Chat:World/2020-04-07
MadKnight: what's the problem thaibien ?
Riyuk: Hey
Riyuk: @MadKnight you don't sleep :p
MadKnight: i slept at daytime
eulerscheZahl: he's like a dolphine: sleeping with half of his brain
MadKnight: how's your submarine Riyuk ?
Riyuk: Still a lot of things to improve
Riyuk: I need to fix some bugs on my moves in order to get to silver
Riyuk: How's yours?
Riyuk: You're more or less like him @eulerscheZahl :D
eulerscheZahl: i just work up after an 8h sleep
mark7: How to get to gold??:sweat_smile:
eulerscheZahl: rank above the silver boss
Neon_Tuts: I not sleeped at
eulerscheZahl: beat captain Haddok
mark7: That knda hard
Riyuk: You're more like a shark then
Riyuk: PP oO
Zandy156: Silver boss has a few weaknesses.
Zandy156: I don't know about the other bots at that rank though.
mark7: But only beating silver boss is not enough
Zandy156: everyone who's in gold now seems quite strong.
mark7: Right now I analyze every replay in which i lose and trying to improve
thousandSun: so how exactly do i complete these challenges?
Neon_Tuts: Hello World
AnasShami: hello from the other world
Uljahn: hello Automaton2000
Automaton2000: do you go to legend
Neon_Tuts: https://www.codingame.com/clashofcode/clash/1178054716d0b938afa03a331ec2ffcb446c088
Riyuk: Going to Silver :raised_hands:
Riyuk: Hola Automaton2000
Automaton2000: no, but you can see what i need to check the referee
MadKnight: gogogo
Th1nhNg0: có ai có gợi ý nào cho "ocean of code" không ?
Th1nhNg0: sorry wrong channel
AntiSquid: ya nobody here speaks vietnamese
Riyuk: You can have some suggestions/Ideas here as well @Th1nhNg0
Cappefra: I'm kind of stuck on skynet 2
Riyuk: That's a valid readon to start Ocean of code :p
Cappefra: I'm already doing it :grinning:
Cappefra: still bronze league but I hope I'm moving up
Nerchio: how do I go about precise tracking like some people do even in bronze? It will probably take twice as many lines of code as I already have :D
Cappefra: yeah that's one of the issues
Nerchio: I feel like I am not advancing any further without adding tracking atm
Cappefra: my code's decent on tracking but not great on movement and decision making
Nerchio: I am only doing some tracking when I hit the enemy and i shoot randomly :P
Cappefra: yeah tracking's important. Work on an external IDE if you're not already doing it
Nerchio: how many lines of code you have in bronze? I know languages can differ a lot :p
Cappefra: http://chat.codingame.com/pastebin/ec45c9fc-5ac8-4050-a506-5efdd6c18ffc
Cappefra: wot
Cappefra: anyways, around 1300
Cappefra: with that style you see in the link's snippet
Cappefra: a lot of it is architecture though. I feel like having lots of classes simplifies the algorithm a bunch, so tracking comes easier and all
Nerchio: yea for sure I need to fix my mess before I continue making this more sophisticated
Cappefra: It's my first contest so I underestimated the issues at first, I started coding imperatively until I got to position 800 or so, then I noticed it would become way too messy and I deleted everything and did it OOP
Cappefra: it was worth it though cause I went from 1300 to 680 in one submit
Nerchio: nice i do OOP and I am around 500 but it's a simple bot
Nerchio: my next step will probably double it
Nerchio: I am guessing people keep a list of moves by the enemy and recalculate each round until they find something?
Cappefra: that sounds like a way to do it
Riyuk: You need to track the enemy as much as possible and hit him only when you're more or less sure in order to get to silver
Cappefra: but there's many things you can consider: moves, your hits and your misses, its hits and its misses, its surfaces...
Nerchio: yea i know :S
Nerchio: lots of stuff
Nerchio: that's quite a challenge
Cappefra: yeah but if you study what happens to the map of possible position one can have after a particular move
Cappefra: with a little bit of graph (grid) theory you can get pretty precise
tobk: "If you fail to output a valid action, you will SURFACE in that turn." This should be reworded to "If you fail to provide a movement action (MOVE or SILENCE), ..."
Uljahn: why
tobk: (i.e. if you e.g. TORPEDO or just MSG, you will still surface if you have no way to go)
tobk: at least so it seemed when I just tried
tobk: also meaning you an not delay the SURFACE if you have just one health left...
Uljahn: actions are not just moving and silence but all possible commands besides MSG
Uljahn: "Each turn you must perform at least one action."
tobk: hm, I just tried, but surfaced despite doing "torpedo ..." but without moving
Riyuk: i.e you surface automatically if you don't move or silence ?
Uljahn: check the syntax
Uljahn: i guess you have a bug
tobk: Well, that's what happened when I just tried. Maybe someone else can confirm?
Uljahn: show the replay
tobk: don't have a replay ready right now, and takes a while to reproduce one
Uljahn: your target was out of range
tobk: weird, not it worked. could drop a mine without moving and did not surface.
tobk: okay, forget about it...
Lily123: how is it possible to see other's solution and compare it to your's?
Uljahn: solve the puzzle then submit and get 100% correct
Uljahn: then you'll be able to see the community solutions
Uljahn: for that language you solved with
KiwiTae: hows the contest going? :D Uljahn your on it?
Nerchio: for enemy prediction do you guys use a queue for enemy moves and repeat it for every position on the map?
Uljahn: hi Kiwo, im stuck in silver
KiwiTae: Uljahn im in Wood2 hehe
KiwiTae: no time :(
Quidome: @Nerchio, ik keep a trail for every possible starting position, throwing out the ones that are impossible
Quidome: After a silence you can have multiple trails for a single starting position :(
Nerchio: yea but at the start there is so many starting positions so you have like 150 trails? :P
inoryy: maintaining possition masks is more efficient rather than queue
Nerchio: I thought about that but feels like a queue gives you more precise information
inoryy: the position distribution process is markov so you don't need to care about history except for when a mine is triggered and your mask is wrong
codingWhale: When you remember where the enemy has been before, though, and are able to narrow it down enough, isn't that an advantage because you know where the enemy cannot go (because he's already visited some cells)? I feel like it'd help.
Nerchio: I am thinking that way that's why I am leaning towards a queue. In my mind it's much simpler to implement
Nerchio: This is my first competition so I am not sure if at any time you are close to reaching the response time limit?
Ifthel: I'd store both. Keep a mask if you want to know their visited cells but otherwise rely on a mask
Nerchio: I can imagine that for some maps and queue of 20 moves it can get quite long
Ifthel: queue*
inoryy: yes, maintaining a mask that you update on each step is helpful. a full history of actions in addition to that is also helpful. a full list of actions on its own, not so much.
kovi: maintaning a mask is importatn in these days
Riyuk: fore sure @kovi :p
T-Rone: THE RIVER I i don't understand
Quidome: What do you not understand?
Quidome: Still not sure whatr you ask
Quidome: Do you understand the problem statement or need help with the solution?
T-Rone: yes but i can't watch it
Quidome: What do you mean by can't watch it
Quidome: You can only see the other solutions after you solved it yourself
T-Rone: this ,i know ,thank you
Quidome: You don't need a queue to solve it though
yfprojects: have you seen the teaser for the spring challenge?
Quidome: a simple while loop will do
Quidome: nice teaser right !
Quidome: Can't wait :D
T-Rone: the loop must be required
Quidome: while the 2 numbers are unqual calculate the new number for the smallest
Quidome: when equal, you found it
T-Rone: that is right:joy::smiley:
yfprojects: where is dbdrs chatlog?
T-Rone: i am going to have dinner ,bye bye ,thank you very much--
wlesavo: https://cg.spdns.eu/wiki/Main_Page
Nerchio: I think my calculations got so heavy that it lags the browser :D
wlesavo: nice sub chuck
chucknorris: thx. yeah finally some real improvement
chucknorris: but lets wait for pb4's submit, I have terrible winrate vs him usually :p
pb4: don't worry about this one
pb4: I'm trying to fix my abysmal surface code
pb4: I basically self-kill myself a lot of the time by surfacing excessively
chucknorris: yeah thats my main problem too atm... shoots the opponent and then surface without a silence.
pb4: not exactly mine
pb4: see lots of space in front of me
pb4: think that if I surface that's even more space
pb4: --> surface -_-
chucknorris: yeah it's really hard to balance. if you reduce the space coeff than it will go into small cornes too shoot the enemy. and then you'll have to silence in few rounds :p
chucknorris: *surface
pb4: https://www.codingame.com/share-replay/446444445 step 116
pb4: rhaaaaaaaaaaaaaaaaaaaaaaaaa
chucknorris: but I think many of us has this problem :p frame 307
chucknorris: https://www.codingame.com/share-replay/446446769
pb4: oh that's a nice one :D
pb4: I've disabled all torpedo code to debug
pb4: and I got a win vs euler :D
chucknorris: I'm thinking about feeding the features of the game to a mini eval NN. and have multiple versions of my bots running in a local trueskill system, and keep the parameters of the winner bot :p idk if it'd overfit against itself tho
dbdr: euler is abnormally low no?
dbdr: maybe he has broken sth
chucknorris: running games directly throught the api is considered as cheating ?
Liftyee: bruh wtf https://www.codingame.com/replay/446450967
Liftyee: also im a noob so don't judge
Nerchio: can I torpedo and silence ?
AntiSquid: Liftyee at least you didn't hit the single square lake :D
AntiSquid: mine doesn't dodge small lakes either no worries Liftyee
mark7: How do you store possible mines? Do you avoid going there?
yfprojects: AntiSquid didnt you make a pacman multiplayer game?
mark7: I am facing a weird bug. Lets say I triggered a mine, op shoot a torpedo in adjacent cell, so op health decreases by one, but my bot think op life decreased because of mine and My tracking system loses op. How to tackle it. https://www.codingame.com/share-replay/446469009 move 85. Number on left is opponent possibilities, no in right is my possibilities.
tobk: mark7 keep track of both mines and torpedos, yours and opponents, and add up the damage for all cells, then compare those to the actual damage inflicted
mark7: Okay means, I should check damage by both possibilities? and only remove possibilities that didnt occur in any situation?
mark7: tobk
chucknorris: read this postmortem. most of us use something like this to locate the enemy
chucknorris: https://github.com/Agade09/Agade-Wondev-Woman-Postmortem/blob/master/Agade_WW_Postmortem.md
mark7: chucknorris thanks
tobk: let's say you trigger a mine at 3/3, opp sends a torpedo to 4/4 and surfaces, than the cell 3/3 gets 4 damage.
mark7: Okay now I understood
mark7: remove op that has less than = 6 life
mark7: but Now i need to rewrite whole tracking system, bcoz I only store op path
mark7: remove op with 0 lif *
MichelLaviron: is there a known algorithm for going through as many tiles as possible without surfacing ?
MichelLaviron: is there a known algorithm for going through as many tiles as possible without surfacing ?
wlesavo: you dont really need one
wlesavo: but google hamiltonian path if you really want to
MichelLaviron: thank you, i think i will start with that for my first submit :)
shubhamGoswami: join the clash https://www.codingame.com/clashofcode/clash/1178423a63354a5a63ba9fd40a4d6b9c3c6137a
Nerchio: I worked 6 hours doing tracking i am so bad at this
AntiSquid: shubhamGoswami don't post clash links in chat please @_@
Nerchio: it works now but it breaks if enemy uses too many silence :D
AntiSquid: it's bad enough people get spammed invites
AntiSquid: if anyone likes that stuff they will join it otherwise no need to pester people with it shubhamGoswami
Astrobytes: Yeah he spammed it in 2 discordd channels too
Astrobytes: *discord
AntiSquid: well done Nerchio, now you can get bronze maybe :D
AntiSquid: just add silence in wood 1
Nerchio: I am in bronze :P
AntiSquid: ah
AntiSquid: nvm sorry :p
Nerchio: that's why I needed to do a bigger step since my bot was fairly random
AntiSquid: you could do a cut off, when you have way too many just limit it
Nerchio: and now it tracks quite well
Nerchio: after I added tracking for some reason my bot moves into islands now
Nerchio: #programming
Nerchio: woops
pb4: Can anbody suggest an opponent against which to test who lays mine everywhere ?
pb4: Preferably a weak opponent since I have disabled all evasion code and would like to see long games anyway :D
wlesavo: lol
2571540: Kannely?
2571540: Pb4
pb4: Thanks, exactly what I'm looking for :)
2571540: Sorry for double messages my xmmp app looks buggy
2571540: My code could work too but idk how this version is doing
2571540: I'm trying to increase the map coverage of mines but it's not working as intended
pb4: never is...
pb4: oh
icecream17: so here's a bug i found... whenever i trigger a mine the "explodage" occurs on two places. how would that work....
pb4: Let's add "deterministic" to the specifications...
pb4: weak bot, lays a lot of mines and is deterministic ?
kovi: pb4 without randomize?
pb4: Yes
kovi: maybe if you mine only we can consider it weak(er than normal)
chucknorris: maybe you could alse submit a working version of your bot and comment out trigger, silence and torpedo
2571540: I think Kannely is deterministic, isn't it?
chucknorris: that would be useful for me as well :D
pb4: Would it be weak though ? :japanese_ogre:
pb4: hehe :D
2571540: XD
2571540: It's 100+
chucknorris: :D
pb4: the 100+ one can't torpedo, and doesn't try to stay stealthy
pb4: only torpedoes allowed are for instakills
wlesavo: pb4 if you dont shoot and trigger, probably week enough
pb4: But yes more seriously it would be weak enough :d
wlesavo: and if its not you can be either really proud of your old version or dissapointed in a new one :wink:
Nerchio: I don't get to see what they power up right?
wlesavo: right
Nerchio: so MOVE S SILENCE is MOVE S for me
wlesavo: exactly
Nerchio: thanks
wlesavo: np
Nagatwin: eulerscheZahl https://www.codingame.com/share-replay/446526839
Nagatwin: Zoning + spam silence looks strong against you :D
pb4: Yes it does
2571540: I tried to spam silence when I'm ahead
2571540: Now when it's ahead my ai spams silence until it times out
2571540: I'm not sure I reached my objective
chucknorris: submisson nr806 :D
eulerscheZahl: right now everything is strong against me, i modified the "character" of my bot to make my mine avoidance stats fit closer to top bots
eulerscheZahl: received 1.47 damage by mines per game
pb4: What does that mean ?
pb4: oh
wlesavo: at what cost?
eulerscheZahl: now i'm at 0.49 last time i checked
eulerscheZahl: cost = more surface :D
eulerscheZahl: that's the next TODO
wlesavo: lol
pb4: It's a hard TODO
pb4: at least for me -_-
eulerscheZahl: but i have bar charts in my stats now
chucknorris: https://i.kym-cdn.com/photos/images/newsfeed/001/465/425/3fd.jpg
pb4: :D
wlesavo: average damage by surfacing would also be a nice stats to learn
chucknorris: avg dmg / surface ?
pb4: avg surface / game
2571540: Xd
2571540: Dmg/surface = 1?
eulerscheZahl: pb4 with 1.54 surface per game
wlesavo: yeah i mean cummulative per game obviously
eulerscheZahl: note: these stats are from yesterday
eulerscheZahl: agent id: 2757348
2571540: I m wonderin what my stats are now xD
eulerscheZahl: the one with these crashes: 446243053 446213878 446096644
eulerscheZahl: (replay IDs)
pb4: Thanks eulerscheZahl, may I re-ask a bit later when I have fixed this ?
eulerscheZahl: not sure how often i should aquire these stats
pb4: heavy for CG servers ?
eulerscheZahl: downloaded 3k replays got blocked temporarily
pb4: haha
eulerscheZahl: server refused my requests
2571540: Haha
eulerscheZahl: Nagatwin i don't have you in my stats
2571540: I don't need those I will do it myself someday
eulerscheZahl: only top15
2571540: Oh
wlesavo: i now want to be in top10 just for the sake of seeing stats on myself at the end of marathon, to bad im probably not gonna make it
2571540: Omw
eulerscheZahl: for the final results i might lower the bar a bit
2571540: I should have my Sim ready by the end of the week :p
eulerscheZahl: but will i get a more pretty UI till that? https://imgur.com/a/hhciThu
icecream17: whoa that data thing looks cool
eulerscheZahl: gets more interesting when you have the comparison to other bots
eulerscheZahl: like some are shooting torpedoes all over, while others hit all the time when they shoot
eulerscheZahl: but do it less often
GingerFan301: Is there still a contest on May 7th? :thinkingface:
GingerFan301: I see that one started Mar 20th, didn't get an email for that one
icecream17: really? i got an email....
eulerscheZahl: yes, there's another contest
eulerscheZahl: https://youtu.be/63iMGOS6_u0
eulerscheZahl: i think it's 11 days, even if it say 10 in the video
GingerFan301: Cool thanks
Keyser_D_Soze: epic contest
GingerFan301: and this one is a month?
GingerFan301: How do we like it?
eulerscheZahl: yeah, you missed half of it already
eulerscheZahl: interesting game, has some depth
eulerscheZahl: i don't see a way to simulate the entire game and run a search on it, as you know too little about your opponent
eulerscheZahl: for of war
eulerscheZahl: fog
GingerFan301: :thumbsup:
Envious: how would you a solve a maze problem
Envious: you always start from (1,1) and end at (3,3)
Envious: but there are walls in betn, and you can only go 4 ways: up, down, left ,right
AntiSquid: bfs
Envious: i have no clue how to do that
chucknorris: I saw siman once charging next to me putting a mine down-> my turn. i had everything on cd. ->his turn. moves away, torpedo, trigger. 3hp down, win. idk if its luck or he go deeper than 2 turns
AntiSquid: https://www.geeksforgeeks.org/breadth-first-search-or-bfs-for-a-graph/ Envious
AntiSquid: or maybe this https://www.geeksforgeeks.org/rat-in-a-maze-backtracking-2/ Envious
Envious: would i use map() in this case ?
AntiSquid: you can use a normal list
AntiSquid: ok i know, python lists aren't normal
2571540: Frog of war
Envious: this must be leetcode medium
Envious: and i can only do some of easy on leetcode
AntiSquid: remove comments and empty spaces it's short
Envious: https://www.geeksforgeeks.org/rat-in-a-maze-backtracking-2/
AntiSquid: yes
AntiSquid: scroll down, click python in the tab
Envious: 2 ppl submitted 0 so far lol
AntiSquid: what are you solving?
Envious: it's a clash of code question
Envious: but i dont care about getting it
AntiSquid: then do normal puzzles
wlesavo: just check labirint
AntiSquid: they don't disconnect you from the IDE after 15 minutes
wlesavo: those are much harder
Envious: ok relax, i dont want something thats harder lol
Envious: i can barely solve this question
AntiSquid: https://www.codingame.com/training/easy Envious
Envious: which one do i choose
AntiSquid: just do onboarding, the descent, thor, temperatures
AntiSquid: good start
AntiSquid: then pick randomly
Envious: aight im doing onboard right now
Envious: what do you usually coe in
AntiSquid: python3 is perfectly fine
AntiSquid: use that
Envious: oh yea but i wanted to what u use
Envious: onboarding one basically told me the answer lol
AntiSquid: just use python3 least complicated
AntiSquid: and you can use it for anything
Envious: lmao yea looks like i have to do those games, these flash qs are hard
Envious: AntiSquid do you know difficulty these questions are in Leetcode
AntiSquid: uhm seriously forget leetcode?
AntiSquid: forget leetcode
AntiSquid: because it's irrelevant
eulerscheZahl: i thought about trying a contest there but then i saw that it's in the middle of the night
Astrobytes: leetcode is leetcode, codingame is codingame. There may be some crossover but it's hardly worth constantly comparing them.
AntiSquid: ya 2 hours only
eulerscheZahl: that's not bad, can be a fun format too
eulerscheZahl: but the 3am part is a huge turnoff for me
mark7: Try codeforces then
AntiSquid: there's also codechef
mark7: there are many cp sites, hackerrank, hackerearth etc..
Envious: acc ill try to figure it first
mark7: I can't see ant game information?
mark7: it is just empty
eulerscheZahl: ant game? you mean Langton's ant?
mark7: any *
mark7: that game information window
mark7: I clicked
eulerscheZahl: play and pause the replay
mark7: hard refreshing fixed it
eulerscheZahl: happens sometimes
mark7: bcos long debug msgs?
Envious: http://chat.codingame.com/pastebin/4d5bb9f8-59bb-4210-91db-05e7b1081e85
inoryy: I've heard from a few ppl that they used leetcode to prep for interviews which they passed successfully
mark7: leetcode is helpful to prepare for interviews, specially company specific questions
NotKid: leetcode has interview questions
pincedu17: Hello everyone, i'm trying to override equals (java). It seems that it is not working. Printing error in my equals is never called when using Map::containsKey. Any clues ?
rwilson: pincedu17 - sounds like you may not have ALSO provided an overridden 'hashCode' method
mark7: Another java problem, HashMap<ArrayList<Point>,ArrayList<Point>>(); not working.
mark7: in the sense, gives null pointer Exception
dbdr: pincedu17 to override, you need to use the same argument type
dbdr: i.e. equals(Object)
dbdr: a good practice is to use the @Override before the method
pincedu17: rwilson might be right, thanks for your reactivity guys
dbdr: that will give you a compilation error if you don't actually override another method
pincedu17: Yeah i had no compilation error, just the containsKey() isn't using my equals
pincedu17: stupid me to miss hashcode x)
dbdr: blame a bad API
mark7: How could I share some code here?
dbdr: juste copy paste it into the chat
mark7: http://chat.codingame.com/pastebin/1c2bd5a2-6af2-41e6-bc2c-16f7809192de
mark7: first list in map is for paths, second is for mines in that paths. but I get null in last line
AntiSquid: how long is the spring contest? is it a bot contest?
JFB: 10 days
pincedu17: Hashcode did the trick :)
AntiSquid: it doesn't say 10 days here: https://www.codingame.com/contests/spring-challenge-2020
AntiSquid: also looks like pacman
JFB: They say 10 days here: https://youtu.be/63iMGOS6_u0
icecream17: Oooh a bot contest. but so few time
AntiSquid: you have a month until it starts ^
rwilson: mark7 - the KEY in your MAP doesn't look like it makes sense. Double check what you're trying to do.
AntiSquid: also if it's pacman then it will be super similar to kutulu contest :/
rwilson: mark7 - If you're trying to build a map of all PATHS available at/from/to a given POINT then you'll want the key to be POINT not ArrayList<Point>
icecream17: kutulu?
mark7: rwilson I am using hashmap to store mines in a given path, ArrayList<Arraylist<point>> to store paths
AntiSquid: https://www.codingame.com/multiplayer/bot-programming/code-of-kutulu icecream17
rwilson: mark7 - look for PM chat
icecream17: 4 players!?
AntiSquid: yes
JFB: eulerscheZahl - you just like timeouts?
yfprojects: AntiSquid didnt you make a pacman multi?
AntiSquid: no
struct: I wonder what the contest will be
AntiSquid: this was the multi i made with 2 others https://www.codingame.com/multiplayer/bot-programming/botters-of-the-galaxy yfprojects
AntiSquid: that was something a staff uploaded for 5 minutes or so
AntiSquid: member of staff *
yfprojects: oh, do you still have the link?
eulerscheZahl: of course JFB
eulerscheZahl: i thought using arrays instead of lists would help my performance and reduce timeouts
eulerscheZahl: spoiler: no
greenthingsjump: Is this thing on?
AntiSquid: too loud greenthingsjump
greenthingsjump: -sadly shuffles off stage-
icecream17: I wish the world was simulated, and also we could program in alien....
ItachiEU: When does one get a rank?
greenthingsjump: If the world was simulated: http://chat.codingame.com/pastebin/0eae2bea-7e94-4c6e-8375-fecc96e436c7
dbdr: eulerscheZahl, what are you doing to your poor bot? :/
eulerscheZahl: trying to fix timeouts
eulerscheZahl: my path is a list<cell>
eulerscheZahl: tried bool[] and hashset
eulerscheZahl: both was worse
eulerscheZahl: which confuses me
dbdr: yes, sounds counterintuitive
eulerscheZahl: maybe i screwed up
dbdr: the devil might be in the details
dbdr: yeah
pb4: hey
pb4: sorry for the basic question
pb4: I usually launch python scripts this way : python myscript.py
pb4: This closes the shell when the script is done
pb4: I'd like to keep the shell open
dbdr: which OS pb4?
pb4: I'm thinking about just writing "python" and then find a way to launch myscript.py from the python shell
reCurse: Run it from bash?
pb4: linux mint
reCurse: Or whatever command line
NotKid: How do I increase my rank ?
dbdr: python should not close your shell
reCurse: ^
dbdr: does not make sense
pb4: Then maybe I used the wrong word
dbdr: are you in a terminal?
pb4: when i just write "python" in the terminal
pb4: I can write python code line by line
pb4: I'd like to lauch my script and then still interact with the variables created by the script
mark7: okay
dbdr: right, that's the REPL
dbdr: yes, then you should start python first (the REPL)
dbdr: and load your script
dbdr: https://stackoverflow.com/questions/5280178/how-do-i-load-a-file-into-the-python-console ?
AntiSquid: why not use jupyter notebook ?
pb4: thanks :)
pb4: python -i myscript.py
AntiSquid: or anaconda, it's so stupidly convinient
AntiSquid: same effect as running a script
pb4: I use anaconda
dbdr: I should have lmgtfy'd it :D
AntiSquid: so what's the difference? i don't get it
Nerchio: i don't get it I do counter 0..1..2.. formyself which is 3 power but my weapons load only until 2 xD
pb4: dbdr : I had googled with the wrong words :D
dbdr: yeah, that happens
dbdr: sometimes knowing the keywords is the hardest
mark7: programmer - pro-googler
mark7: :joy:
pb4: AntiSquid : to me anaconda is a big package with many things : python, some libs, spyder, jupyter, etc...
mark7: pb4 you are right
pb4: using anaconda might only mean using the python version that comes bundled with anaconda ?
pb4: (which is my case)
eulerscheZahl: my prof used to say: a programmer always is a librarian too
AntiSquid: yes, but when you run the code in jupyter you can interact with the vars easily
AntiSquid: no it doesn't mean you are stuck to one version
IjustKilledutoo: hello
IjustKilledutoo: how is everybody here
AntiSquid: create a new python environment and use as parameter the version of python you need to use for that new project
AntiSquid: you can have lots of diff environments
AntiSquid: https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html
AntiSquid: conda create -n myenv python=3.6
AntiSquid: conda activate myenv
ParticleBeam: Now have to remember how to code in C#...
reCurse: I'm surely a corner case but I find jupyter confusing to work with
reCurse: Hard to know what got executed and what didn't
reCurse: Always a mess
kovi: best strategy is not to submit
ThePythonian: Honestly, I've never had such trouble with Jupyter
Astrobytes: The few times I've used it I did find it clunky, maybe just doesn't suit my workflow
AntiSquid: few times!
pb4: My problem with jupyter is the difficulty to launch a bunch of code selectivly
pb4: like, those 5 cells only
pb4: I prefer spyder for this
reCurse: And god forbid you restart
AntiSquid: guess you need an extension for that pb4
AntiSquid: btw how's your bot reCurse ?
JasperV: is there a possibility to create a private leaderboard for clash of code? to do custom clash of codes with a group and have a score with just that group? now every custom clash is individual
eulerscheZahl: there is a slack bot for clash of code i think
eulerscheZahl: https://www.codingame.com/help/codingame-slackbot
eulerscheZahl: you can even select the mode there (fastest/shortest/reverse)
JasperV: very cool.
Envious: how do you split "00:00" into hours and minumute
JasperV: we're on ms teams though at the moment :)
JasperV: I'll look into the bot
AntiSquid: Envious https://lmgtfy.com/?q=python+manipulate+time
Kukiss: xD
Kukiss: I have a questions too regarding OoC Can I make an action with only Silence?
Envious: yup np, i solved it thank you though AntiSquid
Kukiss: I thought it is mandatory to make one "MOVE" every turn
jacek: it is :?
Kukiss: https://www.codingame.com/replay/446604105
Kukiss: I just print out opponents' commands and I cannot see any with "MOVE W| SILENCE"
JBM: it's mandatory to make one "action" per turn
Kukiss: Ok, thanks :)
Envious: anyone know why when i do :
Envious: I get a number thats not binary
Envious: assuming both are integers
mlomb: what do you mean "not binary"
Envious: oh i think its converting to decimal
rwilson: n1 | n2 is binary in terms of the bits.. but the result is still going to be an INT
Envious: so if i do 0001 and 0110
Envious: assuming first number is n1, second is n2
Envious: actually lets say string
Envious: so would it be something like:
Envious: but that included "0b" at the front
mlomb: int n1 = 1; // 0001 int n2 = 6; // 0110 int n3 = n1 | n2; // 0111 (7)
Envious: except it doesnt give 0111
Nagatwin: eulerscheZahl gib stats :D
Envious: a = "".join(bin(int(n_1)|int(n_2))) a.replace("0b","")
eulerscheZahl: stats on what?
eulerscheZahl: your bot?
Envious: n_1 is first binary number, n_2 is second
Nagatwin: Im in top15 now :D
Envious: why does that code not work
Nagatwin: joking :p
eulerscheZahl: ah :D
MadKnight: improve your top15 Nagatwin
eulerscheZahl: easy to add you...
Nagatwin: I might drown soon
Nagatwin: could have been a lucky submit
inoryy: what kind of stats are you tracking?
eulerscheZahl: https://imgur.com/a/O5QD3Ms
Nagatwin: :D
rwilson: Envious - a = f'{n3:04b}'
eulerscheZahl: as you see i haven't finished the UI :D
Nagatwin: how is that compared to others ?
rwilson: Envious , assuming you want 4 bit output based on your question/example.
eulerscheZahl: that's a secret for now, create your own stats :P
eulerscheZahl: that's my way to analyze my bot and compare to others. sorry, i'm in contest mode and that can be an advantage
Nagatwin: haha legit
JBM: is contest mode allowed for a contest creator? :P
inoryy: I like that you added the avatar
eulerscheZahl: yes, i'm allowed to participate
JBM: is there even anything to win
eulerscheZahl: i'm allowed to win
eulerscheZahl: honor for sure
eulerscheZahl: glory
eulerscheZahl: which i already have as a hypnotoad
inoryy: nerd points
JBM: maybe a bunch of eternal gratitude too
eulerscheZahl: i only knew about the game 2 days before the contest started
eulerscheZahl: not an advantage
JBM: 2 days' worth of advantage
inoryy: ah so your contribution is the typical paper co-author level :P
eulerscheZahl: 2 days that i used to extend an incomplete statement and add tooltips to the viewer
jacek: where is that frog
eulerscheZahl: and made the default AI stronger - but that doesn't help in my argumentation of not having any advantage
JBM: nobody ever understood what the default ai was there for
eulerscheZahl: bosses..
eulerscheZahl: same code
eulerscheZahl: but indeed my contribution was small
eulerscheZahl: wanted to work on the contest during the weekend
eulerscheZahl: and then it started on friday already :D
florinpop17: https://youtu.be/qrIldeXkrBI watch us play live (and join us if you want)
Astrobytes: CoC I believe
AntiSquid: COC
AntiSquid: JBM they are playing COC
Astrobytes: With donations...
JBM: i feel like i'm getting old
Astrobytes: Apparently streaming anything entitles you to ask for donations these days
JBM: and there i was, doing it for free
Astrobytes: lol, now you know. You'll make millions from streaming Haskell I'm sure.
AntiSquid: no offense but donating money to a coc player/streamer is really stupid
Astrobytes: I completely agree.
JBM: i suppose i've yet to see it to believe it
AntiSquid: the idea behind donations in general is to sort of appreciate/support the entertainment if you see value in it ...
AntiSquid: never donated anything btw
Astrobytes: Nor have I
Astrobytes: But this guy has his donations on screen rn
AntiSquid: there are some useful guideline videos on youtube, they ask to donate on patreon, i wouldn't mind donating for that
AntiSquid: or i would mind donating for something that is actually entertaining on twitch (very rare occurrence)
AntiSquid: wouldn't*
Astrobytes: I just don't like the whole idea. If people wanna donate it's up to them, I won't judge.
Astrobytes: Well, I will. Silently.
AntiSquid: i guess it's like with the ads, some people leave ads on to not affect videos on youtube then some content creators put the advert straight in the video so they get the money
JBM: a contest creator still around?
JBM: the converse question to the ambiguous end case doubt i had he other time just popped up in my head
Astrobytes: Add it to the forum thread JBM
Astrobytes: :)
icecream17: Uh... help? my if statement timeouts...
icecream17: Oh I see, nevermind
RockyMullet: Oh god, if I shot a torpedo the same turn an enemy surfaced I though I hit him and totally screwed my data hahaha
Ifthel: Ooh nice, I'll be sure to learn from your mistake
RockyMullet: I fail for the greater good
AntiSquid: i feel like starting over
YannT: in Clojure!
AntiSquid: if you stream clojure to silver
YannT: well I know the theory of this game now so I'd be a stream of me googling and figuring out language things :p
AntiSquid: same
Astrobytes: You have to charge 1 euro per 10 minutes
AntiSquid: alternatively implement it in Lisp and run it locally
Astrobytes: Tho it ith
AntiSquid: so it's a fair alternative
Nagatwin: is legend opening the 17th ?
AntiSquid: 10th
Nagatwin: rofl
Nagatwin: cache miss rofl
AntiSquid: 1 week of legend
Nagatwin: thanks
AntiSquid: you can click in IDE at the top on your rank and league
Nagatwin: yes i know
AntiSquid: it shows next league opening during contest
Nagatwin: but it showed 17th because of cached info
AntiSquid: ah
Nagatwin: How many players will be going to legend ? 20 or so ?
AntiSquid: 15 ?
AntiSquid: maybe 20, dunno
Nagatwin: Ill tryto keep up
RockyMullet: oh no, inorry passed gold ! I was testing against you :P
RockyMullet: I guess I'll have to be gold too then :triumph:
RockyMullet: well my bug fix got me from 48th silver to 21th, theres hope
inoryy: haha wow I got pushed over the day for ~40th place
inoryy: from*
RockyMullet: so youre telling me there's hope I get pushed too ? :D
RockyMullet: well I lost a lot against you, thats why I was testing against, so I might have pushed you a bit hehe
Zenoscave: whoa top 20 is super dense right now
Astrobytes: So am I, frankly :scream_cat:
icecream17: uh, i pressed a random keyboard thing and now the textcursor is a yellow rectangle
icecream17: Oh, turns out I pressed the "Insert" button. Didn't even know that existed
Astrobytes: ...
icecream17: I wonder what the other keys do....
Astrobytes: I heard the 'any' key is great if you wish to continue
icecream17: :grinning:
icecream17: wow i've never seen the PrtSc key
Nerchio: I just want to advance to silver and I will probably be done, my first competition and my bot became too big mess
Nerchio: on rare occasion it breaks and debugging it is hell since there is so many paths to track
Ifthel: If it's messy, this is a great opportunity to practice code refactoring Nerchio
Nerchio: I am refactoring some along the waybut i would probably have to rewrite this completely to make sense haha
Astrobytes: Still 12 days and a bit left ;)
Illedan: Yeah, too long -.-
**Ifthel hi
elsinnombre2d2: hi, im new here and id like to know if it is possible to do a machine learning approach
elsinnombre2d2: possible= feasible
RockyMullet: most people who try machine learning in here do it offline
RockyMullet: aka reproduce the game outside and train it
RockyMullet: since you cant keep data from game to game
Ifthel: Isn't there a way to get results from online matches?
Uljahn: you can download replays through API, then parse them
elsinnombre2d2: and how they import the ai?
Uljahn: ai?
elsinnombre2d2: Artificial Intelligence
elsinnombre2d2: the model
Uljahn: from __future__ import ai, Automaton2000
Automaton2000: for the sake of it
Uljahn: ide has limit of 100k chars, so you can copy your model's weights
icecream17: nice, turns out my code has all the letters of the alphabet
Swagboy: Hahah noice
MadKnight: hihi Swagboy icecream17
MadKnight: are u guys playing contest ?
Swagboy: Yes yes quite
Swagboy: working on a mcts
MadKnight: how're u gonna use mcts here ?
Swagboy: with branches weighted with probability of game states
Swagboy: But i'm still at the mcts-0 part
icecream17: oh hi yes im playing contest
icecream17: I just found a huge bug, lucky me
icecream17: unlucky me, i lost the first 8 games in a row
rwilson: icecream17 , but now... can you squish the bug quickly and get your bot doing better ;) ... ( or worse :( sometime )
icecream17: i'm not even sure there's a bug....
icecream17: oh wait nevermind
Swagboy: There was a bug or not then?
icecream17: many many bugs
icecream17: unbelievable
Swagboy: Nice in a way
Swagboy: that leaves room for improvement
RockyMullet: fight of the century https://www.codingame.com/replay/446659344
icecream17: lol
TidyMaze: excellent xD
icecream17: I got distracted by many bugs....
icecream17: but I finally figured out why my code was bad!
icecream17: submitting
icecream17: Huh? 186th is "HelpMePlis" and 187th is "HelpMePlz"