Chat:World/2020-06-19
Reksio: Thank you for your answer. This is good to hear!
Reksio: I have two more questions for whomever knows and would be kind enough to answer.I also wonder if there is a way to get an error output after submit.
anid: No that is not allowed for puzzles Reksio
anid: You may print the validators and hence it should not be allowed
Reksio: oh, thanks. It's all right I suppose
anid: but in the case of bot programming, it is accessible from the replay
Reksio: thanks
Reksio: The second question. I did some dump of Console.ReadLine and WriteLine and it looks like this: http://chat.codingame.com/pastebin/cb58630d-8d37-4ebb-8a42-527ad08a5865
Reksio: http://chat.codingame.com/pastebin/0284f5d4-974e-46f1-9c9c-c01e9eddedde
Reksio: http://chat.codingame.com/pastebin/aa9876a9-2db2-40d1-90a9-77ebf20785b3
Reksio: Why it shows this link...
Reksio: http://chat.codingame.com/pastebin/f2faef03-bb34-431c-860c-90c881950418
anid: If you paste over 5 lines, it goes to pastebin
Reksio: Ok, sorry for spamming but I'm trying to get rid of the link thing
Reksio: ah
anid: NP
Reksio: Read: 6715 microseconds Read: 2 microseconds ... Write :5 microseconds
anid: So if you print too many error messages, it only shows the first and last
anid: otherwise it would be too many lines
anid: I think there is a char limit
Reksio: Read: 6715 microseconds Read: 0.2 microseconds ... Write : 0.5 microseconds
Reksio: Actually this is correct. I don't have ...
Reksio: I just didn't want to post too many lines
Reksio: I wonder if it is normal for first read to take so much time
Reksio: I specifically talk here about the descent
Reksio: it's the early puzzle
Reksio: I'm trying to create my own tests but I wonder how game treats reading of input for 6.7 miliseconds
Reksio: while the algorithm itself takes < 1 microsecond for example
Reksio: or are those slowdowns only in IDE mode but after submit there are some guarantees about execution time of ReadLine and WriteLine
Reksio: I guess I just wonder how is execution time being calculated.
Reksio: Is number of reads and writes always set in stone and time it takes to run them always counted as fixed value (possibly 0) ?
tutubalin: Some times input is like that: 1st line: number N Next N lines: blablabla
tutubalin: So input length may vary
dbdr: execution time is not calculated, it is measured with a clock
Reksio: I'm trying to say that ReadLine which is provided to us can take x100 more time to execute than our own algorithm.
dbdr: you should start measuring after you first readline, not before
Reksio: which seems ridiculous and hopefully is accounted for somehow by removing the impact of the input/output on final "speed" score (or some solution of sort)
dbdr: otherwise you are measuring things that are not your code
dbdr: readline can "take time" because it is waiting for input
dbdr: but that's not counted as your time by the referee
jacek: ^
dbdr: v
Reksio: ok, but if input time is not counted then how is solution which reads line 1000 times scored against solution which reads line only 10 times. In real life the solution which waited for input 1000 times would be much slower.
dbdr: think of it from the point of view of the CG code
dbdr: it sends the input, then it starts its clock and waits for your response
dbdr: don't worry about the time readline takes, it's negligible
dbdr: (once the input is present)
Reksio: thank you. I will treat it as such then. This helps quite a lot!
dbdr: :+1:
jacek: for puzzle does time really matters except for timeouts?
jacek: wether it is 20ms or 50ms
tutubalin: Reksio the very first ReadLine may look pretty long time (from your code perpective) but referee doesn't take it into account (it;s zero for it). So start you clock after first ReadLine - that will make measures much more realistic.
tutubalin: Also take into account that time functions are not accurate, so if timeout is 100ms, try to fit in 95ms
Reksio: Is it just 1st read line or all read lines and write lines? The write line takes whole 5microseconds which is longer than my whole algorithm takes for the whole game (0.5microsecond) if you substract input/output from it.
tutubalin: just first
Reksio: So other readlines + writelines which can take x5 amount of time of my algorithm will be counted to the score from what you say. Then why not all of them? Is there anywhere CG actually describe all of this?
dbdr: time does not influence score
Reksio: really? Is it just timeout based?
dbdr: yes
Reksio: oh, interesting but good. Is it also true for bot competitions? Some people are complaining that C++ has advantage for CG puzzles but with just timeout as score I don't see how...
dbdr: they were speaking about multis, not puzzles
dbdr: time matters when you want to run a search to optimize something, and run your algo as many times as possible in the allowed 50ms
Reksio: makes sense
Reksio: So I just need to clarify as I get mixed responses if all input/output (readline/writeline) is cut from time I spent running my puzzle or just the first one in the whole game or the first one for each loop cycle.
Uljahn: i think the first one in the whole game (common time limit is 1 sec) and the first one for each loop cycle
Uljahn: ofc you can check it yourself
tutubalin: don't measure time of the first read of every loop cycle
tutubalin: because it takes much longer due to switch from another bot
Reksio: thanks all for your answers
jacek: darkhorse64 do you have some bias in your rollouts?
darkhorse64: For Othello, no at the moment
jacek: my early results with moves at corners trice as likely as others have much better winrate
darkhorse64: That makes a lot of sense because taking corners is almost always a favorable move
darkhorse64: I am trying to write a boss for Othello based on my best bot. I have nerfed the perfs, removed some features and it performs better :nerd:
jacek: nerfed bot is better?
dbdr: lol
dbdr: a top 1 bot is born
darkhorse64: Yes but it still rank 6. The good news is that by removing code I probably fixed a bug in my bot. The bad news is that I won't get promoted against my own creature if struct creates a league with my boss.
CodinBotChatRelay: Pratyush#9392: ok
darkhorse64: *it is*
CodinBotChatRelay: Pratyush#9392: @CodinBot leaderboard coding
CodinBotChatRelay: Pratyush#9392: @CodinBot clash
CodinBotChatRelay: Pratyush#9392: @CodinBot contest
CodinBotChatRelay: Pratyush#9392: ok
CodinBotChatRelay: Pratyush#9392: hi
CodinBotChatRelay: Pratyush#9392: bye
CodinBotChatRelay: Pratyush#9392: hello
CodinBotChatRelay: Pratyush#9392: how are you
CodinBotChatRelay: Pratyush#9392:
- angeremote:
AntiSquid: spam? @_@
AntiSquid: could do that in a private chat you know
Astrobytes: darkhorse64: did you fix your skip nodes issue?
Astrobytes: AntiSquid: isn't that Benjamin's account for when he can't log in
AntiSquid: yes, but should he test his bot on the main chat?
Astrobytes: No I agree, just checking you knew it wasn't some random crap
AntiSquid: i am just thinking about this as what if everyone did the same thing
Astrobytes: Well, still random crap ;)
AntiSquid: what if everyone posted clash invites
Astrobytes: No I agree
darkhorse64: Using struct expert inputs helped a lot yes
Astrobytes: Nice one :)
AntiSquid: expert inputs?
AntiSquid: extra inputs?
Astrobytes: No, expert
Astrobytes: Well, bthe really :)
Astrobytes: *both
Astrobytes: 'If you output "EXPERT followed by the move"...'
AntiSquid: aha
Astrobytes: Just for all opponents moves and passes from the previous turn
AntiSquid: would be really useful in many multis, less of having to reverse engineers opponent's move
Astrobytes: Yeah it didn't occur to us before we approved, was only when darkhorse was having trouble with pass moves nodes in his MCTS we figured it would be a good idea, so struct very kindly patched it for us
darkhorse64: Yes, fog of war for opponent moves is not handy
Astrobytes: euler has it in Vindinium too
darkhorse64: We should have guidelines for board games. "Opponent moves must be in the inputs". It's actually more useful than game state because you wlll have to write an engine anyway
Astrobytes: That's a fair point actually
AntiSquid: make it retroactively so XR has it too :p
jacek: and oware
CodinBotChatRelay: Proven Angel#3425: http://c-help.tk - STUDENT PAPER ASSIGNMENTS DONE BY ACADEMIC WRITERS
Astrobytes: ehm
Astrobytes: AntiSquid
raulet: Hi, new here, Whats "the Boss Validator" ?
CodinBotChatRelay: Proven Angel#3425: http://c-help.tk - STUDENT PAPER ASSIGNMENTS DONE BY ACADEMIC WRITERS
CodinBotChatRelay: Pratyush#9392: What the heck are you all doing here
CodinBotChatRelay: Pratyush#9392: ??????
CodinBotChatRelay: Pratyush#9392: I will kick you then have fun
raulet: hello
jacek: boss validator? where
struct: hi
raulet: https://www.codingame.com/training/easy/a-mountain-of-a-mole-hill
raulet: I don't know what kind of procedure i'm expected to do to pass that.
jacek: every puzzle has test cases in validation that are not available in IDE to avoid hardcoding solutions. you can't know them
jacek: this probably means your solution has some bugs
raulet: they should give a clue or something on what expected to do or use, stuck at 87%, and don't know what I have to fix ; ;.
Astrobytes: They usually use edge cases or test efficiency. Make some custom test cases and make sure your algo can handle them properly
Astrobytes: e.g. large garden with lots of enclosed fence areas and lots of molehills
raulet: I will try that, thank you.
Astrobytes: Perhaps use "How did THAT happen?" test case as a starting point for your custom case
Astrobytes: Also "Include and infiltrate" because of all the nested fenced areas
tutubalin: your goal is to implemented what is specified in Puzzle Statement Test cases are just a way to test your solution (that's why they are called tests). So watching tests and trying to cover only cases described there is wrong way.
tutubalin: In real life part programming some test cases are provided by QA, but a lot of test cases are created by developer themselves.
Astrobytes: Hey tutubalin, it's nice to see the #ru users being more active in #world recently :)
eulerscheZahl: you could just go to #ru if you miss them
Astrobytes: My Russian isn't so good ;)
eulerscheZahl: they are more relaxed about english speaking intruders than our french friends
Astrobytes: :D
darkhorse64: I use google translate. Funny results sometimes
Astrobytes: It's not too bad for Russian mostly
Astrobytes: I wouldn't rely on it to translate my English (or French for that matter) to Russian though
Uljahn: do you miss MadKnight? i guess he got offended by being compared to a bot :(
wlesavo: Astrobytes you have to be working on uttt to join #ru though :slight_smile:
Astrobytes: The idea of Maddy being so offended that he left is a wonderful irony
Astrobytes: lol wlesavo, seems so!
wlesavo: this uttt fever going for a while, i got caught by it even in self isolation
Hjax: wait, you guys scared away madknight?
eulerscheZahl: sometimes i had the impression that cyberpunk used google translate
Astrobytes: Yes, for some things I think he died
eulerscheZahl: he was really understandable here on world (general back then). but his private messages were glibberish sometimes
Astrobytes: *did, lol
wlesavo: and died as well
eulerscheZahl: nah
eulerscheZahl: he just left us :(
Astrobytes: He pops up in #ru discord with random ML related things now and then
eulerscheZahl: but ru to en works surprisingly well
eulerscheZahl: i never tried ru to de
Astrobytes: afk
wlesavo: Ich denke, Russisch zu Deutsch wird schlechter sein, da es viel weniger Benutzer gibt
tutubalin: ja
wlesavo: tell me if that makes any sence :smiley:
tutubalin: Basically German is closer to slavic languages than English
Hjax: ? isnt english a germanic language
tutubalin: also many words came to Russian from German
eulerscheZahl: this sentence makes perfect sense wlesavo
eulerscheZahl: was about to ask you if you speak German :D
wlesavo: :smiley:
tutubalin: German is closer to English than to slavic. But it is closer to slavic than English to slavic
eulerscheZahl: i read an interview of a google lead engineer for translations 1 or 2 years ago
jacek: omelette du fromage
eulerscheZahl: he's a fellow German and disappointingly said that german<->english doesn't work as good as he'd like
eulerscheZahl: machine learning approach, just feeding lots of texts and translations of books
tutubalin: gurken - ogórek - ogurets
tutubalin: eulerscheZahl probably they parsed it as words
tutubalin: so they couldn't find a proper mapping for Donaudampfschifffahrtselektrizitätenhauptbetriebswerkbauunterbeamtengesellschaft
jacek: a cat sat on your keyboard?
wlesavo: tutubalin there are tools for feeding whole sentences i believe, no need to parse it
Astrobytes: "Association for Subordinate Officials of the Head Office Management of the Danube Steamboat Electrical Services"
Astrobytes: Google version: " danube steam shipping electricity main plant construction subordinate company"
jacek: try deepL
wlesavo: russian version looks like translation from english as well
Astrobytes: jacek, deepL just gives the same text back
tutubalin: i think that most of emails of this company cannot pass spam filter
Astrobytes: :D
NASADeveloper102: How do links work in Ghost in the Cell
Astrobytes: factory1 is linked to factory2, factory2 is linked to factory1. distance is the distance (int turns) between them
Astrobytes: *in turns
eulerscheZahl: and the distance is an int
Astrobytes: :P
eulerscheZahl: btw where did the space maze guy go?
Astrobytes: Just disappeared. Thankfully
Astrobytes: Working on his ray casting NN I suppose.
NASADeveloper102: I don't mean that.
NASADeveloper102: I mean how do we know the ID of the factory we're talking about.
NASADeveloper102: ?
Astrobytes: factory1 is the ID of factory1 and factory2 is the ID of factory2 ;)
NASADeveloper102: Oh my lord
NASADeveloper102: I don't know how I didn't realize this.
Astrobytes: Too long staring at it no doubt :D
NASADeveloper102: Probably.
Astrobytes: Easily done
NASADeveloper102: Thanks for the help!
Astrobytes: No worries, good luck :)
NASADeveloper102: :thumbsup:
beinebruestearschgesicht: hey all
NASADeveloper102: Hello
NASADeveloper102: I'm doing well
NASADeveloper102: and you?
NASADeveloper102: Nice :thumbsup:
Hjax: NASADeveloper102 you took my suggestion?
Hjax: ill be working on a ghost in the cell bot this weekend as well
NASADeveloper102: Yes I did!
NASADeveloper102: Thanks, this game seems pretty simple, but it allows for a lot of different strategies.
AggYz: i lern python 2 min ago
NASADeveloper102: :clap:
NASADeveloper102: Good job
tutubalin: Python 2 is a bit outdated
NASADeveloper102: I don't even think you can get it in codingame anymore
-Gamma-: i think he/she meant...
-Gamma-: I leant python "2 minutes ago"
-Gamma-: learnt*
NASADeveloper102: Oh I see.
tutubalin: No
tutubalin: I leant "python 2" a minute ago"
NASADeveloper102: Why do you say that?
Hjax: they definitely meant (i learned python) (two minutes ago)
NASADeveloper102: That's what I thought.
NASADeveloper102: Why would it be, " I learnt python 2, a minute ago"?
wlesavo: learn python2 min()
Hjax: learn min(python2, go)
NASADeveloper102: Huh?
NASADeveloper102: :joy: I wish I understood this.
Hjax: we are just being silly
Hjax: pay no mind :P
NASADeveloper102: Uh okay then
muy31: for some reason recently the console in various arenas have been giving me: Warning: your code did not read all available input before printing an instruction, your outputs will not be synchronized with the game's turns and unexpected behaviour may occur.
muy31: enev though i always read all inputs
muy31: is it some bug
Hjax: yeah pretty sure its a bug
Hjax: ive been seeing it too
wlesavo: this is a recent feature, i think it was introduced in some forum post
Hjax: i definitely am reading all of the inputs
muy31: but then it actually messes up my execution
muy31: and i timeout
wlesavo: are you sure this is the issue? i got that message only when i had an actual bug in my code
muy31: http://chat.codingame.com/pastebin/3719a39a-4369-4251-9bfd-78ca860eea0c
muy31: when I delete the 'depth' on the return statements i get the bug
tutubalin: i have it every so often
muy31: as in return -50, 50, etc.
tutubalin: when i press Play button again it disappears
Hjax: maybe they fixed it, im not getting it anymore
Hjax: i was getting it a lot the other day
muy31: no didnt fix, now i get it randomly
jacek: lets fix it and release on friday afternoon
struct: :)
Astrobytes: I get it randomly, sometimes when I crash, sometimes if I print to stderr, othertimes just out of the blue. Never once, may I add, for the reason it is intended.
karimdjemai: Has someone got an idea why bitwise operators in js like | might return something like 1010274303 ? shouldnt it be only 0s and 1s?
jacek: w00t?
karimdjemai: for example (1010011101 | 1010110010) evaluates to 1010274303
Hjax: because you arent calling it on a binary number perhaps
Hjax: i phrased that poorly
jacek: no way, in python its also 1010274303!
Hjax: it takes two numbers, regardless of a base, and returns a number, regardless of base
Hjax: base is just how you display the number, it doesnt change the value
jacek: karimdjemai try (0b1010011101 | 0b1010110010).toString(2)
karimdjemai: ohh so youre saying it interprets the numbers as base 10 by default?
Hjax: yes, but you are also inputting base 10 numbers
Hjax: 101 is one hundred and one, not 5, unless you specifically tell it you are using base 2
karimdjemai: thank you a lot i was very confused :D
Hjax: 1010011101(base 10) binary or 1010110010 (base 10) is in fact 1010274303 (base 10)
tutubalin: Also, remember about no very well-known feature of JS: bitwise operations always cast number to 32-bit signed integer.
tutubalin: 1e10 | 0 == 1410065408
tutubalin: (2e10 * 2) != (2e10 << 1)
jacek: https://stackoverflow.com/questions/26323776/why-does-this-print-out-10
tutubalin: jacek https://tutubalin.github.io/
tutubalin: watch page source )
Astrobytes: haha very nice
tutubalin: script works in Node.js as well )
tutubalin: https://stackoverflow.com/questions/22588223/how-does-this-magic-javascript-code-work
tutubalin: also some nice slides from Martin Kleppe
tutubalin: https://speakerdeck.com/aemkei/the-art-of-obfuscation-frontcon-riga
code_maniac: This website is cool for such braifuck tricks http://www.jsfuck.com/
code_maniac: This website is cool for changing any normal code to brainfuck code http://www.jsfuck.com/
jacek: you mean to change into js
code_maniac: You can change both way
Astrobytes: Martin Kleppe created JSFuck
code_maniac: He has given a talk on this as well
Astrobytes: I liked his 'inceptions'
Astrobytes: jacek y u still #2
jacek: i dont wanna tric to rage quit
Astrobytes: :laughing:
tutubalin: Astrobytes yeah, inceptions is great
tutubalin: alas there are some XSS attacks are based on this technique
Astrobytes: Yeah, that's always gonna be the way whatever you do. It gives you some knowledge to be able to detect such attacks however.
NASADeveloper102: Do you recommend light mode, for chat and coding
jacek: :scream:
NASADeveloper102: What?
NASADeveloper102: Did I say forbidden words or something?
NASADeveloper102: Is "light mode" a construct that is aggravating to some people?
NASADeveloper102: Probably honestly.
jacek: eeyup
CPea: ey
jacek: https://blog.codota.com/wp-content/uploads/2019/02/04-Better-than-me.jpg
Astrobytes: "light mode" - what is this witchcraft
tutubalin: come to dark side
NASADeveloper102: I'll just continue using Light++
NASADeveloper102: *Light+
Astrobytes: You don't know the power of the dark side
NASADeveloper102: Uh, sure.
Astrobytes: :P
Astrobytes: Just ribbing.
NASADeveloper102: :joy:
NASADeveloper102: Okay then.
NASADeveloper102: Also why would you use vim for C#?
NASADeveloper102: Not you specifically, but anyone.
Astrobytes: Masochism?
jacek: because i cant leave
NASADeveloper102: y i k e s
NASADeveloper102: I see.
Astrobytes: Finished your boss yet darkhorse64?
darkhorse64: Yes. I"m trying to understand why it works so well. I removed the solver, nerfed the engine by a factor of 2
Hjax: boss for what?
Astrobytes: How 'well' does it work?
muy31: breakthrough
Astrobytes: no, Othello
darkhorse64: othello. It ranks #6
muy31: oh
Hjax: why are we calling it a boss?
Astrobytes: Because it will be
muy31: for leagues, right
Hjax: othello is going to have leagues?
Hjax: do community multis have leagues?
Astrobytes: Well, 2
Astrobytes: Yes. Only Wood leagues though
darkhorse64: If struct wants for a boss for a league, I could offer one
Hjax: i see
Hjax: why only wood leagues?
Astrobytes: But you can call them whatever you like
struct: Sure darkhorse64
Astrobytes: CG issue Hjax
darkhorse64: I think a tough may foster some interest from other players. Give me a few days for testing. My eekend is going to be very busy (read not much time for CG)
darkhorse64: *tough bot*
Astrobytes: Yes, I'm all for decent bosses
Astrobytes: And take your time, no rush
Hjax: im too motivated by codinpoints to try the smaller multis
Astrobytes: lol, there's such a thing as fun you know. F'ing SC players :P
Hjax: isnt fun just watching numbers increase?
Astrobytes: :D
Hjax: you must have a weird definition of fun :P
Astrobytes: lol
Hjax: idk, smaller multis are spooky
Hjax: because they are almost exclusively populated by really good people
Hjax: makes it hard to measure your own progress
struct: Site broken or only clashes?
Astrobytes: I heard there's gonna be pylons and cannons in Othello...
struct: It was 60+ people clash and it broke :p
struct: 80
Astrobytes: I'd report that struct, could be a capacity issue on CG side
Astrobytes: btw, wtf clashes have 80+ people?! Stream?
struct: yeah
Astrobytes: Madness.
jacek: maybe ru guys they moved from uttt
Astrobytes: Hjax, I do smaller multis and I'm shit, don't worry about it :)
Astrobytes: lol jacek
struct: darkhorse64 don't worry too much, I would code a bot, but I also don't have much time
jacek: maybe make my bot a boss
struct: I'm spending time on learning sdk modules
jacek: i would have more motivation to improve myself
jacek: :v
Astrobytes: They don't document the JS modules struct?
struct: It's a bit poor
struct: Only real documentation is UTG referee
Astrobytes: They're updating docs soon right?
struct: No idea
Astrobytes: Sure I saw that in forum
Astrobytes: I don't wanna work on SDK wiki until they do that
darkhorse64: I would rather be worried not be able to beat my own creature
struct: well it can use half the time
Astrobytes: Yeah, reduce depth or calculation time, or both
struct: easier solution
struct: // #pragma...
Astrobytes: Might make it toooo dumb ;)
Astrobytes: Don't want it to fail on turn 1 precalc
struct: Ideal bot would be 30.00 points atm
struct: boss*
struct: wait 30.18, so no java allowed
Astrobytes: hahaha
Astrobytes: But yeah, 30 seems about right given the gaps
Hjax: why no java allowed
darkhorse64: No problem for that but I ant to understand how features removed improved (or not) my bot strength. For instance, doubling the rollout count helps but is approximately + 1 ELO point
jacek: +1 with confidence [-100,101]
struct: Hjax Im joking, I just don't like Java
struct: And Im forced to use it on sdk :(
Hjax: :(
Hjax: i like java
jacek: kree java!
darkhorse64: I made several runs
struct: you use mcts right?
darkhorse64: I do
struct: early game might be a bit weak right?
darkhorse64: Usually things start to look wrong against the top by move 35 but it may be due to a weak start. I don't play othello myself
jacek: neither do i
Astrobytes: I only ever played casually myself, doing a lot of reading
jacek: i think i would have some problem playing it with real board
Astrobytes: Nah, you'd pick it up pretty quick I think
darkhorse64: I am only a chess player
Hjax: i think my goal for this weekend will be to go from bronze to legend in ghost in the cell, with 1 submit
Astrobytes: Stream it! Stream it! Stream it!
Hjax: lol, why
Hjax: my development process is not very fast, or interesting
darkhorse64: I did that nearly for Tron
Astrobytes: Just a stream-worthy comment
Hjax: its easy to get into the habit of spamming games against top people, or just spamming submits
Astrobytes: This was during a contest darkhorse64 no?
Hjax: i think it will be a fun challenge to not do that
Hjax: plus i think ghost in the cell is a pretty easy game, so handicapping myself a bit is more fun
darkhorse64: No
Astrobytes: Or after or something. I think you applied something you'd used in another contest or multi, and went gold or legend
darkhorse64: I read the right PM
Astrobytes: Oh I see
darkhorse64: BFS for the in
darkhorse64: win
Astrobytes: You don't even have to correct your typos anymore, I'm quite used to filling them in myself :P
Astrobytes: Need to get your ankles working properly to fix them.
darkhorse64: my w key is not in great shape. Technically, it is not a typo but a hardware malfunction
Hjax: the hard part is going to be knowing when my bot is strong enough enough
Hjax: uh, scratch one enough
Astrobytes: OK, I'll give you the benefit of the doubt darkhorse64
darkhorse64: It's so kind of you
Astrobytes: I'm a generous type :D
Astrobytes: Hjax, you can use replays from top bots etc etc
Hjax: sure, but i cant play against them
Hjax: i read agades pm
Astrobytes: Pretty sure you can CG bench against them if you fiddle around a bit
Hjax: no no no, thats cheating
Astrobytes: Nope, technically it is not.
Hjax: hmmmmm
Astrobytes: But I get your point ;)
Ratchef: Is there a way to invite my friends to a private battle?
Hjax: i really like agades solution to the question "can i get punished by increasing here"
DestroyTasmania: yeah you can create privat lobby and send them link or just invite on right
DestroyTasmania: @Ratchef
Hjax: just play against a clone of yourself that doesnt increase, and see if you get rekt
Astrobytes: I don't remember if I ever read the PMs for that game tbh
Hjax: i played a lot of galcon back in the day
Hjax: on my ipod touch
Hjax: fun game
Astrobytes: And yeah, self-play also a solution. Just don't overfit. brutaltester is good for that
Astrobytes: Yeah galcon was great
struct: I tried to get legend in 1 submitg
struct: did not end well
Hjax: i also played the 2010 planet wars ai challenge
Astrobytes: something something something account deletion
Hjax: back when i was brand new to programming
eulerscheZahl: public github repos are a great source for bronze to legend in 1 submit
struct: I think the closest I got was uttt
struct: Where i got like top 3 gold
struct: from bronze to legend in 1 submit
Hjax: lol, as test opponents, or as copy/paste steal the bot euler?
struct: c&p
struct: https://www.codingame.com/profile/5f47dac1fe5458d3d6d596f391d8c8cb7572263
struct: If there is a bot
struct: its in this profile
eulerscheZahl: i was expecting this user when i clicked the link
Astrobytes: :D
struct: I think I sitll have the link of his old account
struct: give me a sec
darkhorse64: When you see a top 20 guy ask what is a BFS, you figure out why C/P is a bad thing
Astrobytes: To paraphrase jacek "eeyup"
Hjax: lol, he stole all of his submissions?
dbdr: 16th?
dbdr: copy paster I guess
darkhorse64: This is the guy I was thinking of
Astrobytes: yeah
darkhorse64: Look at his CP profile, especially the time scale
Hjax: im surprised theres such strong bots publicly available
dbdr: 4996CP in clash
darkhorse64: guru in to weeks
eulerscheZahl: struct claiming that it's anirudh? i don't think so
struct: yes
eulerscheZahl: darkhorse64 sorry to bother you but what is that BFS you were talking about?
darkhorse64: :smile:
dbdr: Big Fing Secret
dbdr: nobody really knows
eulerscheZahl: Fing? Chinese?
Hjax: theres really that many strong open source bots? im kind of surprised lol
dbdr: effing
Astrobytes: Bird First Search
JBM: isn't effing spelled with an E?
dbdr: yes but that ruins the accronym
Astrobytes: f'ing would be more correct
darkhorse64: Hjax: yes
struct: https://www.codingame.com/profile/db6321ede0cfc71dba63fb778cc61f677390781
JBM: can we ever be too correct?
struct: This was aniru I think
eulerscheZahl: https://www.youtube.com/watch?v=bawmxQE_Fj0
Astrobytes: Who knows JBM, we can try
reCurse: That's really sad
reCurse: Was talking about the copy paster
Astrobytes: :D
eulerscheZahl: no struct. there should be a tagline "almost alll languages". with 3 L
Astrobytes: JBM: "Who knows, JBM. We can try"
eulerscheZahl: old actually, account deleted years ago
eulerscheZahl: he claimed to be able to code in all languages
eulerscheZahl: as the bots he copied were in different languages :D
Hjax: amazing
Astrobytes: ani d profile looks about right then "I am a 12th grade student at Indus Bangalore, I have learnt several languages: python, C++, C, C# and Java. Dabble in JS, ruby. Now learning D"
struct: So this one ueler?
struct: https://www.codingame.com/profile/9b6f917d8ba745e1194e9922b5ef3dd67406451
struct: Platinum Rift - Episode 2 0th / 0 / 0
struct: This is the first time I see this in CG
struct: https://i.imgur.com/Tipp1Ig.png
Astrobytes: It's not True CG unless it's 0nd
struct: No score on multi
struct: Not even language shows
eulerscheZahl: you found him struct!
struct: yes he is bottom of hypersonic legend leaderboard
struct: last player
struct: Doesnt even show language
eulerscheZahl: iirc thibaud did some cleanup
jacek: a purge
eulerscheZahl: he got DQed in a contest (stole python bot, added bash wrapper and claimed it's pure bash)
eulerscheZahl: then rage-deleted the profile after he got lots of hate on the chat
struct: also he used bot on CoC
jacek: who rage deletes their account?
Astrobytes: :thinking:
eulerscheZahl: absolutely barbaric
Astrobytes: :D
freeman42x: streaming clash of code: http://twitch.tv/freeman42x
Hjax: why did struct delete his again? because he was upset with his UTG bot?
struct: I said legend in 1 submit or delete
Hjax: ah
Astrobytes: Waited until last day
Hjax: UTG is a pretty hard game to get legend in 1 submit with
jacek: just like you, 1 submit from bronze to gold
Astrobytes: , night even right struct?
Hjax: because its so metagame dependent
struct: I cant remember well
Hjax: planting fake bombs and guessing your opponents
struct: But maybe
Hjax: hard to test properly with just selfplay
Astrobytes: Pretty sure you waited til the night
Astrobytes: Don't do that again eh
struct: We need more anonymous contributions
struct: *undefined
freeman42x: private clash: https://www.codingame.com/clashofcode/clash/1262076003528789ee2605e32db72839683f30f
Daher: private clash
Daher: https://www.codingame.com/clashofcode/clash/126209199251b6823a67af3f99a8c9e48c2b461
Daher: for java, c, c++ and c# coders
Astrobytes: Not very private if you er, post it in world chat, is it?
EliTheCoder: he just doesn't want to get bots in his game
EliTheCoder: and he can control what languages he wants
Astrobytes: The bots will be there if not enough players anyway. And yes, I'm aware of that. The feature is supposed to be for sharing with friends you want to clash with.
muy31: what's the deal with short time limits anyway? does making it too long cause issues with servers and whatnot or what?
Astrobytes: Yes. Imagine a leaderboard of 2000 bots all using 1 second of calculation time, playing matches against each other
Astrobytes: For the CG configuration this would be a Very Bad Thing.
muy31: i see
muy31: that is why we cannot do chess on CG
Astrobytes: We could have a chess game on here for sure, it would have to depend a helluva lot on book moves though :)
Astrobytes: Fischer-Random/Chess960 might be doable though
Astrobytes: And it still might be a nice addition just to practice on
muy31: i dont see it happening and any of the ais submitted to the leaderboard actually being good
Astrobytes: Oh yeah? I'd bet otherwise
muy31: unless like someone hard codes a move for each state
Astrobytes: like I said "it would have to depend a helluva lot on book moves though"
Astrobytes: Which is about 80% of chess anyway right
muy31: in that case, one would preprogram a chess bot locally and the CG bot would just a 'scribe'
Astrobytes: Not quite, bit more to it than that
CykaBo: share code on hjava pls
Astrobytes: Neumann, I never noticed your 5th place in CSB, nice
Neumann: Thks, SR gave me a slight eval improvment :D
Astrobytes: "slight" :D
Astrobytes: CykaBo 1234abcd!"£$ - now decipher the code
CykaBo: at the end I meant 4Head
Neumann: What do you expect us to understand with "share code"
CykaBo: show code or whatever the option's name is
Astrobytes: Qml0ZSBteSBzaGlueSBtZXRhbCBhc3Mh
JBM: that password is going to need replacing
Neumann: ?
JBM: hadn't seen you in a long time
JBM: always a delight tho, until proven otherwise
Astrobytes: So as not to pollute #fr JBM, I suggest your trolls should have been Korrigans
JBM: you're so welcome to pollute #fr
Astrobytes: I try not to, sometimes I "cross over"
JBM: i personally welcome it more when it's expressed in #fr
JBM: as bad as it comes (hi MK)
Astrobytes: lol
JBM: the effort is what makes it worth it
Astrobytes: on that note, MK left it seems
Astrobytes: Anyway I'm out for tonight
Astrobytes: gn all
tomatoes: gn
NASADeveloper102: Okay
NASADeveloper102: With what
NASADeveloper102: -HUh
NASADeveloper102: What?
NASADeveloper102: I don't know whether or not to take this seriously, but I assume if you were dying in real life you would call the police and probably not come here, the probability that you would come here instead of any other place is very low.
NASADeveloper102: I don't know why I just explained my train of thought to a person who was saying, " Help, I need a notch apple "
JBM: police will be with you shortly
NASADeveloper102: I wish I knew how to respond to that.
JBM: "is habeas corpus even a thing anymore in my country?"
NASADeveloper102: Uh....
NASADeveloper102: What?
JBM: sorry i thought you were american for a while
muy31: Why would this man cheese my bot like this? https://www.codingame.com/replay/473615813
NASADeveloper102: Uh I'm American.
NASADeveloper102: I don't know what a "habeas corpus" is.
NASADeveloper102: I probably should, now that I think about it
muy31: right to go before a court when arrested
muy31: 6th/7th amendment
USE-TO: SUPPORT PYTHON3.8!!!!!!!!! PLEASSSSSEEEEEE
tomatoes: what so good 3.8 got?
NASADeveloper102: %= ?
NASADeveloper102: I don't know
LelouchVC2: Python 3.8 introduces a feature that lets the code fondle your balls
scareware047: can i downgrade
scareware047: my league
scareware047: and start over in another language
tomatoes: no, but you can start with another language in your current league
scareware047: ye but current league is hard xD
Hjax: 3.8 has the walrus operator
Hjax: :=
Reksio: Hi All. Is code golf counting a source code length after or before the compilation ?
Reksio: I also wonder how time it takes to finish the puzzle is taken into account if at all.
Reksio: Anyone?
SPDene: time taken to finish is only relevant is it takes (i think) > 15s. then it will time out/fail
SPDene: if*
struct: code length is show on ide
struct: its before compilation
struct: all chars count, spaces, tabs, comments
Reksio: haha :) this is great
Reksio: and you say 15s 8-)
Ratchef: Hello I have a question about Java
struct: Also compilation time doesnt matter
Reksio: +1
struct: Since the time will get is the time you press on submit
Ratchef: Can I add a variable to itself like this? int value = value +1;
Reksio: yes
struct: Ratchef yes, but you can declare it again
struct: like you just did
struct: you did int value = value + 1;
Ratchef: It often tells me that the value isn't initialized
struct: yeah, if you never assigned a value to it
struct: it might happen
Ratchef: So I should do int value = 0; value = value + whatever;
struct: or value += whatever;
Ratchef: okay I'll try next time thanks
Ratchef: Is there a way for the code to know if a character is uppercase or lowercase?
struct: yeah
Ratchef: Is that a java function?
struct: https://www.tutorialspoint.com/java/lang/character_isuppercase.htm
Ratchef: ty
Danger189: any python people have a second to help me with something?
Hjax: Danger189 what do you need?