Chat:World/2020-10-16
doduyphuong: :expressionless:
DaNinja: you print the number of the mountain to shoot
SINGHALBROTHER: https://www.codingame.com/servlet/urlinvite?u=4009074
MadKnight: what is that link for
pluieciel: Hi!
pluieciel: one question: clojure macro doesn't work here in IDE, right?
pluieciel: (defmacro prefix [exp]
(list (second exp) (first exp) (last exp)))
pluieciel: (println (macroexpand '(prefix (5 * x))))
pluieciel: Standard Output Stream: (prefix (5 * x))
pluieciel: in my computer:
pluieciel: my-test.core> (println (macroexpand '(prefix (5 * x)))) (* 5 x)
Shrimpster: Some random Asian kid just followed me, should I worry? :rolling_eyes: (he's my first follower haha)
magaiti: he liked your name, no doubt
Shrimpster: hahaha thanx, that's probably it
Astrobytes: if you don't know him, simply don't follow back Shrimpster
Shrimpster: I would never, he just followed and I thought it was weird (^.^)
Astrobytes: Yeah, it happens fairly frequently unfortunately
Shrimpster: ok good to know
Remembered_Goodbye: hey guys, how do you do the ascii art puzzle
Remembered_Goodbye: nah screw it ill search it online cause dead chat
Kamatcho: Damn, trying out python for the first time
Astrobytes: Nice to see you waited less than a minute before making that decision.
Kamatcho: It's quite hard.. feeling dumb here xD
Astrobytes: If you're totally new to python (and programming) codingame *might* not be the best for practising Kamatcho
Astrobytes: It assumes a basic working knowledge
Kamatcho: Which I don't posess
Kamatcho: What might be better? I've done some lua programming earlier
Astrobytes: Hm, you could try freecodecamp? I think that's supposed to be decent for learning
Astrobytes: Or there's always youtube video series and what not, depends how you prefer to learn
Kamatcho: And then again mostly by trying to solve things
Astrobytes: freecodecamp should work then I reckon
Astrobytes: Then, when you're a bit more comfortable you can come back here and solve some puzzles etc
Kamatcho: Might check that out before I go in here doing puzzles and feeling stupid lol
Astrobytes: Yeah, it's advisable. So you can focus on just solving the problem rather than trying to figure out the language :)
Kamatcho: Thanks for the advice!
Astrobytes: No problem at all
Sylarb91: Astrobytes... does freecodecamp have something for c# too, cuz i just looked and couldnt find anything ?
Sylarb91: i know the basics of c#... but there is always room for improvement
Astrobytes: Sylarb91: No I don't think so, it covers JS, python, HTML & CSS, some other web-stuff only I believe
Astrobytes: Not sure where's best for C# tbh
Ramdeath: hi all
Sylarb91: ok. thank you
Astrobytes: hi Ramdeath
geppoz: hi
geppoz: there is a way to collect all the scores of the tests?
geppoz: without looking each one, but getting as a table?
magaiti: you can make a js script for that
magaiti: i dont have one ready available
magaiti: but it should be as easy as 3 lines
MadKnight: geppoz u can run custom js scripts on web pages
geppoz: after i play all testcases i find result in some variable?
geppoz: or you mean launching test with my js?
GaethofsQ: Is there any more documentation on coders strike back? I'm looking for the formulas used to define the speed, acceleration, etc... of the drones
KiwiTae: check the forums
Uljahn: or CSB landing page, there are some links to external resources
735487: geppoz: I have one. it show all test case results and sum it up
735487: var sum=0;var i=1;$('.testcase').each(function() {
var a = Number($(this).attr('title').split(':')[1].slice(0, -1)); console.log(i++ + ": " + a);if (i>2)sum+=a; }); console.log(sum);
735487: I used it for 2048 game. that is why it skips first test result
ooh: is there a way to erase submissions in history?
magaiti: why would you want to do so?
Unkmar: Are you ashamed of your progress?
magaiti: progress is a process of trial and error, and there are shameful errors
ooh: i'm ashamed of pressing submit on the same code
Shrimpster: there's no shame in that, happens to everyone
magaiti: yeah, i for once, want color tags for my submissions, and maybe even editable descriptions
Astrobytes: magaiti: https://github.com/Azkellas/cgenhancer
magaiti: hmm, intersting stuff. but no
Astrobytes: It has editable descriptions
magaiti: ok ill look into it someday. thanks
Astrobytes: np
wlesavo: it would be quite good if enhancer wouldnt reset all the descriptions from time to time
Astrobytes: agreed
eulerscheZahl: does it still break the UI for puzzles?
Azkellas: Never heard of this issue wlesavo
MSmits: mmh for some reason c++ doesn't like it if i set a 100 million sized uint8_t array to the value 3
MSmits: I'll just set it to 0 before run time and change it to 3 during runtime when needed
MSmits: works.... dunno why
Alshock: xD That's a bit weird
MSmits: yeah
MSmits: I think it's some kind of optimization
Alshock: you have an initialization list of 100 million {3} ?
MSmits: you cant do that I think
MSmits: if you set it to 3 that way its just the first element
MSmits: the rest will be 0
MSmits: i was using a function to set them all to 3 at the start of the program
MSmits: but i think the compiler noticed it and decided to do this before runtime
Alshock: Ik, that's why I was asking if you had an initialization list with 100 million threes, wouldn't actually write them all down ;P
MSmits: which took too long somehow..
MSmits: oh lol, no.
MSmits: awww the first symbol of my endgame book is a ✻
MSmits: damnit, i finally finish my size 12 endgame book for Bandas and I crash my VS because the string is too big
YurkovAS: MSmits do you know the original name of the game Bandas?
MSmits: nope
jacek: bandas isnt original? D:
Dan4Life: \o/
Dan4Life: |
Dan4Life: |
jacek: wtf? how can i reply to a forum post
asuna321: i kinda suck at this
eulerscheZahl: use the "reply" button
jacek: i clicked reply, typed text but i cant see any send or ok button
eulerscheZahl: screenshot?
eulerscheZahl: is there a bar at the top?
eulerscheZahl: if yes, close it
jrke: oh trailer out for fall challenge thats nice
jacek: https://i.imgur.com/tjaYopF.png
jacek: ahh, thx euler
eulerscheZahl: and close that windows bar as you are on it
jrke: yeah euler is right
jacek: windows bar?
jacek: i only use windows for games, so i dont bother customizing browser much
jrke: euler have you registered now,i mean you were having problem in registration of fall challenge is it fixed?
eulerscheZahl: was fixed minutes later
eulerscheZahl: maybe they broke it again, no idea
jrke: oh thats to quick
eulerscheZahl: 9,194 registered
eulerscheZahl: let's check again in a few minutes
jrke: yes will cross 15,000 this time i think
eulerscheZahl: 2 more registered now. seems to be working
jrke: yeah
jrke: i am off now gn bye bye :)
Neumann: eulerscheZahl: is it me or does 2048 have the same issue as SearchRace regarding timeouts when you use more than 40ms ?
eulerscheZahl: i didn't even check the time i take per turn, don't know
MadKnight: Neumann people on #ru were saying they also had the same problem in both
Neumann: Ok :/
MadKnight: amurushkin
eulerscheZahl: that's disappointing :(
Apocrypha_blank_01: Question. Does your clash rank update even if you play privately with friends ?
jacek: ahh so thats why im not #1 at 2048 :v
MadKnight: Neumann could it be related to some kind of late timer stopper in the SDK ?
eulerscheZahl: Apocrypha_blank_01: no
Neumann: Now I need to figure out why my BeamSearch sucks
Neumann: MadKnight: maybe, hard to tell without any backend information
MadKnight: like it takes 15ms to read all moves that u print + proceed them
MadKnight: try splitting your entire solution in 600 turns
Neumann: Meh, I don't think so. Even with a really under-optimised engine, processing a hundred instructions shouldn't take that long
MadKnight: yea but reading from stdin might be slow
Neumann: It's all on one line
MadKnight: maybe just try splitting in 600 lines ?
MSmits: man... VS goes mental, it's only 24209 characters
MSmits: one one line I mean
jacek: why dont you do it normally in VS and the unicode stuff only in the ide
MSmits: cuz you hadn't suggested it to me yet
jacek: its obvious
MSmits: also, how do i play CG bench matches this way? I tried putting in these unicode strings, but it doesnt take them
MSmits: something to do with encoding of the text file
jacek: i dont use cg bench
MSmits: mmh ok
MSmits: oh btw, what i did so far, was use a different program to produce the string, output it to console and then copy paste it into the other program. For some reason this inserts random spaces
MSmits: only very few, but still
Schwase: anyone here do "the lucky number" puzzle before?
Schwase: its the one with 6s and 8s
jacek: i use python to do te encoding
MSmits: mmh ok
jacek: how big endgame can you solve in reasonable time?
jacek: if not for the 100k limit
MSmits: hmm
MSmits: I think using an endgame db, I can solve full maps
petr_step: What is your favourite language?
MSmits: with 3x3 as the biggest, I already solve at ply 9-15
MSmits: if I make a bigger db offline and run the algo on the full map, i can solve it in a few mins to hrs probably
MSmits: depending on the map
MSmits: I do this with mcts so it's not the most efficient solver
MSmits: probably could be done faster with a proper negamax
MSmits: but the database I hope to get working I should be able to have endgames of 2x6 or 4x3 instantly solved also. It's just a really big jump. From roughly 30k states to nearly 2 million
MSmits: it fits easily in 24k unicode characters, but having some problems debugging
jacek: 24k unicode? thats just 48k normal
MSmits: i could use ascii encoding sure, but thats not really easy either
MSmits: you cant use the full 8 bit, it's 7 bit
MSmits: and i dont really get the extended ascii or understand whether i can use it here
MSmits: might be i can also fit it with base64 even
MSmits: but then the string gets long also,
jacek: mmh ok
MSmits: I'm thinking the unicode range has some more gaps that I am running up against
MSmits: I know there's a 2k gap in the high ranges that I am skipping
Nerchio: hey guys is there any UTTT local referee for all levels i think i could find only for 1 board
Nerchio: or i don't know how to change it to higher level :D
YurkovAS: Nerchio readme has uttt referee https://github.com/dreignier/cg-brutaltester
YurkovAS: run command: java -jar ./cg-brutaltester.jar -r "java -jar -Dleague.level=2 tictactoe-referee.jar" -p1 "./uttt-1" -p2 "./uttt-2" -t 10 -n 1000 -l "./logs/"
Nerchio: it says my player does invalid input when it clearly doesn't :P
Nerchio: my player can play boss in the uttt referee but when i try to play myself it doesnt work idk
Nerchio: ah i guess it had a problem with reading the same file so i created a copy xd
Nerchio: nvm it doesn't really work when i play against myself doesnt make any sense ;p i guess i will just leave it
tobk: So, from now on all challenges will be called "[Season] Challenge [Year]"? Kind of dull...
Astrobytes: they switched to 2 official contests per year so I guess that's the way they want it :neutral_face:
tobk: I don't have a problem with the two contests thing, and even using that name for the contest, I'd just wish they would re-brand it when it's released for MP
Astrobytes: valid point
Majeck: Wait, so no more awesome names like "Coders Strike Back", "A Code of Ice and Fire", "XMAS RUSH", "Code a la Mode", etc?
Majeck: Not kidding, the names were one of my favorite parts hahaha
Astrobytes: Sure seems that way yeah
Majeck: :/
Majeck: Woah Google is registered for the fall contest cool!
Majeck: Amazon too
MSmits: all it requires is some players that work there
Astrobytes: ^
MSmits: doesn't mean jeff bezos made a decision :)
Astrobytes: lol
Majeck: Oh you're right hahaha I thought it was a company decision lol
jacek: meh imagine deepmind playing contest
MSmits: it would get claustrophobic and we'd beat it
jacek: whats this called? https://img-9gag-fun.9cache.com/photo/aQdNKg8_460svav1.mp4
jacek: would make a nice puzzle
Astrobytes: no idea, seems nice though, yeah
codeing: hi
codeing: for this puzzle
codeing: https://www.codingame.com/ide/puzzle/count-of-primes-in-a-number-grid
codeing: look at this 100%
codeing: http://pastebin.fr/69139
codeing: any feedback ?
Allis: It's super-verbose. :/
Astrobytes: almost unreadable
codeing: what verbose ?
codeing: like ?
Scarfield: many words
hiljusti: "wordy"
codeing: comments.
hiljusti: it can be positive, like "very well explained"
Scarfield: VerbosestroBytes
codeing: was using java
Astrobytes: Wordfield, hello
jacek: very bad for codegold though
jacek: golf even
PatrickMcGinnisII: codeing isprime function not optimal
mcjordan2002: https://www.codingame.com/clashofcode/clash/1423060a6c70aa460f22c1045e8f858d4865387
Astrobytes: mcjordan2002: post in #clash
jacek: oO
Astrobytes: mcjordan2002: you can ask for participants in here but keep links to #clash or PM please
codeing: not optimal like ?
PatrickMcGinnisII: only have to loop to sqrt of n, not n
codeing: mhm
jacek: anyway, whats the problem. does it work always?
codeing: sure. the code is ok
codeing: 100%
jacek: so, whats the question? :thinking:
MSmits: to be or not to be
codeing: to be for me.
codeing: to be => to code
Dan4Life: to be not to be
magaiti: 2B | ~2B
MSmits: ||
MSmits: or it would be bitwise
PatrickMcGinnisII: or you could be holding a pipe
Dan4Life: 2B || !2B
Dan4Life: or 2B | ~2B
MSmits: https://en.wikipedia.org/wiki/Shakespeare_Programming_Language
Dan4Life: negation doesnt work ~
MSmits: it's the way the chat shows the tilde
Astrobytes: it's a font issue don't worry
MSmits: it doesn't do squiggly
MSmits: You know what's annoying. The community games usually dont have swap, but some games are really lopsided in favor of p1
MSmits: Bandas is 70-30 when I test it. Then I submit and i get 7 games vs karliso of which he is player 1 6 times
Astrobytes: Yeah, and no way to change it as the creator
MSmits: time for some spam !
MSmits: I finally have my huge engame book working
Astrobytes: for Bandas?
MSmits: am in ballpark with Nagrarok and karliso now
MSmits: yeah
MSmits: it has almost 2 million endgame states with L/W/D values
jacek: and loops are D?
Astrobytes: didn't know you were making one lol, I should just assume you always have/are working on an endgame db for any game you're currently playing
MSmits: apparently there are no loops that are affected by the turn limit unless the turn is beyond 190 or so
MSmits: so that means i can just look them up, loops arent always draw
MSmits: depends on how many dudes there are
MSmits: of p1 and p2
MSmits: what i did was retrograde analaysis for every state, from turn 200, backwards. Then I tested if any state has a different value for any turn between 1 and 190 and they did not
MSmits: but a state at turn 195 or so may have a diffeent value than it has at turn 190, so you cant look up the states at the very end
MSmits: this is very different in oware btw. You have to always take the turn into account
MSmits: (to solve I mean, obviously not to win in most cases)
MSmits: earlier I accidentally forgot to reset my time at the beginning of my turn, which meant I only used calculation time in the first turn. The rest of the game it would do at most 1 expansion and then stop. I still got rank 11, mostly because it used the reused tree from turn 1 :P
MSmits: was kinda hard to figure out though, I thought I messed up my unicode stuff again
MSmits: jacek, as for your loop question. This game was solved at frame 29:
MSmits: https://www.codingame.com/share-replay/493501704
Astrobytes: I usually forget to start the timer and wonder why it times out
MSmits: it shows as value 0 solved
MSmits: yeah
MSmits: this one is way more extreme: I solved it as a loss at frame 14. It turns into a loop with tric trac having 2 more dudes
MSmits: https://www.codingame.com/share-replay/493501556
PatrickMcGinnisII: ass https://www.codingame.com/replay/493502975
MSmits: yw :P
MSmits: was solved at frame 12
MSmits: this is why dudecount is a bad evaluator for this game, you had way more guys, but they were nearer the edge
PatrickMcGinnisII: umm, i only looked at my 4 moves... nothing deeper...so....lol
MSmits: yeah I know, this game heavily favors performance too
MSmits: like uttt
PatrickMcGinnisII: I still in top half of leaderboard, so more than half aren't even trying
PatrickMcGinnisII: so, you're saying goal to make game last longer may be more beneficial?
PatrickMcGinnisII: or end in your favor i guess
MSmits: hmm
MSmits: in my case it doesn't matter, is solve very early. Basically, either the game is lopsided (many maps are) or you just make the right decisions the first few turns
MSmits: I lose to lower ranked players sometimes, mostly because of the maps
MSmits: the mcts node value will be -0.4 or something, which is really bad and that shows that no matter what you do, your position is crap
Metanoob: when your algo thinks its right but it's actually having a seizure: "Found: abdarbdacaaadadabdarb... Expected: abracadabra"
MSmits: lol
Metanoob: a bit like Automaton2000
Automaton2000: but i just don't know what i'm doing
Metanoob: sry mate
PatrickMcGinnisII: hmmm, not over my head. But I'm always in the situation of determining the overhead execution cost of iterative deepening/pruning or just dropping a heuristic rule. Bandas is one of those like you said like uttt, where sure you can gen 20k nodes, but if you try to prune and go deeper you lose 25% of the total nodes you can explore
MSmits: true
Metanoob: in those kind of challenges, timeout counts a lot, right? Does that mean some languages have the upperhand slightly?
MSmits: depending on which languages you compare, it might be a lot
MSmits: C# and java, small difference, C# and C++ large difference, C++ and python HUGE difference
PatrickMcGinnisII: I put an opening book together for UTTT, and it helped alot. I didn't properly mcts after that ... and left a basic heuristic in place because I didn't have the time to fix the MC.
MSmits: yeah that would certainly help, if the book is good
Dan4Life: quick question, why is shortest mode a mode in clash of code considering its very inconvenient for C++, C# and Java Users?
MSmits: easiest is to just watch top players and copy their moves
MSmits: use the right tool for the job Dan4Life
MSmits: it's not a goal of CG to make all languages equivalent
Dan4Life: So i should learn Ruby then?
MSmits: python should really be enough
PatrickMcGinnisII: i didn't run the sim for long enough locally to really refine the book, and I couldn't get more than 50k board states
Dan4Life: oh, ok
MSmits: 50k moves in the book?
Astrobytes: feel free to learn Ruby if you want ofc
PatrickMcGinnisII: i think I ran it for 2 days. I'd have to look it over, but 50k was the goal
PatrickMcGinnisII: boardstate X, best move =.
Metanoob: makes sense yeah MSmits, thanks :)
MSmits: thats pretty awesome PatrickMcGinnisII, it's not that easy to fit this many moves in the codesizes
Astrobytes: don't forget Rust in that speed comparison
Metanoob: oof yeah
MSmits: I am guessing you stored a tree PatrickMcGinnisII, as opposed to fully defined states?
MSmits: that way you only need to store each move
PatrickMcGinnisII: http://chat.codingame.com/pastebin/0b94ce08-42c2-46c2-b241-1d1bdc888b4f
MSmits: doesnt work
MSmits: bugged
Astrobytes: still broken
PatrickMcGinnisII: damn, still not fixed
Astrobytes: use pastebins from poland like MSmits
MSmits: yea :)
Astrobytes: :P
PatrickMcGinnisII: pastbin.pl wouldn't load
PatrickMcGinnisII: lol
MSmits: pastebin.pl
PatrickMcGinnisII: A Database Error Occurred
Astrobytes: you must have broken it
Astrobytes: All these years it had no traffic, then suddenly: BOOM
Metanoob: haha
Astrobytes: I always liked hastebin
Metanoob: thanks for sharing, didn't know it
PatrickMcGinnisII: my lookup https://hastebin.com/erajuvazaz.pl
Metanoob: yup, perl confirmed
Astrobytes: Steganography challenge!
PatrickMcGinnisII: it's compressed, but it is what it is
rohanjain: do we also get difficult questions in clash of codes ?
Astrobytes: Difficulty is subjective rohanjain
Metanoob: ^
Astrobytes: If you feel they're too easy try some puzzles instead
rohanjain: so as we increase in level, the difficulty increases ?
Allis: No, that wouldn't make sense for Clash.
Allis: You don't get to choose your opponents.
Astrobytes: No, I mean it's a personal experience. You may find it easy, others may find it hard.
Metanoob: also PatrickMcGinnisII I've had to do that so many times too, most frequently with exporting encoded images to data:base64 in js because of canvas tainting
Metanoob: s/encoded//
PatrickMcGinnisII: if size isn't an issue md5 works
rohanjain: if after coming after first in clash of codes, if we leave in between, would the progress still show or it won't ?
Astrobytes: rohanjain: puzzle section is here, varying difficulty levels https://www.codingame.com/training
PatrickMcGinnisII: the first turn having a decent amount of time, decrompressing the book isn't a problem
Astrobytes: Yes, it will show rohanjain
Metanoob: ah ye I forgot about the first turn, nice one
Metanoob: Hi JimmyJames
rohanjain: https://www.codingame.com/clashofcode/clash/14231499760341acf5c3367e78ddde92abbdae8
Astrobytes: rohanjain: there's a channel for clash links - #clash
PatrickMcGinnisII: MSmits I think your terniary works better than what i did
Astrobytes: You can ask in here but keep links to private msg or #clash
Astrobytes: It's OK don't worry
MSmits: PatrickMcGinnisII when i write an opening book i never store gamestates
Astrobytes: plies + best moves no?
MSmits: I store a tree made of moves. In the first turn, my bot turns that tree into a map with states
MSmits: it looks like this:
Astrobytes: aha gotcha
MSmits: 4 { 3 5 { 6 } }
MSmits: i play 4, best response is 3, if opponent plays 5, i play 6
Astrobytes: elegant
PatrickMcGinnisII: I just was running into issues with addressing the correct branch/node
MSmits: btw, i never bothered compressing an opening book
MSmits: a single move requires so much calculation time that i never hit the limit
MSmits: my oware book looks like the example i just shared, including the spaces and such. It's still only a few kb
MSmits: despite a run time of months
PatrickMcGinnisII: so 2 bits per position x 81 + 5 bits for the move... umm compresses into ... crap I forgot
Astrobytes: small pecan brownies
Astrobytes: :P
PatrickMcGinnisII: yea gotta be crack brownies at this point
Astrobytes: hahaha
MSmits: for example, if you do oware, you basically need values between 0 and 5 for the move and each move needs information for the follow up (sibling, go deeper, or go up)
MSmits: so 18 possibilities
MSmits: means you can fit 3 moves in a single unicode character
MSmits: so I am guessing 200-300k moves is realistic
PatrickMcGinnisII: MSmits but don't you end up with alot of empty spots in the data?
MSmits: why empty
PatrickMcGinnisII: cause the game would be over
MSmits: how so?
PatrickMcGinnisII: i guess it's how you are referrencing them
MSmits: oh usually unordered map
MSmits: this doesnt have to be fast
MSmits: i just do one lookup per turn anyway
Einwickler: MSmits you there?
MSmits: the conversion from text-tree to unorderedmap takes some time but its ok
MSmits: didn't you see me spamming Einwickler?
PatrickMcGinnisII: langs get caught up trying to find the right index in an array if it doesn't have a mathematical or incremental 'method'
Astrobytes: He's not Einwickler, it's an imposter, don't believe the hype
MSmits: it's not an array PatrickMcGinnisII
Einwickler: yeah I just noticed my chat was scrolled up :sweat_smile:
MSmits: unordered map is a hash table
MSmits: well i am giving up spamming, karliso's bandas bot still beats mine
MSmits: i'll be happy 2nd
Einwickler: Didnt want to interrupt you guys. sorry for that :stuck_out_tongue:
MSmits: np
Astrobytes: I just noticed the time, I'm off. gn all :)
PatrickMcGinnisII: ahhh, cg doesn't support SplHash for php
MSmits: what did you want to ask Einwickler ?
Einwickler: cu astro
MSmits: gn Astrobytes
PatrickMcGinnisII: and everytime php makes an object the memory overhead is just stupid
MSmits: yeah, some languages are like that =/
Einwickler: I just implemented basic minimax for uttt and was remembering your advice of not going into the arena
MSmits: ah right
MSmits: does it always win in wood?
PatrickMcGinnisII: so your bandas finally finished computing
MSmits: cuz it should if you did it right
Einwickler: Im pretty sure its working as intended now. Anything I should test before going to next league with 9x9?
MSmits: PatrickMcGinnisII i submitted many times, a few times 2nd, sometimes lower, but never 1st
PatrickMcGinnisII: pretty darn close, gratz
Einwickler: Against the wood boss at least, yes :)
MSmits: Einwickler you can try to add ab pruning. It's not particularly hard to implement. Takes a while to wrap your head around why it works though
PatrickMcGinnisII: i looked awhile back you were hovering around 7
MSmits: ah yeah, that was before the end game book
PatrickMcGinnisII: so, what did karliso do I wonder
MSmits: i always wonder that. In uttt he has a 3k line bot and I have no idea what to write about uttt in that many lines
PatrickMcGinnisII: unrolling some loops can gain a few ms...maybe i guess
MSmits: usually the compiler does that and it barely matters
MSmits: definitely not why his bot is better
MSmits: probably he has some smart heuristics
Hjax: did you guys see that Rémi Coulom (author of crazy stone) took #1 on Dots and Boxes O_O
The_Auditor: what? RickRoll ?
jr7vn: i come from vietnam, and you?
jr7vn: because your name is GameBot
GameBot: When did you join CodinGame?
jr7vn: please review my website, i have compliti it . this address is: www.htblog.tech
Metanoob: nice jr7vn I don't understand vietnamese but I google translated your "about"page :)
Unkmar: Hello jr7vn. Life good?
thanhhv317: Hi