Chat:World/2021-01-17
 reCurse: Finally
 Coder_Raghuveer: https://www.codingame.com/clashofcode/clash/report/154867836128af97471cfd8327656e0c1fc1d39
 Coder_Raghuveer: see this
 Coder_Raghuveer: simplest clash
 Coder_Raghuveer: only i got 100%
 RyanGilbert: u played against all bots
 RyanGilbert: or nvm sorry i didnt read
 Mr.Ma: hi, anyone has built a genetic  algo model on'code vs zombies' ?
 jacek: good morning
 pipebits: 2 AM no morning here lol
 jacek: reBless doing d&b now? :scream:
 eulerscheZahl: happy 8:43am
 sayto: happy 9:50Am :p
 Salted: Happy 2:02 am!
 jacek: those canadians
 Salted: It's a great day in Canada, and therefore the world!
 jacek: this google suggestion: "if you die in canada you die in real life"
 firepro20: I'm stuck at power of thor optimal angle test case. any hints?
 KiwiTae: firepro20 fix it!
 k4ng0u: firepro20 when you have both a vertical and horizontal diff beween thor and the lightning are you moving in diagonal (1 move SE instead of 1 move S + 1move E)?
 AlgoRhythmus: hi nerds! ;P
 nqsa00: hi
 firepro20: k4ng0u I wasn't, I realised NW SW NE SE are more powerful moves. I think im on the right track now
 Stilgart: yes you are
 zhoubou: YES! I got 100% on Code vs Zombies with 70k. Finally found a bug in my code. Woohoo :D
 emetafora: :evergreen_tree:
 Stilgart: gg zhoubou
 zhoubou: I'm so thrilled and proud. It was such a stupid bug :P
 DomiKo: zhoubou good job
 zhoubou: I'm wondering if I can be better in Python
 jacek: can you show your python?
 jacek: python :v
 LinhT.Nguyen: what u mean show python
 LinhT.Nguyen: you mean python as pet?
 zhoubou: Yes, he means pet :sweat_smile:
 jacek: it code of course. if he made the code in D, I would ask him to show D
 LinhT.Nguyen: hmm this is strange i believe theres another meaning
 LinhT.Nguyen: this is so strange
 zhoubou: I would never code in D *sweats*
 LinhT.Nguyen: i thought people'd just use C instead of D
 zhoubou: Oh God this is not good
 Astrobytes: jacek are you taking the P out of his D?
 LinhT.Nguyen: just programming languages
 zhoubou: I heard C++ was better
 jacek: oh my https://en.wikipedia.org/wiki/P_(programming_language)
 LinhT.Nguyen: the world of programming is amazing
 Astrobytes: Did he get his python out? Or show you his D yet jacek?
 LinhT.Nguyen: you mean the code?
 Astrobytes: Of course.
 jacek: yes. its over 100k limit :o
 **BlaiseEbuth slaps Astrobytes arround a bit with a large :eggplant:
 Astrobytes: :grin:
 zhoubou: Aanyways, Astrobytes. Student finally beat his teacher :upside_down: I'm 520th :D
 Astrobytes: Ayyy, congrats zhoubou!
 eulerscheZahl: your teacher is on CG too?
 Astrobytes: You had a bug in the end?
 zhoubou: Yeah, it was a stupid one too
 Astrobytes: Most bugs are tbh
 eulerscheZahl: that's a better way to parse it than a stupid teacher :P
 Astrobytes: lol
 zhoubou: I forgot that if 2 zombies are on 1 human, human count is decremented only once lol
 zhoubou: I mean, I didn't account for it
 Astrobytes: depriving yourself of points eh
 zhoubou: I'm wondering if NumPy might increase the score though
 zhoubou: Never used NumPy, so this could be a good reason to get into it
 zhoubou: Or maybe there aren't enough vectors and dimensions to make a difference
 eulerscheZahl: https://github.com/glouw/tinn/blob/master/Tinn.c#L119
->
https://github.com/dvdplm/rustinn/blob/master/src/lib.rs#L75
 eulerscheZahl: something doesn't look right in this port
 BlaiseEbuth: The name ? :p
 eulerscheZahl: the array size in the lines i linked
 eulerscheZahl: some pointer ninja magic in C
and the one who ported it to Rust didn't really understand what was going on
 jacek: common with the rusties
 BlaiseEbuth: Hmm. But I think the name is a greater problem.
 eulerscheZahl: the naming is annoying, yes
 eulerscheZahl: 1 letter vars
 eulerscheZahl: and what's "nips"
numberOfInputs it is
 BlaiseEbuth: I meant the lib's name :p
 jacek: nip could mean something else
 jacek: well its tiny after all
 BlaiseEbuth: nvm, the joke doesn't works in english...
 BlaiseEbuth: too bad...
 eulerscheZahl: something with tintin and melou?
 BlaiseEbuth: Nope. in french "rustine" mean "patch"
 BlaiseEbuth: And we say "mettre des rustines" when a project is barely fixed with approximative solutions...
 eulerscheZahl: yeah, joke completely gets lost in translation
 BlaiseEbuth: Yeah... I thought there was a "rustin" like english word... Meh
 n4gb07: where can we put collision pod radius in code to avoid collisions
 BlaiseEbuth: At the right place.
 eulerscheZahl: if you aren't in gold or maybe even legend yet, don't bother with collisions for now
 derp22: Do damaging items bypass ward in LCM?
 eulerscheZahl: in least common multiple?
 Stilgart: euleur: locam
 eulerscheZahl: aah
 eulerscheZahl: but what's euleur?
 Stilgart: just me being brainless
 BlaiseEbuth: Heulleeuurrrrrr with rolled R
 eulerscheZahl: locam is so long ago. I don't even remember what a ward is
 eulerscheZahl: now I got the question at least. But I don't remember, looking for replays where it happens...
 derp22: Don't bother, I can do that. Just thought I'd ask first to see if someone who knows and remembers happens to be around
 eulerscheZahl: https://www.codingame.com/share-replay/484942826
is frame 76 what you are looking for?
 eulerscheZahl: the contest was 3 years ago. I remember nothing
 derp22: The item also removes ward so I'm not sure if that counts
 eulerscheZahl: true
 eulerscheZahl: i think otherwise it would remove the ward and not deal any damage
 eulerscheZahl: just like an attack from a normal card
 derp22: I'll just assume that and see how my eval works
 DomiKo: https://github.com/CodinGame/LegendsOfCodeAndMagic
 eulerscheZahl: that's an RTFM move :D
 DomiKo: you can check it there
 zhoubou: 7th in Python :scream:
 zhoubou: Sorry for spamming the chat. I'm so proud of myself lol
 BlaiseEbuth: :clap:
 BlaiseEbuth: Sorry, there was a mosquito.
 DomiKo: zhoubou you know you can hardcode some testcases?
 zhoubou: I know the validators are available, but I don't want to use them for some reason :P
 DomiKo: nahhh
 zhoubou: What does hardcoding a testcase mean then?
 DomiKo: exactly this
 DomiKo: if I remember there are 3 testcases
 DomiKo: where you can do BIG combo
 DomiKo: and you can either spam submit
 DomiKo: or hardcode them
 jacek: :F
 zhoubou: Well I do neither. I fix bugs and submit once :P
 zhoubou: But I guess I could hardcode them
 zhoubou: So I would run sims offline until I get a very high score?
 DomiKo: yes
 zhoubou: Hmm okay. I guess that is my next step. But it kind of feels like cheating
 DomiKo: not much
 DomiKo: because it's not really easy to find that big combo
 struct: its either that or spam submit
 struct: and hoping you can get the combo
 zhoubou: Guess you're right, yeah
 DomiKo: and 
 DomiKo: you are using python
 zhoubou: Which is slow, I know
 DomiKo: so you have low number of sims versus c++
 DomiKo: so it's like a little help for your python program
 zhoubou: Yeah, you're absolutely right
 BlaiseEbuth: :see_no_evil:
 jacek: sounds like opening books
 BlaiseEbuth: "My job does not pay so much, so I can steal some food. That's not really illegal, more like a little help for live." :smirk:
 DomiKo: :money_mouth:
 RyanGilbert: Could some people mind looking at my contribution and giving feedback or approving it?
 RyanGilbert: https://www.codingame.com/contribute/view/5956223929b6a2443a0b3816f7b421beea07?comment=38741
 eulerscheZahl: as soon as you review mine https://www.codingame.com/contribute/view/5118d71262a5772672a3a604a0a53942eb12
 Zorg1: ^^
 eulerscheZahl: checkmate :P
 jacek: catchy
 Illedan: Wow, I actually used time to debug why my last bit was not set to 0 :D Guess I shouldn't be coding today. 
 jacek: and why was that
 Illedan: It was set to 0. It was just the tostring didn't print the leading 0s :P 
 eulerscheZahl: because smits did the bitpacking for him
 eulerscheZahl: :D
 eulerscheZahl: then it's the first bit, isn't it?
 Illedan: Index 63
 eulerscheZahl: ah got it now
 Illedan: https://tech.io/snippet/wexEwY5
 eulerscheZahl: you didn't count bits
 eulerscheZahl: what are you up to?
 Illedan: I'll keep it for myself :P Better show by doing than by telling and not finishing
 eulerscheZahl: number shifting
 eulerscheZahl: just kidding, won't happen
you try the NN hype train too
 jacek: hype?
 Illedan: Where is the hype?
 Astrobytes: The hype's sleeping
 eulerscheZahl: right, the only hype is about CoC
 struct: they just need to convert the 5%, then everyone is happy
 reCurse: The training train 🎵
 eulerscheZahl: :train:
 eulerscheZahl: that doesn't look like a train at all
 eulerscheZahl: congrats for beating that horrible league2 btw ;)
 Illedan: What was the binary representation of int in C# again. Head not on the right place :P
 eulerscheZahl: ?
 Illedan: 0x111 => Hex
Binary?
 eulerscheZahl: 0b10101010
 Illedan: thx
 eulerscheZahl: and don't forget about the underscores for formatting
 jacek: huh, youre actually gonna NN the d&b?
 Illedan: Those I wont forget. I just couldn't figure out what that annotation is called to google it properly
 eulerscheZahl: so CG chat beats google in this case
 reCurse: Ah yes, the good old "you can only search when you know"
 eulerscheZahl: but it's pretty easy to code in java or JS when you know C# and have a stable internet connection
 reCurse: True, though why the "when you know c#" part?
 eulerscheZahl: definitely missing some syntactic sugar but at least getting working code
 eulerscheZahl: or have some prior knowledge in any other coding language
 reCurse: Oh, true
 eulerscheZahl: sure works without too. but less efficient
 reCurse: I guess C# is quite good for discovering common concepts actually
 eulerscheZahl: abstract enough that you can concentrate on algorithms and design patterns and not deal with pointers
 Knee-Gears: ey euler i have a question
 reCurse: It did get me acquainted with functional concepts
 eulerscheZahl: number 3 on the list again, i have that already
 eulerscheZahl: go ahead Knee-Gears
 Knee-Gears: i learn stuff and understand it, but when i do a clash here I just frick everything up
 Knee-Gears: how do i stop that?
 Illedan: frick?
 jacek: stop clashing then
 reCurse: Red X button, top right
 jacek: [solved]
 Knee-Gears: xD
 eulerscheZahl: keep training. there is no other way i think
 Knee-Gears: okok
 Illedan: Better do easy puzzles
 Illedan: Because once you solve them, you can see other solutions
 Illedan: and learn
 Knee-Gears: ok thxx
 reCurse: Keep in mind doing clashes only gets you good at doing clashes
 eulerscheZahl: exercise is really important to see reoccurring patterns
 eulerscheZahl: good point from reCurse
 Illedan: Indeed, I have never needed to code some string pattern creation at work
 eulerscheZahl: but you also know a bit of the language (loop but also lists, dictionaries, ...) if you don't know them already
 reCurse: Though with some far stretching you could say dealing with reverse is like dealing with clients
 Illedan: haha
 eulerscheZahl: :D
 Illedan: And submitting what you think work and only getting 50 % is the perfect analogy 
 kovi: i wonder when recurse own clashes with NN
 reCurse: For? The captcha part?
 eulerscheZahl: and solving the captchas ...
 eulerscheZahl: yeah
 eulerscheZahl: known pool of questions makes it meh
 reCurse: Why NN when you can DB
 kovi: well, i never clash, so i dont know how repetitive those are
 eulerscheZahl: 954 clashes in my DB
 eulerscheZahl: some may be removed by now (low rating = delete)
 struct: how many puzzles?
 eulerscheZahl: 1390 total contributions, subtract 954 clashes
 eulerscheZahl: but that's only community contributions, no official puzzles
 AntiSquid: DB ?
 reCurse: database
 BlaiseEbuth: Double Bacon
 AntiSquid: ah ok that makes sense BlaiseEbuth
 BlaiseEbuth: :ok_hand:
 eulerscheZahl: dbdr
 BlaiseEbuth: ho
 Salted: anyone know how to format ascii art for feedback in a comment on a puzzle contribution?
 Salted: http://chat.codingame.com/pastebin/8c1a9997-0da6-4975-920d-482f5f55e535
 AntiSquid: "i want to shitpost, but lack the talent"
 Salted: It's helpful I think
 eulerscheZahl: try nbsp
 AntiSquid: ah nvm, that's not very entertaining ascii art
 eulerscheZahl: not sure if it works
 Salted: Hmm, I just used the pastebin link. Maybe that will just work for a bit
 Angecide: ok I'm totally feeling it, this might be the submit that gets me to legend in uttt
 Angecide: 32 point, not even half ways, something has to be totally rigged if i dont make it
 struct: How many points does boss have?
 Angecide: too many, I am stagnating
 jacek: it reminds me of someone https://img-9gag-fun.9cache.com/photo/axM2Pzb_460swp.webp
 jacek: are you sure you sent new bot? :v
 Angecide: legend is impossible
 n4gb07: got 27.12 in this round.. i will be in gold after an hour
 Angecide: submitting a bot is like gambling
 Angecide: where you pay with your time
 reCurse: :clap:
 Angecide: hm, I guess it takes time for it to update on the actual leaderboard
 Angecide: but i did get the notification
 struct: you need to wait for promotion
 struct: Older multies have that problem
 VizGhar: congrats :) MCTS?
 Angecide: yea, a really shitty one I think
 VizGhar: hmm what language?
 Angecide: it updated, oficially legend on the leaderboard
 VizGhar: well ask reCurse, I've seen his post somewhere about how you should use C++ on CG
 VizGhar: key is no STD
 VizGhar: so you might try ansi C right away :D
 Angecide: yeye I am doing BMI2 and stuff like that 
 reCurse: uh?
 derp22: If you have an STD, you're fucked. So the key is to not have one
 jacek: oO
 Angecide: I started learning c++ 4 months ago when I reached a bottleneck with python
 VizGhar: ah it wasn't reCurse sry :| https://www.codingame.com/forum/t/c-and-the-o3-compilation-flag/1670
 VizGhar: :D you reached it pretty fast didn't you?
 Angecide: I guess, but I did ask for some help occasionally on the discord server and also a bit in here, which helped me a lot
 Astrobytes: You were already going in the right direction, well done
 VizGhar: here, have a rank of Jedi knight... I'm not 5% so not looking for competitions that much. But earning legend in such short time is nice
 Astrobytes: Some people just go for certain multis or optims (or puzzles), so XP and rank is not always a helpful metric
 VizGhar: If I go for multi. I'm going for such multis, I can show to my wife :D yeah I'm more like puzzle guy
 zhoubou: I'm currently following the quest map. And getting hung up on getting better scores, while I'm at it.
 zhoubou: I think I'm done with CvZ. 180k is good enough for now I think
 darkhorse64: stl bad for CG is an hoax
 zhoubou: They are brainwashing you through puzzles.
 darkhorse64: The truth is many people don't know how containers are behaving and use them wrongly
 RoboStac: there were real issues with it on that version of gcc (4.9). 
 Severed: Is the Power of Thor game broken for anyone else? Giving me wrong coordinates for the player
 derp22: https://www.codingame.com/training/easy/power-of-thor-episode-1  this one?
 Severed: yes, making a post there now.
 darkhorse64: Look at the number of people who solved it. The problem is in your code
 SPDene: my solution from several months ago still passes everything
 Severed: No. all Ive done is run debug to get the coordiantes
 Severed: its assigning the lights Y as the players Y as well, which its not
 darkhorse64: You have to update Thor coordinates yourself
 Severed: No, you tell him to go a direction
 Severed: His location is Y=0, but debug says its 4
 darkhorse64: You have an error reading the inputs or you issue too many commands per turn
 Severed: im doing nothing yet besides running 4 lines of debug to get each X and Y
 Severed: and its wrong
 BlaiseEbuth: darkhorse64, listen to him ! He KNOWS. So stop saying bulshits to him.
 RoboStac: if you want to paste your code in here we can tell you whats wrong
 Severed: http://chat.codingame.com/pastebin/a9ae6ec8-952c-4d8b-a7fc-65bfe0ba493b
 zhoubou: You're not updating the coordinates.
 darkhorse64: BlaiseEbuth: no bullshit, he is doing exactly what I said, too many commands
 SPDene: For "Straight Line" test case, 4 is the correct Y coordinate
 BlaiseEbuth: No. You're wrong darkhorse64. He KNOWS.
 struct: he is printing to stderr
 struct: he only has 1 print
 struct: to stdout
 Astrobytes: Ferme ta bouche BlaiseEbuth :rofl:
 darkhorse64: Sorry, my bad, shoud go to sleep
 darkhorse64: *should
 zhoubou: He is just not updating x,y coordinates of Thor
 Severed: Oh, so map is 45 degree?
 struct: Severed the coordinates are only read one time
 Astrobytes: Severed: You are not updating Thors coordinates, but you are printing "NE" every turn
 struct: you need to update them yuorself
 Astrobytes: So Thor is moving, but you are not updating the position as has been said
 mzbear: Severed, press the rewind button below the display to get to first frame. Check thor's coordinates again. They should match now
 Severed: Gotcha, I was looking at map wrong.
 Severed: Thanks
 mzbear: does Thor have debug mode in the display?  A lot of puzzles have it, you can enable debug and hover things to see their coordinates
 zhoubou: I don't think it does
 mzbear: That's a shame
 SPDene: not a "debug mode", but it does give "game information: Thor position = (31,4). Light position = (31,4). Energy = 74
 Astrobytes: Just be careful you don't get a Thor finger from pressing 'Submit' too often.
 darkhorse64: OK, I have been a bit blunt, sorry about that but my first hint was the right one
 darkhorse64: Hammered by captchas
 BlaiseEbuth: The man who already KNOWS don't need hint.
 Astrobytes: darkhorse64: Don't worry, BlaiseEbuth is taking the piss :D
 Astrobytes: So out of character! :P
 darkhorse64: Bad day ?
 Westicles: ^^ I can't tell if Salted is trolling me or not
 BlaiseEbuth: Trollception !
 Astrobytes: hehehehe
 mzbear: everybody be kungfu trolling
 Astrobytes: "Hah! Who? Wha? Who's that?"
 Westicles: https://www.codingame.com/contribute/view/597138ad120ca73e7e45ee52da5a280aba3d
 Astrobytes: lol, must have misunderstood
 mzbear: meanwhile, i am very sad at There is no Spoon ep 2, since my code says the Expert testcase is unsolvable :(
 BlaiseEbuth: If you code say that...
 BlaiseEbuth: It must KNOW !
 Astrobytes: THEN IT MUST BE TRUE. You should make a forum post immediately.
 Astrobytes: And ping all the staff on discord.
 Astrobytes: And PM euler.
 Astrobytes: Obviously.
 VizGhar: lol :)
 BlaiseEbuth: AutomatonNN ! Red alert !
 AutomatonNN: annoying news about that :)
 VizGhar: expert testcase is last one? it was much easier to solve then CG testcase
 Astrobytes: lol NN speaks sense in World for once!
 VizGhar: AutomatonNN is right guys o.O
 AutomatonNN: bend the first one
 mzbear: viz, the CG testcase times out currently, but expert testcase doesn't, so i'm looking at fixing expert case first.  just can't figure out what's wrong with my code ;_;
 BlaiseEbuth: AutomatonNN ! Ping euler !
 AutomatonNN: eulerscheZahl is there a way to see the code and it will say that the compiler will probably create 
 BlaiseEbuth: Thx
 SPDene: Westicles I think it might help to say A-Z represent the DEPTH of the pool, not the height - if it was "height", 26 would be above 25
 SPDene: and if he doesn't see that it's the other way round, it's VERY confusing
 VizGhar: mzbear maybe you are computing closed graphs incorectly?
 Westicles: SPDene, 26 is above 25...
 SPDene: then it's really confusing... you say "Outside the pool the height is 0", and the uotside isa square of Zs
 mzbear: nah, something else is wrong, because my algorithm says there's a node that cannot be satisfied no matter what.  so perhaps my heuristics are buggy.  but i've gone through them several times already.  maybe i'll need to postpone this for another day >_<
 Westicles: The Z's are the edge of the pool, not outside
 Astrobytes: Yeah I think depth would solve the issue. I got it right away but it may help to clarify.
 SPDene: that's really confusing
 Westicles: Nobody been to a swimming pool?  :P
 Astrobytes: lol, no I get it. Some might not though.
 VizGhar: bear have a rest :) ... I solved all except CG case in 2 days. then waited 3 months to finally solve the CG in 1hr
 SPDene: I've never been to a swimming pool when you have to climb up 26 units to get in
 Astrobytes: Getting into my local pool is a fair whack above ground level tbh
 Astrobytes: Anyway, I do think depth would be more appropriate.
 Westicles: https://www.homedepot.com/b/Outdoors-Pools-Above-Ground-Pools/N-5yc1vZcfyt
 VizGhar: mzbear maybe solution will help you somehow https://pasteboard.co/JK7rZrs.png
 Westicles: Depth isn't accurate though
 Astrobytes: Eh, I retract all my previous comments.
 Astrobytes: "above-ground swimming pools"
 Astrobytes: My bad, I missed that.
 Astrobytes: Height is definitely OK.
 Astrobytes: (imo ofc)
 mzbear: VizGhar, it helps a bit, it shows that my heuristics are correct up until it starts the search
 mzbear: if i help my code by hardcoding one of the decisive links in, it finds the answer correctly.  ugh
 struct: These above ground pools seem a bit expensive
 struct: Is there a similar to coc but for photoshop?
 derp22: How would the images be validated?
 mzbear: ffs lol, this is the graph continuity problem after all. This incorrectly flags a perfectly good solution as incorrect
 VizGhar: :thumbsup: told ya :)
 mzbear: i have my own disjoint set implementation and i cant find a bug in it, yet it must have a bug ...
 mzbear: lol, i went to the bathroom and immediately realized where the bug was.  the debugging power of a bathroom break is amazing
 mzbear: right in time before network errors hit
 VizGhar: :poop:
 VizGhar: I see you have solved Shadows of the knight 2... have u used the approach - resolve x first y second, or something more sophisticated?
 struct: x first y second works
 mzbear: i recall i solved it one axis at a time, yes
 mzbear: now it's only the CG case left...
 VizGhar: that's too easy :| I tried Trilateration, but with not much success... Even debug mode marked final node as not solution :D
 VizGhar: sometimes
 VizGhar: I'll solve it on both axis separately then
 VizGhar: It works, but I can't instruct properly where to jump exactly... to costly operation
 VizGhar: *too
 mzbear: oh, i just needed to add one more heuristic and now the CG case solves too
 mzbear: there we go, spoon ep 2 done.  not the prettiest possible code but whatever
 mzbear: and while reading someone else's code, i realize i did so many stupid things in my code :D
 mzbear: i'm upset at how much better DaNinja's code is compared to mine
 mzbear: ohwell
 Westicles: That's how I was with Binary neural network Part 1.  Some big solution then I see a guy did it in 9 loc
 Westicles: Really just one line if you take out the io
 MeNubz: Why sometimes codingame gives me 75% after I have already passed all the code?
 reCurse: Validators are slightly different
 Wontonimo: to see if your code solves the general problem (as opposed to just the visible tests)
 Minato_N: hi
 Minato_N: sup
 Wontonimo: hi
 Wontonimo: what programming language do you use Minato?
 PetarVasiljevic: why is this so easy
 PetarVasiljevic: this clash of code thing man
 mylittlestojan: ide gas