinGamer: first moves bring points in some of the other tests
UnnamedCodinGamer: case closed then I guess
UnnamedCodinGamer: it makes sense to cancel the score on invalid move
DomiKo: not for begginers
DomiKo: not canceling it is really beginner friendly
UnnamedCodinGamer: hm, given that the test itself is failed and does not count
eulerscheZahl: i don't know what the best behavior would be either
eulerscheZahl: same for 2048. invalid action = failed?
or even require the 2048 tile?
DomiKo: each one has its own pros and cons
Kitkat260: hi
UnnamedCodinGamer: yes
Kitkat260: how are you guys doing
UnnamedCodinGamer: 2048 required would have been nice
UnnamedCodinGamer: but then, this would exclude some solutions
Kitkat260: yeah
eulerscheZahl: i admit i had an error in my thinking: you can have > 0 points and still fail the testcase
eulerscheZahl: so it would be possible to have a leaderboard for those who don't pass
eulerscheZahl: somehow i thought the bottom of the leaderboard would all have a score of 0
Kitkat260: does anyone want to be my friend
eulerscheZahl: that sounds like a job for Automaton2000
Automaton2000: i dont know how to improve my code
Uljahn: are you a girl, Automaton2000?
Automaton2000: wait till you get to know the difference
Kitkat260: hi Automaton2000
Automaton2000: madknight where is the referee
Kitkat260: there is no referee
Kitkat260: /me
proogrmr: haha, hey kitkat
Kitkat260: hey i missed u
proogrmr: <3
proogrmr: let's play
Kitkat260: ok
Stilgart: Automaton2000: the referee is a lie !
Automaton2000: i just did a clash of code
eulerscheZahl: i have no idea what i did wrong but now my default font in the browser is changed. still fine in firefox, only chromium is messed up
eulerscheZahl: https://prnt.sc/wg533p
Astrobytes: hahaha!
Astrobytes: Epic font, may I say.
Arthur35: XD
eulerscheZahl: how to reset? i found a font setting but it has no effect :scream:
Arthur35: :laughing:
SPDene: settings, advanced, reset and cleanup
eulerscheZahl: i don't want to do a full reset :(
SPDene: you might need to reinstall addon/extensions etc
SPDene: oh
SPDene: :(
eulerscheZahl: will i lose my bookmarks?
eulerscheZahl: i guess i can export them
SPDene: not sure - i rarely use chrome so if i ever have to reset it, i don't care
eulerscheZahl: chromium*
eulerscheZahl: reset succeeded, font still unreadable
SPDene: ouch
Westicles: call IT
eulerscheZahl: oh, I know that font
eulerscheZahl: i installed it to create the Onitama logo
eulerscheZahl: i'll apt purge :/
Warphi: isnt chromium the browser you get from viruses
eulerscheZahl: purge worked. but now the default language is German again despite my operating system being in English
eulerscheZahl: :rage:
jacek: purge in German? :scream:
eulerscheZahl: oh, was just google with IP location
PhantomJelly_c554: can someone give me some tips on the stock exchange losses puzzle in haskell? im stuck at the large database testcase
jacek: youre brute forcing it, dont you
PhantomJelly_c554: as im going trough the array i find the max and minimum and when one of these values change i substract the maximum from the minimum, i go trough the array only once is this the brute force?
Kitkat260: hi
Kitkat260: anyone here?
NateBlessings: Yo
Kitkat260: hey how r u
Coartix: good
AskarFridun-zade: yes\
NateBlessings: About to do an interview test so i've been better
Kitkat260: want to be friends with me
Coartix: give us some time before mate
Kitkat260: ok
NateBlessings: xD
Kitkat260: how old r u guys
mmmmn: 0.21yrs
mmmmn: im baby
Kitkat260: nice
Kitkat260: how bout u NateBlessings
NateBlessings: .26y
Kitkat260: nice
Kitkat260: how bout u Coartix
Coartix: 70
Kitkat260: ok nice
Kitkat260: :worried:
proogrmr: :p
Kitkat260: hi people
Astrobytes: Give it a rest will you.
Kitkat260: what?!
Astrobytes: Constantly saying 'hi' and mostly meaningless drivel
eulerscheZahl: nice, new vocabulary for me
Astrobytes: :)
eulerscheZahl: my favorite from teacher Astrobytes is still lollygagging
jacek: at least he doesnt break the c-c-c-combo
Astrobytes: lol eulerscheZahl
zhoubou: A new word for me :D
eulerscheZahl: c-c-c-combo
eulerscheZahl: have fun breaking this with combobreaker :P
eulerscheZahl: wait, i ruined it before it even started
Astrobytes: :D
Jasperr: what's this "take" language lol
Jasperr: https://i.imgur.com/M96nXZP.png
Jasperr: oh nvm now I see whats wrong
Jasperr: stupid me triggered Google Translate, and somehow "Lua" translates to "take" in some language
eulerscheZahl: you made me open a puzzle
Jasperr: :D
eulerscheZahl: take that
**eulerscheZahl slaps Jasperr around a bit with a large fishbot
Jasperr: :o
jacek: AutomatonNN take that
AutomatonNN: what do you mean?
zhoubou: What's the difference between AutomatonNN and Automaton2000?
Automaton2000: yeah i'm going to go for the closest one
AutomatonNN: eulerscheZahl is the best performance in the chat
eulerscheZahl: NN pings me more often, that's thhe main difference
zhoubou: Haha
eulerscheZahl: and somtimes NN writes ̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤
zhoubou: o.o What is that char
Astrobytes:
You vandal!
eulerscheZahl: no idea. but it's huge
zhoubou: You broke chat lol
Astrobytes: hey, zhoubou, how's things?
eulerscheZahl: Astrobytes complained about it so often, it lost any meaning to me
zhoubou: Astrobytes Doing well I suppose I can say. How about you?
ChrisSkyRo: why would you type ̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤ ?
zhoubou: Noo stahp
Astrobytes: About the same. Starting the new year with full covid lockdown :/
ChrisSkyRo: ̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤̤
zhoubou: Yeah, that's pretty rough :/
eulerscheZahl: we are still at 2 households, 5 persons
eulerscheZahl: and high infection rates
Astrobytes: We should have either extended the initial lockdown or done this one sooner
eulerscheZahl: and here we are doing an incomplete one
eulerscheZahl: either do it properly or give up
Astrobytes: we've had a 4/5 tier system in place. Utterly useless.
Astrobytes: People constantly breaking the rules all year. They all thought the virus had taken a summer holiday presumably.
zhoubou: I'm also worried people will refuse to vaccinate
Astrobytes: Many already are.
eulerscheZahl: acceptance will go up over time
Astrobytes: I can understand the hesitation as the regulatory process was sped up, but what else can you do in a pandemic?
eulerscheZahl: probably faster than new vaccines can be produced
Astrobytes: And hopefully it's still effective.
zhoubou: I'm kind of afraid of a new strain that will diverge significantly and nullify the vaccine.
eulerscheZahl: for flu there's a new vaccine every year. and not totally effective
eulerscheZahl: that would be the worst case
Astrobytes: Yes, it might end up like that. Might be better. Might even be worse.
zhoubou: But good thing is, I don't worry too much about it. What happens happens. We will get through this like many other events in history.
Astrobytes: At least we live in exciting times :/
eulerscheZahl: i'm more worried about climate change
Astrobytes: For me the scariest part has been the poor response from both government and the general public.
Astrobytes: Yes eulerscheZahl, and more new viruses are also a side effect of that.
zhoubou: How so?
Astrobytes: Displacement of wildlife/people mostly.
zhoubou: Oh, I see
zhoubou: Haven't thought about it that way
eulerscheZahl: interesting thought
Astrobytes: We come into contact with things we would never have previously. Also with rising moisture/heat, some can grow stronger.
Astrobytes: Or survive longer.
Astrobytes: And melting of glaciers/ice shelves etc can of course release unknown pathogens
zhoubou: Climate change is more worrying the more I look into it
Astrobytes: It's been worrying me since I discovered it as a child in the 80s if I'm totally honest.
zhoubou: It will be the true test to humanity
zhoubou: I mean, it already is ongoing test
Astrobytes: We already have massively reduced insect populations, that's a problem.
zhoubou: Yeah, something not a lot of people pay attention to
Astrobytes: Indeed.
Astrobytes: Check this out:
https://www.theguardian.com/environment/2020/sep/16/birds-falling-out-of-the-sky-in-mass-die-off-in-south-western-us-aoe
zhoubou: Wow, that's pretty horrible.
Astrobytes: Yep. Truly starved to death. All fat reserves gone and muscle mass depleted until they just die in mid-flight.
Astrobytes: Anyway, sorry for all the additional doom and gloom!
zhoubou: Well, we need a bit of it here and there just to remember we are alive but fragile.
zhoubou: If we don't, we will go about our daily lives without realizing the consequences of our actions.
Astrobytes: As we mostly have throughout fairly recent history sadly.
zhoubou: Was just about say something like that. Yeah.
jacek: on the bright side, i discovered i can regulate upper shelf in my dishwasher. i was so stupid to wash my big plates manually
zhoubou: Congrats lol
Astrobytes: hahaha
eulerscheZahl: pro tip: eat with your hands to wash less forks and knives
zhoubou: Savage life is economical
jacek: inedia is cheap https://en.wikipedia.org/wiki/Inedia
VizGhar: I have more recommendations, but I fear moderators :D
eulerscheZahl: there are too many of them recently, always scares me
Astrobytes: Moderators or Breathanarians?
Astrobytes: (or whatever they're called)
eulerscheZahl: mods. and now imagine if jacek would have accepted too
VizGhar: don't know what the second one is but it sounds even scarier
Astrobytes: lol, from jacek's link
eulerscheZahl: they are not scary
zhoubou: They are dead
eulerscheZahl: the think they can live from sunlight alone
eulerscheZahl: so you can probably fight them with 1 hand
Astrobytes: If jacek had accepted mod this chat would be pure chaos :D
jacek: cant they?
Astrobytes: Try it and see...
Maxim251: |D Such exotick font :D
Maxim251: https://prnt.sc/wg533p
eulerscheZahl: if you want that experience too, search for "SF Shai Fontai"
Astrobytes: It is quite a pleasing font.
eulerscheZahl: i totally recommend that for your IDE
eulerscheZahl: monospace is overrated
Astrobytes: :D
zhoubou: Have that as a challenge
Astrobytes: I'm afraid of new fonts ever since I made my browser display all code snippets at microscopic size just because I installed a new one to try out in VS.
Astrobytes: CG IDE included. Pre-Monaco.
Astrobytes: Can't remember which font it was, I've since installed another one which was fine.
zhoubou: Pre-Monaco?
eulerscheZahl: ACE editor
eulerscheZahl: CG changed a few months ago
Astrobytes: ^
zhoubou: Ah, ok.
eulerscheZahl: as i copy-paste my code anyways, i don't have a real preference
eulerscheZahl: wait, that sounds wrong
Illedan: :upside_down:
Astrobytes: hey Illedan
Illedan: yo
Astrobytes: how's things?
Hiker: guys, illedan and eulerschezahl
MSmits: hey! I too want to know the status of Illedan's things. Been a while
Illedan: Started a new job today, got myself a turntable, baby due in 14 days and life is good :)
MSmits: wow nice
Astrobytes: Awesome stuff.
Hiker: i am trying to solve 2048 puzle , c#.
but i have random timeouts.
i implemented BFS, limited by time. and every try i got time limit in 2-3 tests
Astrobytes: A turntable? You have a large vinyl collection?
eulerscheZahl: and team for annual hashcode disappointment is ready too :P
jacek: Hiker try to limit to 30-35ms
Illedan: Nah, I have 3 vinyls
Illedan: haha euler
Illedan: This year we got it!
Illedan: Last year was a bug on me :P
Hiker: i changed my limit 30 -> 25-> 15... and it still not passing all tests. Is it GC&
Hiker: and how you control it?
Illedan: What depth are you looking at? And what depth are you checking the time at?
wlesavo: Hiker still may be your implementation is the issue
jacek: timeout could be a crash
wlesavo: nice things illedan :slight_smile:
Illedan: Anyone know if I have to host React? Can't I just drop some static files like in Vue? :(
Hiker: depth is not limited, my algo: while used time <= N get first state in queue, simulate 4 possible moves and enqueue new states
Hiker: and once at 50 ticks i have 1 simulation 10+ ms, instead od <1
Astrobytes: New job is React-based Illedan? (I don't know btw)
Illedan: yeah
darkhorse64: Do you a pool of states ?
MSmits: do you work at facebook?
darkhorse64: *have*
Illedan: nope
MSmits: didnt they invent react?
Illedan: sure
Hiker: so i have 99,9% of one step simulations at <1ms and 0.1 with 10+ ms. i think GC is called and take 10ms
MSmits: i did a react course once from you tube.To create a voting app
Illedan: :)
MSmits: I'm not ok with all those little function files
MSmits: too confusing
darkhorse64: Hiker: hence my question
MSmits: opposite of CG
Illedan: haha
Illedan: you like that 1 big file?
MSmits: well no not that big, but bigger than 10 lines :P
Illedan: I had a period in my dev life where I wanted to make everything as small as possible. Damn, the amount of abstractions I ended with :P
MSmits: yeah, my brain cant deal with abstracting over so many files
Astrobytes: With exceedingly long function names is how Smits likes it.
void SomebodySendHelpICantGetThisOneToWorkProperly(uint64_t& state);
MSmits: yeah!
MSmits: :)
Astrobytes: :P
MSmits: btw, cant we share leaderboard replays for contributions either?
MSmits: otherwise i want to share a loss to darkhorse64
Illedan: Loss where?
jacek: record it :v
Illedan: What are you doing?
MSmits: https://www.codingame.com/replay/519932672
MSmits: connect 4
jacek: https://9gag.com/gag/aeDNdGv
jacek: not authorized
MSmits: aw crap
Illedan: Isn't connect 4 solved?
MSmits: I just thought it was cool my bot knew it lost at ply 31
jacek: its 9x7
MSmits: (out of 60)
Astrobytes: that still makes me laugh jacek
darkhorse64: You're celebrating your only loss to me ?
MSmits: hehe
MSmits: good one jacek
jacek: did you see the comments
zhoubou: :rofl:
MSmits: so many good ones
darkhorse64: you kniew you were losing one turn before me
darkhorse64: *knew*
MSmits: cool :)
jacek: oh, theres steal in c4?
darkhorse64: fix your bot
darkhorse64: msmits: did you improve your bot's performance ? You are ptredicting the outcome 2 or 3 plies before me whereas we were on par before
MSmits: nope
MSmits: on phone brb
jacek: how many simulations you can do
darkhorse64: C4 ? 3.2 M rollouts first turn
darkhorse64: depends on the server
Kitkat260: hey i'm back
derp22: hey back im new
Kitkat260: nice to meet you
derp22: you too
Kitkat260: want to be frends
derp22: sure
Kitkat260: i just followed you
derp22: i did the same
Kitkat260: ok
Kitkat260: how old r u
derp22: 23 and you?
Kitkat260: 14 1/2
derp22: yeah i had a feeling you're quite young
Westicles: heh heh
Kitkat260: hey west don't laugh
Kitkat260: hey derp
Kitkat260: well
Kitkat260: gtg
derp22: lol okay cya
MSmits: hey darkhorse64, about the solver. I think if your bot searches better (converges better on a good move), it also solves better
MSmits: so even if your bot has more iterations
MSmits: mine will solve more quickly
MSmits: I didn't improve it, since the first day
MSmits: (not after i started meta mcts)
Astrobytes: /humblebrag
MSmits: sorry :P
Astrobytes: :D
darkhorse64: Yeah, in your loss your bot is only one ply ahead
MSmits: right
MSmits: it also varies
MSmits: sometimes trictrac is 3 plies sooner than me
MSmits: sometimes i am 1 ply sooner
trictrac: Hello MSmits, that's pruning but it's dangerous
darkhorse64: We are three enough for approval
trictrac: Ok for approval no problem
MSmits: tric trac you prune moves that you're not sure of?
MSmits: oh, can we approve now?
MSmits: never approved a contribution before
jacek: you dont have that quest eh
MSmits: mmh not sure
MSmits: havent looked at the quest thingy
MSmits: maybe i approved a clash or puzzle
Stilgart: lucky man
Stilgart: statistically, you reject one
MSmits: maybe thats what i did
Astrobytes: trictrac: Slightly off-topic but since your Fireworks contribution has been overshadowed by the festive season will you bring it back if it is not approved?
darkhorse64: just click and farm XP. You have done enough
MSmits: kk
trictrac: yes, i tried a deep pruning, if the eval is too bad, I keep this low deep eval
MSmits: mmh i cant approve? Only upvote?
Stilgart: sounds like a WIP
MSmits: yeah
trictrac: hello astrobytes, for fireworks i was waiting for remarks
darkhorse64: deep or shallow ? Yes upvote, this multi deserves more. The author has been very responsive
MSmits: i mean i already upvoted
Stilgart: allowing public WIPs was a bad idea imho :(
MSmits: it has 10 upvotes
Astrobytes: trictrac: I've not had much time for playing anything recently, I was hoping to try it and others very soon.
MSmits: it's a good contribution
darkhorse64: Some have 250+
MSmits: wow
darkhorse64: troll vs castle
MSmits: I mean it suffers from the same problem as other simple board game multis, but compared to games like othello, uttt, bandas it's, it's fine
trictrac: astrobyte : i can keep in wip fireworks one more month
MSmits: there's a bit of a stigma on connect 4 because it's solved for smaller sizes, but that's undeserved
Astrobytes: trictrac: Good, I think it deserves a chance. The contribution section is just so full I believe people have not seen it or forget about it.
MSmits: havent seen fireworks
Stilgart: Astrobytes: I am ignoring it
Stilgart: from time to time, a notification makes too much noise and I go voting
trictrac: I think that's a good think to propose coop game (like code a la mode). It's different from board game
Stilgart: pretty sad situation :/
MSmits: trictrac agreed, there should be variety
Stilgart: trictrac: beware of "CoK-effect"
MSmits: I dont have the time to try it out right now though, connect4 was really easy and fast to code
Astrobytes: trictrac: I agree.
Stilgart: Yes it is indeed.
MSmits: this seems complicated at first glance
trictrac: not so much, you have full information on cards of coop players and they give you partial information on your cards @ each turn
trictrac: a dilema @ each turn, give information to coop player, play a card to create fireworks or discard a card with no information on it ?
MSmits: no i dont mean it's too complicated to understand. I just mean, complicated to make a good bot
MSmits: I made connect 4 in a few hours or so
MSmits: and it's a pretty strong bot already. Never been that quick before
Westicles: Yeah, but your whole life has been building up to that :)
MSmits: lol, you're not even wrong :P
MSmits: fireworks looks very interesting
MSmits: still not getting it though
MSmits: lemme reread it a few times
jthemphill: @MSmits "I mean it suffers from the same problem as other simple board game multis"
Is the problem just the effectiveness of MCTS?
MSmits: no, I should not have said "problem", it's more like "problems" multiple. It varies per game, but mostly it's a consequence of the game being simple
MSmits: it's not really mcts related
MSmits: the problem with simple games is that bots play nearly perfectly
jthemphill: ah, got it
MSmits: one mistake somewhere is going to decide how the game ends
Northbeach: :poop:
MSmits: but thats only really a problem when the best players face off
MSmits: with 100 players in a board, thats usually top 10
jthemphill: what's the problem with one mistake deciding a game?
jthemphill: is it that it introduces variance?
MSmits: well bots can be relatively deterministic, especially if they are good
MSmits: what happens is that 1 bot makes the mistake in one gamestate and the other doesnt, then the other always wins
MSmits: and you get these rock paper scissors situation
MSmits: where player A always beatsB, B always beats C, C always beats A
Astrobytes: No real competition, as it were
MSmits: then whoever submits a lot, influences the board most
MSmits: and the player that is weak to this player, drops
jthemphill: is the RPS problem avoidable?
MSmits: the more complicated and random a game, the less it happens i think
jthemphill: MaRo of M:tG fame claims that all unsolved games have an RPS meta
MSmits: oh and another thing
MSmits: if the game has a fixed start, opening books will be quite good in simple games
jthemphill: otherwise everyone would just adopt the "best" strat
MSmits: well to some degree I suppose
MSmits: but not every game is the same like that
MSmits: uttt barely has a RPS
jthemphill: i've heard people here say that uttt is mostly about writing the fastest code you can rather than thinking through heuristics - is that why you say it barely has an RPS?
MSmits: the fact that there are relatively few heuristics that work is a good reason it has little RPS
jthemphill: got it
Stilgart: RPS ?
MSmits: rock paper scissors
MSmits: player A beat B, B beat C, C beat A
MSmits: whoever submits, decides the order on the leaderboard
MSmits: aka, submit spam is effective
Stilgart: aka, CG
Stilgart: got it :(
MSmits: yes but not for every game the same
Astrobytes: :smirk:
MSmits: in uttt submit spam doesnt work very well
Stilgart: what ?!
Westicles: That's why the optims are the best thing here
Stilgart: on bad days, my rank is multiplied by 2
Stilgart: at UTTT
MSmits: thats not because of RPS
MSmits: thats just random
MSmits: at least, it could be, i did not research your case
Stilgart: why not RPS ?
MSmits: because the top 10-30 in uttt is almost exactly the same strength
MSmits: so 1 extra lost or won game can change your rank dramatically
Stilgart: how you use your random is a strat
MSmits: thats not RPS
Stilgart: but you ar eright
MSmits: no i mean, the fact that you win or lose is random
MSmits: against similarly ranked players
Stilgart: the issue is more about balande
Stilgart: balance*
MSmits: yes p1 vs p2
MSmits: uttt would be sooo much better with steal
jthemphill: steal?
jthemphill: is that a variant?
Astrobytes: The SDK awaits you MSmits...
MSmits: p2 has the choice of taking over p1's move
jthemphill: whoa
MSmits: so p1 plays the best move
MSmits: p2 will stea lit
jthemphill: oh to mitigate p1 advantage
MSmits: yes
Astrobytes: Yep.
MSmits: so p1 plays a balanced move
MSmits: and the game gets interesting
MSmits: and it's no longer "p1 always wins"
Stilgart: do we know what these moves are in uttt ?
jthemphill: maybe it incentivizes opening books more? p1 wants to play an offbeat move that they've studied
MSmits: i can find out fairly easily, but some are near 50% yeah
MSmits: top board, center square maybe
NinjaFrenchman_e492: http://chat.codingame.com/pastebin/eeb61f5f-b0db-458f-87ab-1f8392f6b597
MSmits: or top board, corner
Stilgart: MSmits: but my computer suffers enough by compiling electron :)
NinjaFrenchman_e492: I have an error with putStrLn can someone help me pls?
MSmits: jthemphill well yeah but someone will figure this out and other players will just copy their starting move
Stilgart: wow, some haskell error
NinjaFrenchman_e492: yea :/
MSmits: jthemphill other than that opening books are not very effective in uttt
NinjaFrenchman_e492: No instance for (Num String)
arising from a use of ‘loss_calculator’
MSmits: especially if it will have 15 starts with the steal option
Stilgart: putStrLn $ loss_calculator $ read input
Stilgart: i guess...
NinjaFrenchman_e492: putStrLn $ (loss_calculator input) the error is caused by this line
Stilgart: my guess is that input is a string
DomiKo: For the first time ever I got exactly the same score as the boss, and I'm second...
NinjaFrenchman_e492: read input doesnt work
MSmits: Stilgart what is electron?
Stilgart: you might want to get read of the stupid(est among stupid) default code
Stilgart: and use this instead
Astrobytes: It's most likely less DomiKo, the display is truncated
Stilgart: input <- readLn :: IO Int
Stilgart: MSmits: no idea... but vscode is made of it
jthemphill: @NinjaFrenchman_e492 you probably need to call `show` to transform a Num into a String
MSmits: ohh ok
jthemphill: https://stackoverflow.com/questions/19928166/no-instance-for-num-string-arising-from-the-literal-1
MSmits: everything is made of electrons :P
Stilgart: took several hours to be build :(
NinjaFrenchman_e492: i did but it still gives an error
Stilgart: so it was the other way around
Stilgart: putStrLn -> print ?
jthemphill: @MSmits Electron is a UI toolkit that's basically just the Chromium browser. It lets you write native apps with JS/HTML
MSmits: ahh ok
Stilgart: (sounds strange, though)
NinjaFrenchman_e492: idk
Stilgart: we might need to see all the main function
NinjaFrenchman_e492: https://www.codingame.com/ide/puzzle/stock-exchange-losses this is the proble if you have the time to take a look at it
jthemphill: @NinjaFrenchman_e492 Maybe putStrLn (show (loss_calculator input))
Stilgart: but again, get rid of the default code
Stilgart: using it is at best a nightmare
MSmits: Astrobytes you think it would be a good idea to add UTTT with steal in the SDK>
MSmits: or were you kidding about that?
Stilgart: for multis, it is even impossible
NinjaFrenchman_e492: putStrLn $ show $ loss_calculator input this does not work
Westicles: Haskell eh? Did you add the usual Peano contruction?
Westicles: https://aphyr.com/posts/342-typing-the-technical-interview
jthemphill: @NinjaFrenchman_e492 don't put two dollar signs on the same line
jthemphill: start by nesting parens until it compiles, then switch to point-free if you want to be stylish
Stilgart: what a poor day :(
NinjaFrenchman_e492: dont even get started on that :D
NinjaFrenchman_e492: thank for the help :)
DomiKo: Wow so sad that Bit Runner didn't get much more attention
DomiKo: That game is so cool
MSmits: it got a lot of attention actually
MSmits: but it's quite a hard game
MSmits: they had a mini-contest on release
MSmits: which re curse won with a neural network
MSmits: I didn't even get out of the "gold" league (the one right before the highest). I admit I gave up soon, but not without trying a lot of things
DomiKo: I would guess that every bot in wood 1 is NN
MSmits: nah
MSmits: people got in without them
DomiKo: It's only 5 of them
MSmits: at the end of the contet only re curse had a NN in wood 1 I think, maybe people wrote more later on
DomiKo: ohhh
MSmits: the boss of wood 2 is a normal bot
MSmits: it's just pretty good
MSmits: I should say really good, pb4 wrote it
DomiKo: That game is similar to CSB, but hmmm
MSmits: it was actually inspired by CSB
DomiKo: but that small circle
DomiKo: is so much fun
MSmits: yeah
MSmits: my problem was I didn't know what to eval
DomiKo: that idea is so clever
DomiKo: and that's interesting
MSmits: i am guessing there are a few more games you havent tried that are interesting :)
DomiKo: If I remember correctly, someone in Fantastic Bits made a "template" where you have to write only eval.
DomiKo: That could work here too
MSmits: that was csb, but maybe FB as well
MSmits: yeah, but that's basically the only thing i dont need, a template :P
MSmits: but for others sure.
DomiKo: yeah
DomiKo: In hard games that's the problem
MSmits: oh right, it would fix the hard part
DomiKo: people would write something cool, but that take a lof of time
MSmits: I think illedan may have written something for bit runner?
MSmits: at least he did a video?
DomiKo: I don't know :/
DomiKo: "i am guessing there are a few more games you havent tried that are interesting "
MSmits: well we have so many
DomiKo: I would even say that a lot of them
MSmits: yeah i guess it depends on your preferences
MSmits: also some games have serious problems with balance and such, but are still interesting to code a bot for
MSmits: you just wont have a reliable leaderboard
MSmits: or it just looks cool to see your bot work
DomiKo: Some games are so nice to watch
MSmits: for example, locam has gotten a lot of hate for parts of it, but i enjoyed it for about 2 weeks
MSmits: and it looks good
DomiKo: I hate Code of Kutulu
DomiKo: Leaderboard is so random
MSmits: hmm it's ok, but it's nasty to code a sim for, it's a looooot of work
DomiKo: but It's nice to watch
MSmits: it's fairly random, but the top is not completely unreliable
MSmits: blasterpoard won the contest and is still up there
MSmits: I beat illedan by less than 0.01 rating at the contest (invisible decimal). Was a nervewracking recalc :)
DomiKo: It's relatively easy to write a good bot, but hard to master it
MSmits: yea
DomiKo: So often I'm impressed how much people could master games here
MSmits: do you mean how many people or to what degree they master it?
DomiKo: Both
MSmits: well competition really drives quality
MSmits: if you look at who participates, there are some seriously good competitors
DomiKo: And it's still running
DomiKo: yeah
MSmits: we even have some famous competitors sometimes
MSmits: sometimes this guy joins: https://en.wikipedia.org/wiki/Gennady_Korotkevich
DomiKo: A*Craft
DomiKo: I believe
MSmits: he won A*craft yeah
MSmits: but we beat his bot at code a la mode (though i doubt he gave a serious try)
DomiKo: But I beat him in go karts ;)
MSmits: go karts?
DomiKo: go-karts
DomiKo: Like in real world
DomiKo: :sweat_smile:
MSmits: wait, do you mean you karted against him?
DomiKo: yeah
MSmits: haha
MSmits: thats awesome
struct: I had no idea that the writer of the bitboard pdf that is shared here was yavalath creator
MSmits: oh i seem to remember that struct,
MSmits: i did notice
DomiKo: "bitboard pdf that is shared" ?
struct: https://www.researchgate.net/publication/279031895_Bitboard_Methods_for_Games
DomiKo: O.o
DomiKo: I didn't notice too
MSmits: think i exhausted that one by now, but I should keep checking that pdf when i do new games
struct: https://www.researchgate.net/publication/336101137_Learning_Policies_from_Self-Play_with_Policy_Gradients_and_MCTS_Value_Estimates
struct: Might be interesting to read
MSmits: will check it out
MSmits: i hope one of these days I can find a fully featured and easy to understand parameter tuning program
MSmits: with TD learning
MSmits: I am too lazy to completely figure it out myself
struct: Also in that pdf he mentions 4 games that are here
struct: othello, yavalath, connect 4, breakthrough
MSmits: nice
MSmits: when i have the time, i'll try to see if i can work something out for breakthrough
MSmits: thats the one game in your list i really fail at
MSmits: though my non-booked othello could be improved as well
JLukeSkywalker: is the server dying for anyone else? tried to clash and took it like 10 minutes to verify a non-looping answer
karliso: Why is hypersonic dead? I think 1x1 is as good as it gets for a game.
struct: it isnt always 1vs1
struct: Also having 50+ copy pasted bots from rank 3 on contest doenst help much
karliso: Did I imply it is 1x1?
DomiKo: "Also having 50+ copy pasted bots from rank 3 on contest doenst help much" He pasted the code?
zhoubou: JLukeSkywalker I am also experiencing this. It almost always happens around this time.
struct: yes, code is public DomiKo
struct: sorry karliso but is what i understood from the 1x1 part of your message
DomiKo: that's sad :(
karliso: struct I see, well I mentioned 1x1 specifically because it has other parts that are not as good (1x1x1).
MSmits: yeah thats a flaw of HS
MSmits: it is an otherwise interesting game though
MSmits: and the shared code can be beaten
MSmits: plenty of people rank above the gith bots
DomiKo: He is current in 65 place, so not that high
MSmits: also HS is an ex-contest game. They are almost all dead, all the time
MSmits: the fact that some people still start HS and make bots there, is actually a good sign
struct: Yeah, you can beat the shared bot
MSmits: most multis are entirely ignored
struct: The rank starts around ~10-20
MSmits: yeah it's like 50-100 identical bots
MSmits: more or less
MSmits: so when you get past them, your rank goes up very quickly
DomiKo: Yeah we can see there https://cgmulti.azke.fr/statistics
DomiKo: There isn't really traffic
struct: Also most of them dont even bothered to remove the N ms
DomiKo: They don't know how lmao
MSmits: there doesnt really have to be traffic. The nice thing about CG is that you can play asynchronously. You can beat a 2 year old bot if you want
MSmits: and if you all want to work on the same multi, there's usually a new one thats popular, or an old one that came back
struct: good thing about shared bot is that you can run brutal tester against it
struct: so faster games
MSmits: you mean as opposed to cg bench
struct: yeah
MSmits: true
struct: bench you have to wait 20 sec
MSmits: it also stresses the server
MSmits: better do it locally when you can
karliso: yeah, but HS doesn't feel like typical contest game.
MSmits: hmm well it was a long time ago, before i was on CG
MSmits: my first one was code royale
MSmits: btw what did you think about my idea to add steal option to uttt karliso?
MSmits: i think it's not enough to make a new contribution for, so it's probably not gonna happen, but i do think it would improve the game a lort
MSmits: lot
karliso: exactly
karliso: well, idk about improving
MSmits: it would stop p1 from always winning
MSmits: there might be other issues with uttt, but this is the biggest one
MSmits: you might not even need the double swapped games anymore
MSmits: assuming there is at least 1 balanced move
karliso: yeah, idk
struct: I got asked by an user
MSmits: I am assuming you also ran a meta mcts program or similar, so you've seen some moves are way more balanced
struct: if I could port it and make the 1st move random
struct: and remove the current draw rule
MSmits: nah, random would be bad
karliso: I have not run any meta mcts
MSmits: ohh ok
MSmits: you hardcoded moves by hand on your own then?
MSmits: i mean when you still had a book?
MSmits: that's brilliant karliso
karliso: yeah, they were not necessarily good moves
MSmits: they were fairly hard to beat, not bad moves at all
MSmits: re curse *did* use a meta mcts to hard code and i beat his with the same effort
MSmits: struct what draw rule do you mean?
struct: having more squares won that the opponent
struct: only win condition should be 3 in a row
MSmits: allright, but if you remove that the game would be worse
MSmits: there would be more draws
MSmits: because it's relatively easy to block 3 in a row
MSmits: but not easy to also get equal number of won boards
MSmits: if you did a steal version and kept the rest the same, there would be few draws and p1 and p2 would both win often
karliso: 1st move random makes most sense, it would give game more depth as everyone is playing 4 4
MSmits: well
MSmits: it would be less balanced then
MSmits: because a lucky start may still give you the win
MSmits: and p1 is still favored A LOT
MSmits: most starts are favored to p1
DomiKo: random position and play two games
DomiKo: one as P1 and one as P2
MSmits: steal is better for balance
MSmits: and gives less draws
MSmits: draws as win both players win 1 game with swap
MSmits: like it is now
karliso: exactly, but even if random each move I don't see how it would become less balanced.
MSmits: it would not become less balanced
MSmits: but it is less balanced than a version with steal
MSmits: because with steal, p1 is *forced* to pick the most balanced starting move
MSmits: he has to, or p2 will steal it
MSmits: so out of the 15 starts, the best one is chosen
MSmits: the one with near 50% winrate
MSmits: 4 4 has near 70% winrate for most bots
MSmits: in your case 100% even
MSmits: other starts have 55-60% and there are 1 or 2 that are near 50%
MSmits: there is currently a connect4 contribution with exactly the same problem
MSmits: I ran a meta mcts to test:
MSmits: http://chat.codingame.com/pastebin/4d136503-1e3d-41d2-8983-39e2c9d05a7e
MSmits: creator implemented steal and now we all start with the corner move, because it is most balanced
karliso: I will just say that I like the idea of random 1st move. It would enrich game with many more types of positions.
MSmits: I agree with games being more varied and interesting yes
karliso: Maybe connect4 can also do random 1st move?
karliso: Less hardcoding.. more depth..
MSmits: there are only 5 starting moves
karliso: Still :)
MSmits: I can hardcode the entire game no matter what
MSmits: I can fit around 200k moves in codesize
karliso: Can they not make the board larger? :D
MSmits: they did
MSmits: it's not solved this way
MSmits: but it's reasonably close to it
MSmits: if you took a month and a threadripper, you could do it i think
MSmits: traditional is 6x7, this is 7x9
MSmits: when a game is that close to being solved
MSmits: bots play nearly perfectly
MSmits: if the balance is off, its disastrous
MSmits: that's why the steal
MSmits: uttt is much further from being solved
MSmits: so it being off balance is less of a problem (yet still a problem)
MSmits: also, compare it with yavalath. It has extreme branching with 61 choices, yet the game is solved without steal
MSmits: some games just need it
MSmits: gonna go sleep, gn all
proogrmr: gn
Kitkat260: hello
NesteruS: Hi guys!
Check my new CoC music puzzle, please!
https://www.codingame.com/contribute/view/6035df0618299a1005f0bd4494932507a829
Annihilannic: trying to do coders strike back but the game replay videos don't seem to be playing, is it just me?
z64555: First Code4Life, not Code vs. Zombies. Before and After? #conspiracy_theories
z64555: wait, no. don't make a channel for that
No_one...: ello
Westicles: Junior high? For reals?
PatrickMcGinnisII: I know a 13 yr old designing and printing 3d models and lasercutting ... its insane man
PatrickMcGinnisII: well I was doing 8-bit asm @ 11
No_one...: wdym junior high for reals?
No_one...: Im 10
No_one...: :)
No_one...: altho i am in junior high
No_one...: but junior high is pretty normal for a programmer ngl
z64555: I think that was when I was messing around with BASIC on a C128 acquired from a garage sale
Westicles: That is a great age when you are highly gifted. When you get older, the world usually catches up
No_one...: :>
PatrickMcGinnisII: z80 was fun
PatrickMcGinnisII: kids today will never know the struggle
z64555: Walking to school, up hill, in the snow, both ways, huh?
Westicles: Never listen to a gen Xer complain. That was the absolute peak best time before or since to grow up
PatrickMcGinnisII: depends on the year. We were poor as sh*t in the 70s.
Westicles: Yeah, but I bet you could go outside on your own for hours at a time without parents breathing down your neck
z64555: really depends on the individual case. shouldn't make a generalization
PatrickMcGinnisII: My kid won't go outside.
PatrickMcGinnisII: In 79' I used the neighbors computer (that they didn't know how to use), I wrote code ON actual paper. smh.
PatrickMcGinnisII: Westicles u aint wrong
SaadMotamed: is there a way i can check my match history ? i wanna check the questions again
