Chat:World/2021-08-04
BlackGriffin03: In Horse-Racing Duals , shouldn't the second test case answer be 2 because 17 & 15, is the closet strength with minimum strength difference
BlackGriffin03: But the correct answer according to the test seems to be 1 .How is is it possible ???
eulerscheZahl: because there is a horse with 6 and one with 7
eulerscheZahl: and some more pairs with distance 1
BlackGriffin03: okay :thumbs_up:
Queen29: I don’t understand this stuff.
KiwiTae: o/
tealfrog: How does Edit option in CoC puzzle approval work? Is it like a suggested changes?
Batfly: One question about C++ : why " int h = list[ i ] ; " don't work ?
Batfly: list is... a std::list<int> I created
Batfly: (Oh, and I forget to say : Hi ^^" )
KiwiTae: Batfly use std::vector instead
KiwiTae: u can't access list element like that
Batfly: oh, ok, that's work ^^ thanks
Batfly: (Still, why using list if vector work better ? :s )
RoboStac: list is a linked list so you can insert / remove elements in the middle very quickly but indexing to elements is slow. Vector is a dynamic array so can only add / remove elements at the end quickly but indexing is very quick
Batfly: okey, so it's just a question of convenient speed . Thanks ^^
HornedRabbit: Hi everyone. I have a question about Racing Duals optimization thing. I'm relatively new and I really don't have an idea for how to sort out this massive list without timng out...
derjack: you said it yourself
derjack: 'sort'
HornedRabbit: Well. Any sorting i try to do still results in timing out...
SlayerOfTheBad: What sorting algorithm are you using?
HornedRabbit: I tried "quick sort", but I figure it's not a right solution gere.
HornedRabbit: Or maybe i just screwed the implementation...
derjack: what language
SlayerOfTheBad: Probably that second one, implementing sortin galgorithms can be tricky.
SlayerOfTheBad: Yeah what language.
derjack: and do you still have nested for loops
derjack: once sorted, no need for that
derjack: you can compare horse and horse+1 in single loop
SlayerOfTheBad: You could try merge sort if you're comfortable with recursion.
HornedRabbit: Thanks. I'll try to work more with the sorting and possibly get back once i get somewhere.
Uljahn: use built-in sort
Uljahn: oh, chat froze
SlayerOfTheBad: If the language has it available.
Uljahn: which language does not have it?
SlayerOfTheBad: Far as I'm aware at least.
HornedRabbit: thanks Slayer. Merge sort did it. Now I'll try a few more things to take it apart and understand it better.
SlayerOfTheBad: Glad to hear it.:)
SlayerOfTheBad: Glad to hear it. :)
SlayerOfTheBad: Oh, no editing, only repeat messages huh? good to know.
derjack: oO
Illedan: Hi
Wontonimo: roll-over seems like it is fixed, or is it? :unamused: still cutting off when near the bottom or edge
Wontonimo: hello fellow devs to you too Vortakis and Jothegeier
Xaniouks: Hi guys, is there a way to do some profiling on the code to see what the worst parts are?
darkhorse64: dapo - 9h30 ?
Astrobytes: he has or had a bot in bronze, so I guess yes
Astrobytes: bronze, what am I on about
Astrobytes: don't listen to me, but I think so yes
darkhorse64: ironically, the sogeti challenge has been highly beneficial to my FB ranking (#42 to #23) but I believe the top has improved even more their eval
BlaiseEbuth: Hey there.
BlaiseEbuth: Speaking sogeti, how do you discretize the possibles actions ?
BlaiseEbuth: Eval ? What algo ?
darkhorse64: To be fairly honest, my code is based on Magus sim/GA and pb4 eval with various performance improvements
Magus: yeah my code is alsed based on Magus sim/GA
Magus: *also
darkhorse64: but much more improved
BlaiseEbuth: I thought about ga too... But someone spoke about mcts, and I really don't see how to discretize enough...
Illedan: ga ftw
darkhorse64: You could discretize the same way you do for a GA but the branching would be overkill
darkhorse64: afk for a while
BlaiseEbuth: Anyway, no matter the algo, it's a little utopian to imagine I'll get it working before contest end. :p
AllYourTrees: woohoo top 10 in C4! :scream:
BlaiseEbuth: Top 10 on 11 players? :o gg!
AllYourTrees: :sob:
BlaiseEbuth: Na... 298! gg AllYourTop
AllYourTrees: ty Blaise :P
ProCoder03: what's C4 ??:thinking:
ZarthaxX: connect 4
BlaiseEbuth: :bomb:
ZarthaxX: or that :P
ProCoder03: thanks
BlaiseEbuth: Wut. CG put my arm to uppercase! :scream:
KiwiCoder69: new to python and i just beat power of thor! :smile:
ProCoder03: I fail in temperature code golf validator 1 & 12 , is it a problem with size of code
darkhorse64: No, your golfed code does not provde a valid answer
KiwiTae: KiwiCoder69 o/ are you a fan?hehe
KiwiCoder69: ?
KiwiCoder69: ?
KiwiTae: because we both kiwis
ProCoder03: :thumbs_up:
ProCoder03: darkhorse34
LazyMammal: @ProCoder03 it could also be slow server (if it is a timeout). retry and see if the same test cases fail
ProCoder03: Actually now, it seems the servers are all heavy because every validator execpt for third is failing
MSmits: err temperatures requires no heavy calculation I think?
MSmits: my meta mcts for oware with 36 seed endgame book is functional btw
MSmits: now I need to wait a while to get enough data
MSmits: assuming i don't run into more bugs. This stuff is hard
MSmits: earlier i ran into a problem where my windows forms text box was updating too much because games run so fast
MSmits: led to crashes.
reCurse: You should stick your computation stuff to a console app
reCurse: UI is sometimes a bottleneck
MSmits: that's a really good idea
MSmits: I might do that once I am sure there are no more bugs
MSmits: the ui makes it easier to spot them
MSmits: what confuses me a bit is that I have a 10 core cpu and if I run 10 games at the same time, they all use 6.7% of the total cpu. So 67% of core so to speak. I know cores and threads are different things, but it always seems to be 6.7% somehow
MSmits: thats what the task manager tells me anyway, no idea how accurate that even is
MSmits: maybe it's hyperthreading or something that accounts for the lower cpu use
MSmits: or the UI is just bottlenecking things hard like you said...
derjack: 6.7% for 10 threads? theres some (i/o?) bottleneck for sure
MSmits: per thread
MSmits: so 67% of total cpu
MSmits: "wasting" a third
eulerscheZahl: wasting 2 thrids
MSmits: but possibly not, as i dont know for sure why this is the case
MSmits: lol
MSmits: you know what i mean :P
eulerscheZahl: :P
eulerscheZahl: your constant chatting made me take a quick look at dots and boxes btw
MSmits: oh cool
eulerscheZahl: i won't go into that rabbit hole
MSmits: no it's really something else
eulerscheZahl: and my eyes hurt with that black background and yellow dots
MSmits: thats what re curse said
MSmits: didnt bother me too much
eulerscheZahl: maybe your screen isn't good enough to show the black properly
MSmits: who knows, maybe :)
eulerscheZahl: UI rule: never use #000000 for black
MSmits: why is that
MSmits: too much contrast?
eulerscheZahl: at least not for bigger areas
eulerscheZahl: hard for the eyes when you look at it for a while
MSmits: mmh ok
jacek: thats racist
MSmits: well I dont often look at those replays anyway, I was mostly coding and looking at my own pretty map outputs
jacek: MSmits have you seen https://www.codingame.com/multiplayer/bot-programming/abalone
eulerscheZahl: a hex board
MSmits: i've seen yeah
MSmits: only 5 players so far
eulerscheZahl: and the start setup where wikipedia suggests it's prone to passive gameplay
jacek: you can be top10 easily
MSmits: lol
MSmits: thanks jacek
MSmits: eulerscheZahl you mean he should have gone with a different setup?
jacek: passive opening
MSmits: what does that mean?
eulerscheZahl: there are different ways to arrange the spheres at the start of the game
MSmits: right
jacek: inactive opening
MSmits: I guess i dont know how to define "active" in this game
eulerscheZahl: i read a bit about it because one day i considered porting that game too. decided against partially because of that possibly passive gameplay
MSmits: a jrke just appeared on that leaderboard
jrke: yes
jacek: so lets make meta mcts
jrke: random.choice(moves)
MSmits: no thanks, i need my cpu for my oware NN data generation
MSmits: and later for training i suppose
MSmits: jrke random moves are better than bad moves
eulerscheZahl: is it me or does the leaderboard only show the first player? and load the rest when you scroll
MSmits: happens to me all the time
MSmits: also sometimes if you wait
MSmits: instead of scrolling
Illedan: Same for me euler
eulerscheZahl: i'm inpatient
eulerscheZahl: ah, there it loaded without scrolling
MSmits: in which hospital?
eulerscheZahl: :/
eulerscheZahl: is that bug reported already?
MSmits: i havent
**eulerscheZahl uses Chromium. When I say that, they close the issue right away (only firefox + chrome are supported)
MSmits: I see
MSmits: well make the bug report and says msmits and illedan have the same issue on chrome
jacek: try ie
eulerscheZahl: ok
eulerscheZahl: i try to write a slightly more detailed bug report
eulerscheZahl: "When opening a leaderboard, only the first player will be shown. The rest are rendered when scrolling or waiting a few seconds only."
eulerscheZahl: illedan: "Leaderboards only load the first player."
Astrobytes: Minimalist! :P
jacek: jacek: "not enough ponies"
eulerscheZahl: he was faster
eulerscheZahl: now i won't report
Illedan: Maybe I can shorten to "Leaderboards"?
Astrobytes: LB
Illedan: BUG
eulerscheZahl: "fix whatever the hell is broken"
MSmits: or just: :bug:
Astrobytes: Just !
eulerscheZahl: and illedan reacted with :plus: to his own bug report :D
Astrobytes: lol
MSmits: good, reinforce that
jacek: now add the eyes
Illedan: So you didn't have to find the proper emoji
Illedan: You clicked it instantly
Illedan: #win
eulerscheZahl: sometimes i accidentally react with plug instead of plus
Illedan: Fixed
jrke: so a laptop confirmed for pb 4
eulerscheZahl: wait if dapowill join
Astrobytes: dapowill? His brother? :P
eulerscheZahl: missing space
Astrobytes: You don't say!
jrke: how much time is left for that contest?
Astrobytes: ~5.5 hours
jrke: oh
Astrobytes: ~5.0 hours my bad
Illedan: Code wont climb up again :( juli is harder to beat now -.-
Victor_Baron: Heya
JHowell: How can I see how many non-private clashes I've done? I got the 50-clash quest (a lot of them were private) but I still can't approve/reject clash contributions.
eulerscheZahl: https://www.codingame.com/leaderboards/clash/global
eulerscheZahl: you have 27
JHowell: Thanks!
jrke: btw i have made a clash bot its working should i report this to someone?
eulerscheZahl: why report it?
BlaiseEbuth: Burn him!
eulerscheZahl: it's known to be possible. that's why captchas got added
BlaiseEbuth: Yeah of course...
Astrobytes: Just don't use it.
eulerscheZahl: +
jrke: captchas are now bit rare event as far as i discovered in clashes till now
jrke: yeah i won't
eulerscheZahl: tradeoff between stopping bots and not annoying human players too much
martinpapa69: after 2 days of research.. my quantized dense layer is working... :D totally worth it
FalINTOblivion0112: my ranking is going down...
FalINTOblivion0112: :disappointed_relieved:
AntiSquid: should i make a tutorial on how to make clash bots in a really easy way ?
jrke: which library?
AntiSquid: you're overthinking it already
jrke: selenium?
jacek: magnesium!
eulerscheZahl: use a human bot? outsource the clashing work to someone from fiverr
Westicles: another mod dancing around the precipice
AntiSquid: what like World of Warcraft gold farmers from china? that's basically close to slavery
AntiSquid: either way it's a bot programming site, if people can't figure it on their own, i guess they lack creativity :P Westicles
eulerscheZahl: may I push him and yell THIS IS SPARTA?
AntiSquid: what's better on CV : done clash of code or done a bot for clash !!
xxd: bunch of kids in a basement googling solutions for you
xxd: a Bot Programming challenge where the game is just clash of code
KiwiTae: Daporan a pas submit sur sogeti?
KiwiTae: oups:(
KiwiTae: french sorry
FalINTOblivion0112: if i can create an undefeatable clash of code bot...
eulerscheZahl: and then let it play against itself...
KiwiTae: a better metric would be how long can the bot survive before getting spotted
FalINTOblivion0112: right
FalINTOblivion0112: yes
jrke: i think clash of code bot won't work in shortest mode
jrke: efficiently*
xxd: why
jrke: atleast not the one i have made
SlayerOfTheBad: Is it the case some puzzles do not display console output? I'm expecting either some debug prints or at least a compilation error but I'm not getting anything but the test case is failing regardless.
Wontonimo: you probably did something wrong then
SlayerOfTheBad: Yeah I got that far, but without console output I can't really see what I did wrong.
Wontonimo: is it the test or validation?
Wontonimo: put a print right at the top and see if that does anything
SlayerOfTheBad: I did, and it does not print anything, so I'm assuming it's a compilation issue.
Wontonimo: if it doesn't compile, you should see a red error message.
SlayerOfTheBad: But it's not giving anything.
Wontonimo: what language?
Wontonimo: sorry, I know a few languages but not rust. is there an end-of-line token in rust? like newline?
SlayerOfTheBad: After locally compiling it I figured out it does indeed have compilation errors, but in previous puzzles these showed up in the console.
SlayerOfTheBad: '\n' generally.
eulerscheZahl: i should not try to compute -int.MinValue :/
jacek: huh
SlayerOfTheBad: compute uint.MinValue instead. It's easier.
FalINTOblivion0112: fyi its my birthday today
FalINTOblivion0112: remember
LeviathanProgramming: happy budday
FalINTOblivion0112: august 4th is my birthday
FalINTOblivion0112: thank you everybody here at codingame
LeviathanProgramming: obviously
FalINTOblivion0112: you guys are awesome :grinning:
jacek: we are? :thinking:
BlaiseEbuth: I am. Are you?
KiwiTae: FalINTOblivion0112 time to get tipsy
Wontonimo: i'm in
Wontonimo: :beer:
dbdr: RoboStac, what kind of hiding is this? ;)
RoboStac: the worst kind where it's just a completletly unfinished bot
RoboStac: but still slightly better than my old one
dbdr: sorry to hear
dbdr: what kind of search?
RoboStac: ga
dbdr: lack of time?
RoboStac: that and a failure to notice how long the contest was
dbdr: ouch
dbdr: enjoying rust at least? :)
RoboStac: yeah, I use it quite a bit outside of cg
dbdr: nice! for work?
AllYourTrees: :cheer:
RoboStac: nah, just for other stuff (parts of my nn backend is in rust) + been using it for codeforces / codejam etc.
AllYourTrees: since you have stuff in C++ also, whats your take on efficiency of rust vs c++ on CG? i've been wondering about that lately
RoboStac: I think it's pretty close but probably still a small advantage for c++
YurkovAS: i'm hear that not possible make big static array of structs without constructor in rust (10mln+)
AllYourTrees: yeah last time i tried to do anything with statics in rust it didn't end well
AllYourTrees: i think there's a crate to improve the use of statics, but i don't think CG has it
RoboStac: I tried to get my nn code working in the past but the lack of const generics meant it couldn't optimise as well so was about 10% behind my c++ one. Might be better with the newer versions but i've not tried
mihei: o7 comrades, can someone explain a bit, what I need to do to get "examiner" achievements?
Illedan: Hi
Illedan: I had hopes jolindien :P
Illedan: but gg it seems
jolindien: thanks
jolindien: I tried other things
jolindien: but comed back to previous
Illedan: I have tried so many things -.-
jolindien: good game! it is tied
Illedan: Yeah, very. Gg
jolindien: me too!
Illedan: What algo?
jolindien: I hope it's the last time I tried so many evaluation functions
Illedan: Did you have offline fitting of the evals?
jolindien: AG evaluated on two opponent solutions
jolindien: no
Illedan: AG?
Astrobytes: GA
jolindien: yeah GA
Illedan: k
jolindien: AG is french
Illedan: I have SGA
Illedan: :)
Astrobytes: Algorithme Genetique
jolindien: SGA ?
Astrobytes: génétique
Illedan: https://www.codingame.com/forum/t/sofia-labs-coding-challenge-feedback-strategies/190206/4
jolindien: speak french Astrobytes ?
Astrobytes: un peu
jolindien: interesting Illedan
Astrobytes: Oh same approach this time Ille?
Illedan: yeah
Illedan: but I had to tweak a lot because of having ball vs not having ball
Astrobytes: Yeah, that kinda changes a lot
Illedan: and only 36 angles :(
Astrobytes: Doesn't smitsi/UCT forest work on this?
Astrobytes: Thinking of BR
Illedan: Too high branching afaik
jolindien: I tried, I failed
Illedan: You need the details of the throws
Illedan: to get perfect Bludger bounces
Illedan: which is very strong
Astrobytes: Hm, yeah, it's a little more involved, very true
RoboStac: I've tried on fb in the past and even removing spells couldn't get a version that beat my ga
Illedan: Might work with beam search, but had no time to try
Illedan: Because most of the games are just optim
Illedan: I need more games vs Jolindien :D seems I counter you a lot
Astrobytes: Well, anyways, hope everyone enjoyed it. Not likely to see much more of these contests.
Illedan: Sadly
Astrobytes: Yea. It is what it is I guess :/
Illedan: Life goes on
Astrobytes: It most certainly does.
Illedan: Going to miss my main source of procastination
Astrobytes: There will always be community multis to waste time with! (well, I hope)
ViviCode: http://chat.codingame.com/pastebin/57956cd6-663e-4591-9a8a-9d32d4751903
ViviCode: Can anyone help me solve this mystery in Ghost in the Cell?
See link in previous message :)
ZarthaxX: ViviCode one did an invalid action and the other timeouted
Astrobytes: Your opponent played an invalid move and you timed out
ZarthaxX: Astrobytes ;)
Astrobytes: Zartho :D
Illedan: ViviCode, read the tooltip on the vier line + read the game summary, which tells the reason for the final score.
Illedan: *viewer
RoboStac: I'm guessing neither player handles the case where 1 player doesn't have any controlled factories
Astrobytes: Seems that way yeah
Bon[]Crayon: frame 80 'ViviCode : timeout!'
Illedan: Robo submitted 30 sec before? :P
ViviCode: ah, that's true I didn't account for that, because I figured the game would be over if my opponent doesn't have any factories. But they can still have cyborgs on the move, true.
Thanks!
Astrobytes: No Dapo with ~5 sec left
Illedan: gg
Astrobytes: I think the top 3 are pretty clear :D
Illedan: I'm surrounded by French :O
RoboStac: trying to fix a crash that I don't understand
RoboStac: should have stuck to languages where you can just do silly things and it'll continue on without crashing
Astrobytes: :D
Astrobytes: The joy of C++ :P
Astrobytes: "can't be wrong, it runs!"
dbdr: RoboStac what's the symptom?
RoboStac: getting a nan score from somewhere and then using that in a way that causes an exception
RoboStac: pretty sure it's a bad spell target as it wasn't happening before i tried to enable magic
dbdr: don't all float operations saturate to nan?
dbdr: there are no exceptions in rust, by the way, I guess you mean a panic?
RoboStac: yeah
RoboStac: I'm trying to sort a list of floats using partial_cmp (as floats aren't ordered) and I'm assuming nan is what breaks it
dbdr: ah yes
dbdr: you could always decide to order nan as you want. but the right thing to do would be to reproduce the bug and fix it ;)
RoboStac: yeah, fixing where the nan was coming from was my goal
Dohmanlechx: Can someone take a look on my C code for that Thor game? Solved it with C++ but having issues appending char list with chars in C
ViviCode: OMG, I just fixed my code with an else:
print('WAIT')'
and I went from ranking 256th to beating the boss! Thanks again :)
RoboStac: but with 30 seconds left just doing unwrap_of(f64::MIN) seemed more likely to work
RoboStac: *unwrap_or
dbdr: :D
dbdr: nice last minute hack!
Illedan: :ninja:
Illedan: I'll finally have some break now. Cya around people :)
dbdr: good rest!
ViviCode: cya illedan :)
RoboStac: might be enough to break back into top 10 which I'll be happy with
Dohmanlechx: https://pastebin.com/XBu5hg4U
dbdr: yeah, probably two rust in top 10 :)
ElmerZig: what game is most fun/interesting?
ViviCode: Clash of Code :) at least if it's not fun/interesting, it is over in 15 minutes anyway
RoboStac: Dohmanlechx - you can't use individual chars as strings in c like that as they aren't terminated
RoboStac: also you need direction to be size 3 so it has room for the terminating null at the end
Dohmanlechx: RoboStac - that makes sense. Solved it now with your hints, thanks
Dohmanlechx: http://chat.codingame.com/pastebin/d2ed13b4-7cbf-4d69-9850-36605cd8a9c8
therealbeef: is there any way to retrieve the code for the sogeti contest? forgot to save it :expressionless:
therealbeef: wanna reuse some of it for a FB bot
Westicles: hypnosis maybe?
reCurse: There's no report for semi-private contests so it's lost unless you convince [CG]Thibaud to look it up for you
Astrobytes: Who doesn't back their code up anyway? Shame on you therealbeef :P
therealbeef: got caught up in the excitement
Illedan: I think thibaud can give you the link, because there is a page. You just need the id
therealbeef: thanks, i will ask him. although 90% was trash code that needed a rewrite anyway
Illedan: Mine was enough for Legend :)
Astrobytes: He's on holiday for this week therealbeef so hang on til next week before asking him
easdasd: hello guys
easdasd: I love pyton
Wontonimo: that's great
Wontonimo: python loves you too easdasd
reCurse: With a cold embrace until it chokes you and swallows you whole
ViviCode: lol
Wontonimo: it's hot in my house. i welcome the cold feel.
Astrobytes: Technically anacondas ain't pythons though, so that would be unlikely. They're a type of boa iirc
Astrobytes: (wrt "swallows you whole")
Wontonimo: i thought we were talking about the cult of python programmers
Astrobytes: Both.
Wontonimo: ah
ViviCode: I just googled "death by Python" and got this:
(CNN) An Indiana woman found dead with an 8-foot python snake wrapped around her neck died from asphyxiation, officials said Friday. The snake was a reticulated python, which is native to southeast Asia and is considered to be the longest snake in the world.
reCurse: Um
reCurse: Pythons can certainly do that
reCurse: Though I'm no snakologist
Astrobytes: Swallow you whole?
Astrobytes: Highly unlikely. A very small child perhaps.
reCurse: Did you just assume his/her/its age?
Astrobytes: :rofl:
Astrobytes: The reticulated pythons will struggle with anything smaller than you're average adult
Astrobytes: *your
reCurse: Well python typically comes with an anaconda distribution
reCurse: So there
Astrobytes: There are always exceptions ofc
Astrobytes: I know, that's why I brought up anacondas ;)
Wontonimo: facinating
Wontonimo: *not
Astrobytes: Yes. Something you use to fill your house with very large snakes.
Astrobytes: Anyway, I feel like death with a pulse. See ya's later
Wontonimo: l8R
Wontonimo: rrr
Wontonimo: when do scores get recalculated?
Wontonimo: 1am in France?
Wontonimo: 11pm?
Wontonimo: anyone know?
Wontonimo: Hey ReyZis, are you doing Clash of Code?
Westicles: usually done by 2:15 am paris I think
Wontonimo: thanks Westicles ! I appreciate that
Wontonimo: oh, it's already 1:47 in paris
Westicles: Hmmm, interesting. If your contrib gets approved twice you only get the 50 XP once
Wontonimo: ;)
Westicles: As it should be...
Husoski: I didn't get extra for solving Mars Lander in Pascal, either... boo hoo
Wontonimo: were you expecting extra for pascal?
Husoski: Maybe for bash, I guess.
Wontonimo: i'd give you extra for pascal. i haven't seen pascal used in production since 1995
Husoski: I have a friend who was doing a bunch of consulting in Delphi until quite recently.
Husoski: I haven't used it professionally since '92, though.
Wontonimo: hey, has anyone told you the -3v trick for Coders Strikes Back?
Husoski: I don't think so...but I'm stupid in CSB, so I decided to set aside bots until I learn a bit more about AI.
Husoski: So, what's -3v?
Wontonimo: oh!
Wontonimo: let me make a short replay for you
Wontonimo: just a sec
Wontonimo: Here is a mars lander where it speeds to it's target, and the target is the landing pad
https://www.codingame.com/replay/574102787
Wontonimo: in this example, instead of aiming at the landing pad, it aims at the landing pad MINUS 25 times its velocity https://www.codingame.com/replay/574102797
Wontonimo: The first video was supposed to be this https://www.codingame.com/replay/574102787
Wontonimo: oh.. it was right
Wontonimo: cool
Wontonimo: so, in CSB, you can use this to correct your steering. Instead of aiming for the checkpoint, aim for it minus 3 x your velocity (not thrust)
Wontonimo: The concept comes from PID controllers, and here is the best and easiest to read paper on it https://w3.cs.jmu.edu/spragunr/CS354_F17/handouts/pid.pdf
Husoski: Nice! I wonder where the -3 comes from?
Husoski: Oh...the paper. Yes, I'll bookmark that!
Husoski: Oh rats, I gotta fiddle with it. My old Adobe Acrobat doesn't understand the format. Thanks, though, and I'll get it untangled soon.
dsfjkfdls: I forgor :skull: how to code
dsfjkfdls: I rember :grinning:
jraewilson: omg same:thumbsup_tone2:
Wontonimo: hey Husoski , the -3v means to target checkpoint - 3 * velocity
Wontonimo: the 3 comes from experimentation
Wontonimo: try -1* , -2* and so on
Wontonimo: it's different for every scenario. For mars lander, the factor that works best for me is -25v
xxd: How do I go back to a closed CoC chat
Wontonimo: remember the chat room
xxd: 0/10
Wontonimo: and then put a "#" in front of it and type it here like
- chat12423
Wontonimo: #coc3532
xxd: yeah but like
xxd: i have the clash report
xxd: and i want to access its corresponding chatroom
xxd: like im on the results page
Wontonimo: like you have to remember the name of the chat room
xxd: thats so dumb
Wontonimo: like dude, yeah
Husoski: @Wontonimo I guesssed that when the coefficient for Mars 2 was -25. Thanks a ton for the tip and the pdf. (SumatraPDF reads it just fine.)
Husoski: (stuck scroll buffer)
Wontonimo: hey, no prob
Wontonimo: another cool trick that you'll probably find really straight forward is a simple drift detection like so:
Wontonimo: calculate if at the current velocity your pod would be within 400 units of the current checkpoint within the next 5 rounds, and if so,
Wontonimo: don't target the current checkpoint, target the one after that
Wontonimo: and you'll drift through the checkpoint and look like a pro
Wontonimo: the -3v and the drift will get you to mid gold
Wontonimo: you'll get a list of checkpoints when in bronze or silver
Husoski: I tried drifting like that and made it to a second Wood league. That's about when I ran out of original ideas.
Husoski: (I just made a list of the checkpoints as they came up and noticed when they repeated.)
Wontonimo: nice
Wontonimo: total overkill for wood. crank speed up to 100, implement -3v, and you'll rocket up to silver
Wontonimo: that's it, nothing more or less
Husoski: I think I'll try the ideas on Mars 2 first, but yeah...I see the possibilities
Wontonimo: so, i lied just a little about mars 2.
Husoski: Tsk.
Wontonimo: which is to say, I have 2 levels of sideways thrust
Wontonimo: angle 15 w/ power 4 if within 200 of the adjusted x target
Wontonimo: angle 25ish w/ power 4 if outside 800 of the adjusted x target
Wontonimo: so, on top of the -25v, there is a different thrust.
Wontonimo: that's part of PID though, to have a thrust that is related to the distance
Wontonimo: the -v part is to adjust the distance target
Wontonimo: i don't have anything in either mars lander or CSB that integrates error. that part of PID I've just skipped
Wontonimo: nood gight every1
Husoski: I'll be starting simple and adding stuff as I understand it. I tend to take a lot of this offline so I can crash at a faster pace. :) No graphics, but that's not too awful for someone who started in the teletype and punched card era.
Husoski: Ditto!
Wontonimo: oh! very nice! punch card all the way! I started in the vic20 era
Husoski: I have a VIC20 story for later, then... type at you later!
ProCoder03: Hi for drift detection , see if our current velox and veloy will reach the checkpointX and CheckpointY and make thrust to 0 and aim for next checkpoint .Am i right. because I tried implementing at Gold and it doesn't seem to work for me ??
Michael_48: cool on the output log for only representing printing null if prints follow it. totally doesn't lead to 10 minutes of trying to find an invisible bug...