Rockstar555: hii good morning
dbdr: damn, I pulled MSmits down to #3
YodaMaster123: In which Multiplayer?
eulerscheZahl: onitmaa
dbdr: ^^
eulerscheZahl: i see smits at #2
dbdr: now I beat karliso once, so Msmits back to #2 :D
dbdr: strange, because I made a perf twek but only measured 1% improvement
eulerscheZahl: and you already left the noob ranks where i'm stuck
dbdr: *tweak
dbdr: I was thinking, your eval sounds more expensive, so depth is not the end of story
dbdr: the grouping part, which I don't have
eulerscheZahl: not even sure anymore if i have it or wanted to have
struct: I checked RAIC leaderboard just now
struct: and the 1st guy seems very strong
struct: 50~-% winrate
eulerscheZahl: he won in 2017
Rockstar555: hii
eulerscheZahl: and did well in the google ants challenge
Rockstar555: What was the discussion going on
eulerscheZahl: http://ants.aichallenge.org/
greentae was 2nd
struct: google doesnt do AI challenges anymore does it?
eulerscheZahl: nope
Rockstar555: Mathis Lichtenberger this is the winner of this contest
eulerscheZahl: the user Schneewittchen (mostly active on #De) knows him in person
struct: Never saw im on #world
struct: But saw him on leaderboard on some multies
struct: at least I recognize the avatar
eulerscheZahl: it's from a fairy tale (both name and avatar)
struct: Now I know one more word in german
Rockstar555: hey guys reach gold league in CODE A LA MODE and Coder Strike back
eulerscheZahl: you also know the word "Zahl"
Rockstar555: :slight_smile:
Rockstar555: No
Rockstar555: can u tell
Westicles: and Feuerzangenbowle
eulerscheZahl: very good
Rockstar555: i know eulersche
Rockstar555: But not Zahl
dbdr: I know zahl but not sche
is that posessive, or diminutive?
dbdr: eurlerchen would be diminutive maybe :D
eulerscheZahl: if any, then posessive
dbdr: euler's right?
eulerscheZahl: and yes to the chen
dbdr: eulerchenscheZahl
Rockstar555: eulerscheZahl why u choose a Frog as a avtar
dbdr: your mick is too simple ;)
dbdr: little known fact, but Leonard Euler was a toad
Rockstar555: CODINGAME SPRING CHALLENGE 2021
1,993 CodinGamers have registered to this contest
eulerscheZahl: oh, i'm not even the only registered player from my company
Rockstar555: hey Guys this one is best contest in my codingame lifetime
eulerscheZahl: but iirc i kicked the other one here on the chat. and squiddy later banned him
Rockstar555: u are saying to me
eulerscheZahl: no, not everything is about you
Rockstar555: okk
Rockstar555: i have to invite also
PatrickMcGinnisII: https://imgur.com/a/fXZBFiT
**PatrickMcGinnisII apologizes to eulerscheZahl
PatrickMcGinnisII: Why is everyone approving Million digits of pi puzzle, i hate it
Rockstar555: u hate but i like
Westicles: If I knew how to do emojis, I would say shocked!
Rockstar555: it is intersting
Westicles: Thanks rockstar
Rockstar555: welcome
Rockstar555: :smile:
Rockstar555:
ABUSING THE SYSTEM
in this i see the Hard work of Westicles
Rockstar555: Good Job Westicles
Westicles: :blush:
Rockstar555: guys now seriously which not private chat we have to stop if any body have any topics so we disscussed about it
Rockstar555: :smile:
Rockstar555: it not private chat we have to stop if any body have any topics so we disscussed about it
Rockstar555: sorry my mistake
Rockstar555: :smile:
PatrickMcGinnisII: Last I knew GMP classes aren't included in the php functions installed/allowed on CG, so I'm not sure if I can circumvent and solve it
Radman_737: سلام
Rockstar555: yeah u can try it
Westicles: You don't really need GMP, it is just sort of helpful
PSA28: https://www.codingame.com/training/medium/lumen
PSA28: any one solved this ?
Rockstar555: yeah
Rockstar555: i solved
Rockstar555: this
Rockstar555: Westicles also solved this
PatrickMcGinnisII: there's plenty of big number puzzles I haven't been able to hack solutions for
PatrickMcGinnisII: well at least its not on easy list
Rockstar555: PatrickMcGinnisII it be right to solve it without hack
PatrickMcGinnisII: pretty sure shell_exec isn't allowed
**PatrickMcGinnisII chuckles
Rockstar555: if u start with easy puzzle then midium and hard then very hard
Rockstar555: ot be easy
Rockstar555: it be easy
PatrickMcGinnisII: trying to keepp my easy puzzles to 100% complete
PatrickMcGinnisII: ;)
Rockstar555: which programming language u learn PatrickMcGinnisII
Rockstar555: yeah this is right way keep it up
PatrickMcGinnisII: i know 30 or so, I'm an old man
Rockstar555: ohh
Rockstar555: nice
Rockstar555: u can reach 100% if u consentrate on it
PatrickMcGinnisII: rockstar, u know there is a lang called rockstar?
Rockstar555: no
PatrickMcGinnisII: pretty good video on youtube, talks about obscure langs
Rockstar555: ohh
PatrickMcGinnisII: https://www.youtube.com/watch?v=6avJHaC3C2U
Rockstar555: ok
Rockstar555: i watch
PatrickMcGinnisII: thought your name... might be relevant
Rockstar555: okk
Rockstar555: https://forum.codingame.com/t/a-muder-mystery-puzzle/188155?u=rockstar555
AlexTheGamer: how to do it
AlexTheGamer: im new
PatrickMcGinnisII: Rockstar555 there actually is a similar puzzle, i forget the name of it... it's something like, who is in the graden alone...or something
dbdr: hey MSmits!
dbdr: MSmits #1
dbdr: oops, won't be for long :D
dbdr: he's online but not in chat :thinking:
jacek: hmm?
jacek: :(
ParticleBeam: ?
dbdr: sorry jacek
jacek: ban
dbdr: top 3 seems random, and I stirred things up :)
dbdr: got lucky against you a couple of times I think
jacek: i wonder what MSmits uses in this game. it can't be booked
jacek: ahh, maybe you just got winning positions, that would explain a lot
dbdr: winning start positions?
dbdr: I think MSmits bitboarded the heck out of it, as usual ;)
NoobCod3: hey
eulerscheZahl: he mentioned that he needs 62 bit for the state iirc
jacek: he can compress entire universe in 64bit
jacek: https://i.redd.it/r5nif5jgje361.jpg
eulerscheZahl: a chess clock for that game?
jacek: do they produce connect4 clocks?
jacek: or c4 clocks for short
jacek: you havent watched queen's gambit?
jacek: so havent I :v
PatrickMcGinnisII: This is crazy: https://nanochess.org/chess4.html
Rockstar555: hii everybody
ChampionCoder: I just got to Wood 2 League in Coders Strike Back. Trying to use while loop for the new inputs Distance and Angle, always getting timed out. Can someone please share code
Thanks in advance
ChampionCoder: I know it's cheating, but I've been trying for over 3 days with no progress.
Rockstar555: hii
eulerscheZahl: there is a "reset input" at the top right in your IDE ChampionCoder
eulerscheZahl: click that to get some code that parses the new input
ChampionCoder: ok i'll try it
Rockstar555:
eulerscheZahl hii :smile:
YodaMaster123: Rockstar, avoid writing people`s nicknames exactly as they are so that you don't ping them.
Rockstar555: okk
suleyman-kutukoglu: öncelikle selamın aleykum
Rockstar555: what was happening in World Chat
Rockstar555: Server is so week
wirdes: oncelıkle alaykum selam
Rockstar555: wirdes 05:07PM
oncelıkle alaykum selam
plz don't talk this in the World Chat
jacek: oO
Rockstar555: It is for Group discussion not for this
Rockstar555: if u continue this u are Ban with moderators
Rockstar555: hii Jacek
Rockstar555: nice to meet u
jacek: ohai
Rockstar555: yeah what are u doing today
jacek: just chilling
Rockstar555: YEAH
Rockstar555: ANY TOPIC
Rockstar555: FOR FORMUS
Rockstar555: SO BUSY AT NIGHT
Uljahn: oO
jacek: here something christmas https://img-9gag-fun.9cache.com/photo/adVEPZQ_700bwp.webp
Uljahn: noice
sirsatan: ho ho ho hello
Rockstar555: yeah
Rockstar555: hii
Rockstar555: every one
Rockstar555:
jacek 05:17PM
here something christmas https://img-9gag-fun.9cache.com/photo/adVEPZQ_700bwp.webp
nice poster
Rockstar555: or photo or image
Rockstar555: whatever u says
Rockstar555: :confetti_ball:
Rockstar555: :evergreen_tree:
Rockstar555: yeah guys any body start a topic
Rockstar555: boring
Rockstar555: :thermometer_face:
Rockstar555: guys What about Covid-19
Rockstar555: :mask:
sirsatan: u tell me
sirsatan: :D
Rockstar555: yeah
Rockstar555: every body'
Rockstar555: every one is healthy
PatrickMcGinnisII: gheesh, spent all night trying to figure out why my code wasn't working
Rockstar555: in what
Rockstar555: PatrickMcGinnisII
PatrickMcGinnisII: first my gamestate tree wasn't deep cloning all the changes, so each leaf was operating on bad data
PatrickMcGinnisII: now i fixed that
Rockstar555: ohh
jacek: what game
Rockstar555: yeah nice question jacek
PatrickMcGinnisII: and now my backtracking is broken
PatrickMcGinnisII: There is no spoon2
Rockstar555: ohh
jacek: 2?
PatrickMcGinnisII: simple dfs
PatrickMcGinnisII: ep 2
Rockstar555: episod 2
PatrickMcGinnisII: but there's alot of object referencing going on
PatrickMcGinnisII: and the error_log buffer is never big enough
PatrickMcGinnisII: I gotta run it locally to see everything i need to see
PatrickMcGinnisII: *sigh*
PatrickMcGinnisII: ok, to bed
Rockstar555: It id medium puzzle
Rockstar555: It is medium puzzle
PatrickMcGinnisII: yeah well I had to put in alot of preprocessing because php is so slow
PatrickMcGinnisII: can't get deep enough with full tree
PatrickMcGinnisII: I can if i preprocess
Rockstar555: so choose another programming language
PatrickMcGinnisII: rofl, that's what everyone says
PatrickMcGinnisII: I should be laser cutting xmas presents
PatrickMcGinnisII: glhf, I get so into this stuff i lose time
PatrickMcGinnisII: 'morning...er 'g'nite
suleyman-kutukoglu: s
suleyman-kutukoglu: selamün aleyküm
suleyman-kutukoglu: banlayanın amk
Rockstar555: hii guys
eulerscheZahl: speak English suleyman-kutukoglu
Rockstar555: yeah
wirdes: okey i speak english
wirdes: how r u all
Rockstar555: couldn't counderstand
suleyman-kutukoglu: sometimes i cant speak perfectly english so generally i use turkish
wirdes: +1
Rockstar555: what was this language
eulerscheZahl: but we don't understand Turkish at all
suleyman-kutukoglu: but if it is problem for everybody i can speak english
Rockstar555: yeah
Rockstar555: i am going to use translator
brkn: you can't speak english, you can write in English.
Rockstar555: :laughing:
wirdes: axlikfşlsdakfsd
wirdes: brkn right
suleyman-kutukoglu: sadsakşdjsa
Rockstar555: stop it
eulerscheZahl: by this point i'm tempted to kick. you've been warned
Rockstar555: hey eluerscheZahl what u do at tonight
eulerscheZahl: i'll be sleeping then
Rockstar555: any contribution from your side
Rockstar555: :smile:
Rockstar555: Smart answer
eulerscheZahl: not before RAIC ends
wirdes: rockstar55
wirdes: sdlkfjalksdf is mean haha in turkish
Rockstar555: okk
Rockstar555: stop it
eulerscheZahl: looks more like hitting random letters on your keyboard
Rockstar555: this is too much
Rockstar555: speak english otherwise use translator
Rockstar555: hey eulerscheZahl CODINGAME SPRING CHALLENGE 2021 it is starting at may 06 2021 is it so far
eulerscheZahl: yeah, you have to be patient
YodaMaster123: Is the name yet to be decided? I noticed that most contests are named after movies.
eulerscheZahl: the last 2 have no special name anymore
YodaMaster123: Oh... Spring and Fall challenges 2020
AntiSquid: you can call them Pacman and Spice Girls, most suitable names YodaMaster123
eulerscheZahl: only you say spice girls
AntiSquid: only one calling it as it is
Gabbek: Hello everyone :)
AntiSquid: greetings
eulerscheZahl: :wave:
Gabbek: Could anyone explain to me why it is invalid move: '52' in FreeCell?
https://www.codingame.com/replay/514043421
suleyman-kutukoglu: we cant open this link
AntiSquid: click share replay again
eulerscheZahl: Oups
An error occurred (#UNAUTHORIZED): "You are not authorised to view replay 514043421".
Gabbek: Seems like supermove combining 3-8 cards, but it's invalid for some reason, hmm
Gabbek: https://www.codingame.com/replay/514043421
eulerscheZahl: you can't share replays when the puzzle is not approved
Gabbek: oh, I see
AntiSquid: make a video
AntiSquid: perfect opportunity for a stream
eulerscheZahl: a screenshot will suffice
Gabbek: I'm wondering what will happen - there's one left day for validation. I wanted to thumbs it up, but so far I haven't solved it :(
AntiSquid: trictrac made hanabi ? O__O
AntiSquid: why call it fireworks
eulerscheZahl: CodinBot will refuse the contribution. but the creator can put it up for approval again
Gabbek: https://imgur.com/a/OfoW5Qd
Gabbek: seems like supermove to me, moving 3-8 rank cards from 5th pile to the 2nd one
eulerscheZahl: you want to move 5 cards at once
eulerscheZahl: but only have 1 free spot at the top and 1 cascade
eulerscheZahl: that allows to move 4 at once
eulerscheZahl: moving blocks is just for convenience
eulerscheZahl: you can move single cards only
Gabbek: ohh
eulerscheZahl: move 1 up, then the 2nd to the casecase
eulerscheZahl: 3rd up
eulerscheZahl: 4th somewhere else
eulerscheZahl: then revert the intermediate steps
sirsatan: @AntiSquid 'Spice Girls' is great
Gabbek: I missunderstood it, that explains the issue very well :D
Gabbek: thank you very much eulerscheZahl
AntiSquid: i know SirSushi, but these people don't want to admit it
Romil: Hi Automaton2000
Automaton2000: not sure what i did to get to the top of the leaderboard
AntiSquid: sirsatan * how many sir somethings are there
sirsatan: today im SirSanta ;)
eulerscheZahl: i see satan
Gabbek: hehe
sirsatan: use def Anagram(sirsatan): ;)
Gabbek: anything interesting going on lately? I've been slacking really hard for the last 2 weeks or so.
sirsatan: I wonder is there any problem to save my code after press 'PRESS MY CODE' button
AntiSquid: uncooperative version of hanabi is WIP Gabbek
eulerscheZahl: we have the first 3D puzzle pending now https://www.codingame.com/demo/862009d303ad01e86e3f0537f622e3764fe182
sirsatan: today my changes disappered
AntiSquid: Wolf 3D when?
eulerscheZahl: when you make it
Gabbek: sirsatan never had similar issue
Gabbek: wow, that's so cool :o
sirsatan: i got 2-3 times for last month
eulerscheZahl: it happens sometimes that your code disappears
eulerscheZahl: only submit will store it permanently
sirsatan: next time i will copy to text file
eulerscheZahl: or just submit
Gabbek: I think my next goal will be to go through pending puzzles and check them out
eulerscheZahl: lots of clashes there
eulerscheZahl: i think the questmap is partially responsible for that
Gabbek: I haven't touched any clash, so I guess I'll just check out other puzzles. But before I do that I want to solve this FreeCell, need to change few things because I'm not handling auto-play correctly
Gabbek: How long does it usually take to approve a new puzzle?
Gabbek: I guess it can take a while, right? 3 approvals required, people have to solve it (or I guess that's optional?), hmm.
eulerscheZahl: the fastest approvals are around 10min
eulerscheZahl: the longest is about 4 months i think
eulerscheZahl: neighbor sum grid was pending really long
Gabbek: adding blockout as next goal, after freecell, hopefully I'll be able to provide some feedback on the next weekend.
LastRick: It takes a while because I suspect the pool of users that can and actually do solve the "classic puzzles" is small and they're the same users active (and otherwise busy) on the other parts of the site.
Rockstar555: HII
Gabbek: Good day
LastRick: morning
Kamal-Kumar-Jena: hiiiii
jacek: oh, someone pushed karliso up :unamused:
MSmits: which game is this?
jacek: onitama
MSmits: ahh ok
ParticleBeam: Amazing how far one can get with random moves
MSmits: which game?
MSmits: still onitama?
ParticleBeam: I am trying that at the moment
MSmits: it's a great game
MSmits: it's chess-like
MSmits: except the pieces don't have different moves, the cards do
eulerscheZahl: i like the asymmetric aspect that the cards give. X can capture Y but Y can't do the same with X
MSmits: yeah that's a funny part
MSmits: it's pretty easy to solve though.
MSmits: except there's too many card combos to make use of this
eulerscheZahl: and 50ms isn't that much time
MSmits: yeah I meant offline
eulerscheZahl: i got that
MSmits: I am pretty sure you can solve any combo in less than a day on standard hardware
MSmits: there's so few possible boardstates
Hannnn: gualgualagual
MSmits: I think it's like 10^12 or 13 or something
Hannnn: hey wtsup guys
eulerscheZahl: let's make a new contribution with a larger board :P
MSmits: nah, this one's fine
eulerscheZahl: just to ruin your bitboards
MSmits: I wasnt able to abuse this feature and I would have
MSmits: if I could
MSmits: would sure ruin my bitboards yeah :)
MSmits: it fits so close
MSmits: I use 31 bit lol
MSmits: err 62 that is
MSmits: maybe a larger version would neatly fit in 128 though
MSmits: like most games
MSmits: btw, I am doing a game design course atm, as a student I mean
karliso: I coded a stronger version of my onitama bot somewhere in July, but didn't submit coz my current version was still beating you guys (I think). You have given up in the game already?
MSmits: I have
MSmits: doesnt mean i will never come back to it, but there's soooo many other games
MSmits: probably at least a year before i get back to it
eulerscheZahl: sounds like blasterpoard he declares having a stronger bot for Kutulu than the current #1
karliso: MSmits Which is ur favourite atm?
eulerscheZahl: that's a tough question
MSmits: karliso do you mean which game I like most, or which game I am currently working on?
karliso: currently
MSmits: mmh, well I havent been working on cg much lately, but I have a list of things i want to do
MSmits: i need to get a bandas feature to work
MSmits: D&B has to be finished
karliso: oh, I have not looked at bandas long time. Does my current version still beat you?
MSmits: checkers 3 piece end game db and meta mcts
MSmits: barely karliso
MSmits: probably if i submit 50 times, i will sometimes get nr 1
karliso: I added some endgame to bandas as you suggested, it helped.
MSmits: it certainly does
MSmits: did you go as far as 3x4?
MSmits: I think that's the absolute limit
karliso: I need to look that up.
MSmits: because it's 12 squares and 13 doesn't exist
MSmits: 14 would be 2x7 and require 10x as much space
karliso: yes, I do have 4x3
MSmits: cool we have the same db then
karliso: But...
MSmits: unless you did a fractional db for 4x4 or something
MSmits: or you did some fancy compression
karliso: My 4x3 is coded for one...
MSmits: what does that mean?
karliso: Like there were flippings and rotations
MSmits: ah yes
karliso: But I don't think I implemented all of them
MSmits: ahh ok, I managed to get them all in. It's pretty messy to do though
MSmits: with all the rotations and mirrors
MSmits: I think the rest of your bot is just better than mine
MSmits: so even if my db is a little more complete i still lose a bit more
MSmits: did you do mcts with heuristics?
MSmits: mine is just plain mcts
karliso: I do some stuff I learnt from uttt
MSmits: sure, but that could be anything hehe
MSmits: I always learn stuff between games and try them
MSmits: I noticed you went full on teccles on uttt, cutting the heuristic off at ply 20
MSmits: for both p1 and p2
karliso: well
jacek: try teccles on bandas :v
MSmits: :P
karliso: There is no need to switch for p1.
MSmits: you mean theres no need to stop the heuristic?
karliso: And I think iti s not 100% as p2.
MSmits: hmm i did see some instances where the heuristic was bad for p1, in late game
MSmits: but maybe if you play better, this never happens
karliso: yes, I think you can always win as p1 if you follow this heuristic.
karliso: Meaning you gonna win if the rest of your moves are ok.
MSmits: seems likely
MSmits: but hard to be sure
MSmits: thats why it is not good to have an opening book as p2
jacek: sounds like a challenge
MSmits: yeah it is a challenge and one I have tried a few times
karliso: jacek Is your onitama bot mcts or minimax?
MSmits: so far karliso seems to be right
jacek: jacekmax
MSmits: karliso in other words, UCT guided minimax, backpropagating eval value
karliso: yeah
MSmits: not adding, but replacing highest value
MSmits: ahh kk you are aware of it then
karliso: what?
MSmits: jaceks favorite search
jacek: its world renowned
MSmits: jacekmax
karliso: oh, I misread, I have no idea what is that.
MSmits: you expand like mcts
jacek: consider it MCTS with EPT 0 depth
MSmits: then eval all children
MSmits: and backpropagate minimax style
MSmits: use uct to choose what to expand
jacek: whats your eval in onitama, karliso
MSmits: jacek thats what i use on onitama, but i add up scores, you replace them
jacek: is it complicated
karliso: well
MSmits: mine is pretty complicated, have a bunch of features with params.
karliso: It is bugged and not very good
karliso: not the best part of my bot
MSmits: what do you think is best?
MSmits: without giving it away i mean
jacek: well my bot is pretty slow
karliso: I saw onitama bot on github where it did eval of peaces.
karliso: I don't have that.
karliso: I have coverage scores instead/
MadKnight: hey MSmits what are u doing ?
jacek: peace :v_tone1:
MadKnight: raic ?
MSmits: no, dont have that much time MadKnight, just fixing small stuff about bots
MSmits: karliso coverage score means you can capture opponents using cards?
karliso: My onitama bot could be improved a lot, but I need motivation.
jacek: i was 1st but dbdr pushed you up
karliso: A += (int64_t)popcount(COV) + (int64_t)popcount(COV & 0b00000'11111'11111'01110'00000);
MSmits: karliso, jaceks bot is stronger than yours atm
MSmits: so you have motivation
dbdr: :angel:
karliso: Is it?
MSmits: yes
jacek: hmm https://github.com/maxbennedich/onitama/blob/master/src/main/java/onitama/ai/evaluation/PieceSquareTables.java
MSmits: I did tests a while ago
MSmits: my bot is similar to yours in strength i think, maybe slightly weaker, but jaceks is quite a bit stronger than mine
MSmits: jacek didnt you train a network for this as well?
jacek: ntuple
MSmits: thats pretty much a network
MSmits: in the sense that you do the same things in training it
jacek: yes
MSmits: except you give it patterns in advance
MSmits: instead of letting it discover them
karliso: Oh, I did nt notice jacek submitted in october.
MSmits: by the way jacek, you may have to retrain your oware soon :P
jacek: :scream:
jacek: more endtables?
MSmits: opening book now beats every NN except yours as p1. I beat it as p2 already with 60%
jacek: good, i need motivation
MSmits: wait till i submit, then retrain it
jacek: as karliso knows, being 1st is a curse
MSmits: should be nr 1 again
MSmits: the funny thing about opening books is i can keep growing it. I wonder for oware, if I can keep growing it until you run out of meaningful variations
MSmits: the game is so simple, that there arent so many different moves that are good
MSmits: so the branching for an opening book vs a good bot is very tiny
MSmits: maybe 1.3 or so
MSmits: your random is what's slowing me down. I had agade and robo done as 100% win in days
MSmits: recur se also has random
MSmits: but the bot makes more mistakes than yours
jacek: good thing the winter is coming, you can save money on heating by doing metamcts
MSmits: :)
MSmits: it's a little boring though, for this game
MSmits: I have no clue why moves are good and dont see the strategy thats happening
MSmits: for uttt it's easier even
MSmits: I can look at yavalath meta mcts all day, it's very fun
MSmits: the game is just more strategic
MSmits: as far as a human can tell anyway
123456789LL: Hi
struct: CPU doesnt like my code :'(
MSmits: as long as you like it
MSmits: self confidence
struct: https://i.imgur.com/fErcvHy.png
struct: very bad
jacek: you speculate too much
MSmits: i do ?
MSmits: oh struct
MSmits: well i do too
MSmits: it says bad speculation
MSmits: ohh I see, that means your ifs are screwy
struct: yeah, this happens when i place a piece
struct: and on floodfill
MadKnight: say no to branches
MadKnight: say no to conditional jumps and to variable jumps
MadKnight: ezpz
MadKnight: write code with no ifs
MadKnight: if(stuff) a+=x
just replace it with a+=x*(stuff==true)
jacek: no ifs? so use switches
MadKnight: no conditional jumps jacek
reCurse: If optimizing was that easy
reCurse: It doesn't necessarily mean it's bad
reCurse: Something to look into but any program has to be bound by something
MadKnight: ifs only in top-level parts
MadKnight: like main and reading inputs
reCurse: Ifs everywhere until proven otherwise
reCurse: Caring too much about it is as bad as caring not enough
AntiSquid: if (win.found()) break;
MadKnight: struct was asking that, not me
MadKnight: struct really likes to optimize stuff
reCurse: Yeah but your suggestion can sometimes be misleading
MadKnight: i was kinda joking about no ifs at all
reCurse: Ok then sorry it wasn't obvious
MadKnight: u can kinda say that's a top-level if, AntiSquid
jacek: use structs, not classes. more lightweight
MadKnight: in c# ?
MadKnight: it just makes them value-type jacek
VizGhar: Hey guys... Just looking at the UTTT, how many simulations are you able to run on first 2 runs with MCTS? I'd like to have some "benchmark" beforehand
struct: my bot is not great but I have 20-25k+-?
struct: rollouts
ProCrazyCoder: oh no
ProCrazyCoder: i m so bad at programming((
VizGhar: thanks struct :) - let's dive into it
struct: and that number is for 2nd turn only
struct: on first turn i just play on middle if im player 1
VizGhar: Yeah... maybe some "observed" moves from beginning might help, to reduce tree size and still drawing good moves... I just have to start with regular MCTS
VizGhar: wonder how far can I get... quite new domain for me
VizGhar: Also, yesterday I've seen someone mentioning pragmas for C++ I'm mostly Kotlin dev, that knows how to write basic C++ code. Where do I find how to use these pragmas?
VizGhar: ideally directly with CG
struct: https://www.codingame.com/forum/t/c-and-the-o3-compilation-flag/1670
struct: usually I use these
struct: #pragma GCC optimize("O3","unroll-loops","omit-frame-pointer","inline")
VizGhar: god bless you :D Now I hopefully have all i need
jrke: what is black color int in java basically for CG sdk
struct: what do you mean jrke?
jrke: like i wanna do
sprite.setColor(Black);
struct: you want the int that represents black color?
jrke: so what should be the value of Black
jrke: yes - 'you want the int that represents black color?'
struct: 0
VizGhar: 0x000000 = 0
jrke: oh thanks
struct: white = 0xffffff
jrke: is there any website where i can find many colors int?
struct: https://www.color-hex.com/
jrke: thenks
struct: just replace the # by 0x
VizGhar: Material design colors (some nice selection of colors)
https://material.io/design/color/the-color-system.html#tools-for-picking-colors
reCurse: Generally try to avoid pure white and pure black
struct: Usually I try to replace white by light grey
struct: But Im not a designer so Im not sure if its good or not
VizGhar: And if you are talking about sprites right away, why not to use some shiny nice colors? :) red/green/blue/yellow/orange/purple
reCurse: Doesn't have to be super gray, as long as it's something like 240-255 should be fine
reCurse: Also same applies to colors, try to stay away from programmer colors (anything with 255 and 0 in components)
reCurse: *240-250
jrke: i am doing it first time i will improve it
reCurse: wait *240-245
reCurse: ugh
Presac: Are there only 4 code golf's?
struct: yes
jacek: + clashes :v
DiscoFrenchman_9bfc: hallo
jacek: ohai
dbdr: eulerscheZahl: https://www.codingame.com/multiplayer/optimization/cgfunge-prime/leaderboard
check it before Japan wakes up ;)
eulerscheZahl: :D
eulerscheZahl: congrats, well done
eulerscheZahl: did you follow my advice?
wlesavo: what a battle :slight_smile:
dbdr: eulerscheZahl: no, deterministic
eulerscheZahl: that makes it even more impressive
eulerscheZahl: poor xevi
dbdr: I have to tame him
dbdr: resistance if futile
dbdr: *is :D
dbdr: https://www.solvoyo.com/wp-content/uploads/2018/03/Futile.jpg
jacek: meow
eulerscheZahl: futile or fertile? :thinking:
Westicles: you can do random in cgfunge?
eulerscheZahl: if (random) print("PRIME")
else print actual solution
eulerscheZahl: with very little randomness (maybe 5%) you still have a chance to pass all testcases and cut off a few points from the total score
Westicles: I see, very sneaky
dbdr: if someone starts it, others will too, and the shorter deterministic will still win I think
eulerscheZahl: true
eulerscheZahl: but if only one does it and the others didn't even think about it...
eulerscheZahl: well, now the idea is out in public
dbdr: :D
dbdr: your fault or mine?
eulerscheZahl: 50-50
eulerscheZahl: i just said "advice" here in public and you used "deterministic" in the reply
eulerscheZahl: then i gave it all away
Westicles: If xevi starts on samegame I'll tweet her a hint
jacek: her?
eulerscheZahl: you can share that hint with us
eulerscheZahl: or do you mean hint about random to save your same game lead?
Westicles: yeah, the latter
eulerscheZahl: :D
PeaceMoreZ: hi
jacek: good evening
eulerscheZahl: how are your ATM studies going? do you learn how to spit out money?
DomiKo: hint to SameGame O.o
jacek: o.O
PeaceMoreZ: this site needs to be more popular
PeaceMoreZ: just sayin
PeaceMoreZ: :d
ProCrazyCoder: hi
jacek: oO
BogiBruh: yoo is this the discord or is it the site wide chat
BogiBruh: ah site wide chat
BogiBruh: still epic
DrDrunkenstein: someone from italy here?
PeaceMoreZ: kaney
PeaceMoreZ: how did you solve the last clash
CursingLlama: I find shortest code rather stupid. Never gonna win that against python using c++.
MSmits: then dont use c++
MSmits: the idea is not that every language should be equally good
MSmits: a lumberjack will not complain that his hammer does not take down the tree as well as the axe does
JBM: or the chainsaw
MSmits: or that
JBM: where's that doom multi
MSmits: people have been doing 3D, who knows
AntiSquid: you're reading chainsaw man manga too JBM ?
JBM: errrr no
AntiSquid: for a moment i thought you're cool .
jacek: what is it about?
Maxim251: Helo, I am trying to figureout pikaptcha detective 2, and I have question about how long takes to you to finnish practice games?
Maxim251: I am now working on it 3 days
AntiSquid: it's a story about a boy and his dog, jacek
jacek: like ed...ward?
Maxim251: normally what average of time is to finish particular games?
Maxim251: Is few days to write code for medium practice game to finish it? Like one week?
struct: for puzzles?
Maxim251: yes
AntiSquid: you mean edward scissors hands ? well he can retract the saws back into his arms like wolverine
struct: They are meant to take time
struct: all depends on your skill level
Maxim251: and one week to make it is normal?
jacek: depends on puzzle, could be hours if you have experience with the algorithm. 1 week isnt bad
Maxim251: Almos all games in easy level I have finished with 100%, and now i make medium ones, and my average is one week. or shorter.
AntiSquid: well considering contests usually last 10 days, i guess you have to do it in 10 days Maxim251
Maxim251: I was worry that I am not so good ifI take so much time.
AntiSquid: are you talking about puzzles or multiplayer? confused .
Maxim251: puzles
AntiSquid: ah ok
struct: Just take your time.
Maxim251: Thats good..
AntiSquid: no rush for puzzles, they only give XP anyway
jacek: well, some hard puzzles were easier for me than medium ones
struct: Im sure after you do a puzzle if you have a similar problem
struct: then you will be able to solve it faster
Westicles: That is an easy puzzle, it shouldn't take so long
Maxim251: Thats good, Thank You.
Westicles: You can study the other solutions
struct: rollercoaster and the surface should be medium I think
Maxim251: Yes, on easy puzles it take me not longer that week. Like 3 days, 4 days.
struct: On rollercoaster if you solve it in javascript you dont even need to optimize it
struct: bruteforce works
struct: How long have you been coding for?
Maxim251: I am nooby, I learned c++ 2 months ago, and codingame is greate for exercise.
AlgoN: yeh
Westicles: Oh, in that case you are fine
Maxim251: Good to know.
struct: In 1 year you will solve the easy in under 10 minutes
Maxim251: Thats why I come with that cuestion
struct: It all depends on your experiencey
struct: Hard to compare 2 months vs years
Maxim251: I see on Youtube other programists that they was typing straight from head and that was looking so easy
struct: Also what Westicles suggested can be helpfull, check other players solutions
struct: After you solve a puzzle you can check them
Maxim251: I do, I do :)
Maxim251: Always when I finish puzzle, I check first 15 peoples how they do. Sometimes I fill embarased when I see how short code they have. :)
AntiSquid: some people love to code-golf everything ...
struct: Short doesnt mean good
Maxim251: yes, I see, but I don't feel so strong to do code-golf... But I feel better knowing that I do, as begginer that I am not so bad with time to finish puzzle.
Maxim251: I think I know what You want to say. I notice code that they get data straight from input and they passing to output in one turn. like not using any container to store data,
AntiSquid: code-golf isn't strength, it's madness
Maxim251: I like store all data from input, and then work with it.
Maxim251: Thanks for answares guys :)
jacek: :tada:
therealbeef: code-golf can let you look at a language in new ways. you can learn stuff that could be useful down the road. But I think beginners shouldn't focus on that; it can lead to bad habits.
jacek: but your coworkers will love making the review of your code
Gabbek: JBM hello, do you have a moment? :)
JBM: depends
Gabbek: I'm trying to solve FreeCell, I have one or two questions if you have a minute
JBM: shoot
Gabbek: I'm having big issue with autoplay, let me show you a screenshot, as replay doesn't work until mission is approved
AntiSquid: https://i.stack.imgur.com/qvvKX.gif
Gabbek: https://imgur.com/a/44nXig0
Maxim251: ok, thanks
Gabbek: 2 of spades gets placed on foundation, but 3 of spade doesn't - and my sequence of moves tries to make it into a situation where 2-6 of spades will get placed on foundation, but... I'm confused when I'm supposed to output my move to home and when it will be filled in automatically with auto-play
Gabbek: I manually added "3h" to my moves sequence and then it gets much further
JBM: i see
JBM: for now the agme implements the ms rules (to the best of my understanding of it)
JBM: "Microsoft FreeCell or FreeCell Pro only plays an available card to its homecell automatically when all of the lower-ranked cards of the opposite color are already on the homecells (except that a two is played if the corresponding ace is on its homecell); aces are always played when available."
JBM: from http://solitairelaboratory.com/fcfaq.html
JBM: there's a change in the plan to make autoplay optional, which might make your life easier
JBM: i'd also like to find a way to make it more tolerant to ambiguous moves, but I don't have a clear path for that one yet
JBM: it *is* still best practice to interrupt early games that took a wrong turn
Gabbek: I totally understand that, I just couldn't understand why it wasn't being autoplayed, that explains it quite well for a noob like me :)
JBM: ok
JBM: "aces and twos are autoplayed when possible" is a good summary
Gabbek: I doubt I'll be able to finish it tomorrow, but I'm more than happy to work on it more once you'll resubmit it again - I would like to approve it once I'm able to solve it, as it's really well done
Gabbek: finish it today*
JBM: great!
Gabbek: I've never played freecell before, so the note about aces and twos will be most likely helpful for others as well
JBM: thanks for taking a look
Gabbek: it's a pleasure, the least I can do for so much work put into it :)
JBM: yes i don't mention autoplay details too much yet in the statement
JBM: i intended to dive deeper when I'd actually give the option to swich modes
Gabbek: I think I've stumbled upon this issue earlier, but I just didn't realize it as I had much worse debug, so I didn't really know what was going on
Westicles: no plans to add undo, right?
JBM: no
JBM: i'd lol, but then...
JBM: is it useful?
Westicles: It would probably make it much much easier, so not a good idea. But good for us dumb guys
JBM: i'll think about it ;p
Gabbek: "much much easier" - this made me curious, is the intended way of doing it by using available moves provided as input and nothing else?
JBM: well you can attempt an exhaustive search that way
JBM: but i don't think there's much chance of it succeeding in the allowed turn limit
Gabbek: okey, I see. I couldn't find a use for provided available moves, but it's useful for debugging purposes and verification of your own availalbe moves
JBM: when the referee doesn't crash, it's an easy way to know in advance a move will succeed
JBM: (the referee doesn't crash anymore-AFAIK)
Gabbek: that's a good point
JBM: but that can really only be of help at the start
JBM: i suppose when you get to the point of solving it you've "simply" converged with the referee's idea of valid moves
Gabbek: it's useful at the start for learning the rules and validation - that's for sure :)
struct: Finally I think I finished it
struct: https://www.codingame.com/replay/514190010
struct: I dont think I will optimize more
Maxim251: I have very importand question, i s about pointers
Maxim251: http://chat.codingame.com/pastebin/ec3f2635-5399-4a7a-b027-cc4e56d00565
Maxim251: Why this code is workin???
struct: Why is it working?
struct: Well you are passing the pointer to the function
Maxim251: Yes, function dont use pointers,
struct: char* betty
struct: Its a pointer
Maxim251: Passing betty by reference should work only using with pointers, but this one don't use that.
Maxim251: betty[1] = 'd';
Maxim251: betty should have pointer, yeah?
struct: When you call passbyreference(betty);
struct: you are sending the first address of betty
Maxim251: When I use betty as normal char { *betty = d } it works, but { *betty[2] = 'd'; } dont work
struct: You want to change the value right?
Maxim251: It shouldn't be like thiss ? passbyreference(&betty);
Maxim251: yes.
struct: No because berry is an array
struct: betty*
therealbeef: my advice: use std::array instead of 'raw' array, and references instead of pointers
Maxim251: And when I use array, then this char container is acting diferently?
Westicles: You should never use pointers unless you are porting code by a person who foolishly used pointers
struct: Averaged around 1.4mill in that game
struct: moves per turn
Maxim251: Thats wired property, I am learning code 3 months, but never in tutorials I didn't see this version
Maxim251: And I agree that pointers are bad, and i should use them only in special case. In one game I was forced to use pointer, to prevent function to output array of char, and this one property schock me.
Maxim251: Heh, QTCreator, they use pointers at level of insane :D
therealbeef: Qt is not always nice, in my experience
Maxim251: But I see that many companies are using Qt, and to be able find job in future, Qt should be good choice....
therealbeef: basically, the 'name' of an array can be assigned to a pointer and the resulting pointer will point to the first element. [1] adds 1 and dereferences
struct: [1] == ptr + 1
Maxim251: this is in side of function
Maxim251: void passbyreference(char *betty){
betty[1] = 'd';
}
struct: So what should that function be doing just to see if I understand
therealbeef: [1] = *(ptr+1)
Maxim251: and in side of fuction, I use betty like normal array.
Maxim251: But goot to know that wired property... In future I hope that it will not suprise me if I will do similar thing in future. Thanks for reply therealbeef
Maxim251: #struct The cod e should manipulate array of char, but there is wired property of a pointer.
Maxim251: http://chat.codingame.com/pastebin/24a461d2-0d92-4b54-98d0-a4def0b58dc8
struct: it is manipulating, I dont see anything weird
Maxim251: #struct, play with it little bit. To see how acts
Maxim251: For me it is, first time I see this type. Nobody on Youtube teachers didn't mentioned about that property.
therealbeef: betty is of type char[4], which the compiler can always convert to type char*
therealbeef: so the betty inside the function is exactly the same betty as in main
Maxim251: But when I do betty as char betty then I need use asterix always
Maxim251: http://chat.codingame.com/pastebin/c9a29fdf-f6b9-498e-99e1-60e900393743
therealbeef: (except that the length of the array is lost, but that is a nother story)
Maxim251: No, I was able to add betty[2] = 'd', and output in main function that?
struct: yes, it should be like that
Maxim251: Aha.. Thank You for responce. That's very importand lesson for me to day.
Maxim251: Thanks guys. :)
therealbeef: betty is type char[4]/char*, so &betty is char**, but that can be assigned to char*, so betty inside the function is a pointer to a pointer to char(s), so you need to peel of one pointer-layer with *betty and then [1]
therealbeef: pointers are evil
therealbeef: if you use std::array<char,4> you get a nice object and pass to the function by copy or by ref with std::array<char,4>&
Maxim251: Ok, I will test it too.
therealbeef: succes
Maxim251: #therealbeef I hate pointers from begginning of learning, and future artickles about it give me this same thinking like You about pointers in program. Few days ago, I see on cppcon explanation about how procesor works and how compiler make 'comands' for procesor, This person say, that compiler is enough smart and coders dont have a reason to use pointers to make program faster. Actually, there was no boost of speed.
therealbeef: I agree
Maxim251: I also remember Bios Stromburg, -< Sorry, I know I spell his name terribly wrong, He is father of c++ and he said that He dont understand why peoples are using poiters in that way, if pointers was created for diferent porpouse, and it should be avoided ass much ass posible, and pointers was created for use in function, to prevent big data be copied twice
therealbeef: Bjarne Stroustrup
Maxim251: Yes, Thank You
therealbeef: the idea of C++ is that you get all the nice abstractions like classes without any performance impact over C
Maxim251: :sweat:
Maxim251: He also was hoping that C users will move to C++ as next generation of coding, but somehow didnt succeed. I think it is because of Obiect orientation. Even for me its bit hard to us ethem, but I hardly exercise them. :)
Maxim251: I see that Obiect oriented is very usefull to group bunch staff... Only I need train my self a lot more
Maxim251: Thanks and have Nice Day
nsrsn: asd
Maxim251: I need say that.. I hate that many peoples are using 3 or 4 languages and they use ass c++, Its like mixing styles from C or C# .. Now I now hor recognise that styles, but at beggining of my learning, this types of styles, make my hard days to understand, it confused me a lot
AntiSquid: nsrsn asfjkickdsh
nsrsn: can i hide this chat section?
struct: click on bottom arrow
struct: or add ?disableChat to end of url
nsrsn: when i click bottom arrow i see only Join Us On Discord button
struct: https://i.imgur.com/MwRkuaS.png
AntiSquid: nsrsn add ?disableChat to your codingame URL
nsrsn: lol thanks @struct
nsrsn: that was ez xD
nsrsn: ... and cool :)
struct: Site slow?
therealbeef: church
struct: I think I'm going to report it
struct: Its always at same time
therealbeef: 1am?
therealbeef: cet
struct: yeah
therealbeef: eu+asia is asleep, usa not woke yet, good time for maintenance actions
struct: yeah, but is not usual
therealbeef: actually usa not asleep yet either
struct: yeah its a good time for usa atm
LastRick: maybe something is set to run during the middle of the night in france. Europe is sleeping. But yeah it does seem to hit every night in the evening (U.S. time)
Sergen: it is the same time rankings update
LastRick: there you go
struct: Yeah, but it never slowed the site this much
LastRick: I've not been around that long. This is a new thing?
Sergen: poor scaling?
struct: This slow is recent
CaliforniaQT: probably crypto miners
struct: Has been happening for around a week
CaliforniaQT: jk
LastRick: maybe its me. Ever since the contest I've been playing CSB like crazy :)
CaliforniaQT: this is a great simulation for using dev servers at tech companies
CaliforniaQT: :)
CaliforniaQT: what is csb
LastRick: coders strike back
CaliforniaQT: i would look it up but site is too slow
CaliforniaQT: ah thanks
lopidav: how to report a bug?
lopidav: oh, forum
struct: forum or discord
lopidav: thank you
Rockstar555: hii
MadKnight: yea hi
Rockstar555: how are u
Rockstar555: hii good morning
