Chat:World/2022-02-06

From CG community
Revision as of 05:29, 7 February 2022 by Chat Log (talk | contribs) (Created page with "File:Default_avatar.png Potatobol: hi File:Default_avatar.png Potatobol: i dont know how to code <img src=/a/77753562892354> MrZee: >Makeprogramdothing; File:Defa...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Default avatar.png Potatobol: hi

Default avatar.png Potatobol: i dont know how to code

MrZee: >Makeprogramdothing;

Default avatar.png Kerr.: is there a way to turn off automatically making parens and brackets in doubles

Blokops: yes

Blokops: a answer a bit too late :/

Blokops: go to config in the left side

Blokops: editor configuarition

Blokops: and put false on the autoamtic bracket creaition

Blokops: Kerr.

Default avatar.png Kerr.: tyvm Blokops

Tien2k9: why i can't use auto-complete in codingame?

**Tien2k9 is wondering

**Tien2k9 hmm

Default avatar.png Kerr.: autocomplete, like autocomplete statements?

Default avatar.png Kerr.: I'm only early in, but I'm a little annoyed that every challenge involves string inputs/outputs

Default avatar.png Kerr.: are there any of the games/challenges that have a mini-api?

Tien2k9: nah

Blokops: Tien2k9 vscode has a timer for the autocomplete

Blokops: it went up with an update

Blokops: same way as above you can short it but you need to add the command to do so and i dont know what it is

SSisthebest: yo hi dudes

Dark_Hack: hello

Mortis_666: hi

Default avatar.png BeardedKitten_5f8d: Hey folks, I just started on this and Thor has achieved his goal, on all four TestCases, yet i still have 0%, and i have already solved this in two different ways. is it looking for a specific answer?

eulerscheZahl: no specific answer expected, feel free to share your code. just paste it here

Default avatar.png BeardedKitten_5f8d: http://chat.codingame.com/pastebin/202a7ad7-8718-41e4-bbe6-ecd9f6d0db50

eulerscheZahl: if(lightX == 31 && lightY == 4 && initialTx == 5 && initialTy == 4 )

eulerscheZahl: that's hardcoding

leojean890: euler

eulerscheZahl: you are supposed to write something like if (initialTx < lightX) Console.WriteLine("E")

eulerscheZahl: leojean890

leojean890: read that !

leojean890: http://chat.codingame.com/pastebin/4b412d16-6ac8-4cea-8d9d-a38b099f2fa8

leojean890: I've just approve for those reasons :)

eulerscheZahl: ok, hardcoding time...

eulerscheZahl: still lacks some precision in the statement (how do distances work?)

leojean890: but I understand your point of view^^ but I think that my reasons are a good reason as well to accept it :p

eulerscheZahl: It must be very difficult to approach the optimal score. https://www.codingame.com/playgrounds/40701/help-center/contribution-guidelines

eulerscheZahl: using a website to paste a few lines took me less than 5 minutes

leojean890: ah you're right ! we can compare our simulation with what th referee sends back as a distance

leojean890: to verify euclidian for example

eulerscheZahl: it is euclidian. but written nowhere

jacek: should be eulerian

leojean890: we should update the statement yes

leojean890: the contributor forgot about that

eulerscheZahl: and you didn't read my whole comment

leojean890: do you have news about RAIC ?:P

eulerscheZahl: or another one further down which also mentioned it

eulerscheZahl: https://www.pochta.ru/tracking#RR148020554RU

eulerscheZahl: that's my news from RAIC

eulerscheZahl: and https://www.pochta.ru/tracking#RR148020727RU

Default avatar.png orchidperson: do programmers learn more than 2 languages

eulerscheZahl: stupid German customs. sending my tshirt + hoodie back to russia without even informing me

Default avatar.png orchidperson: lol

leojean890: ah nice for the gifts, I should find a way to see where mine is :P

leojean890: about RAIC 2021/2022 ?

eulerscheZahl: http://chat.codingame.com/pastebin/7224e31c-3f43-4f89-be13-d4aaf2fb621a

leojean890: I will re-read your comment now to see if I didn't see something

eulerscheZahl: don't bother, it's approved already

jacek: im 1st :tada:

eulerscheZahl: somehow that black background depresses me

eulerscheZahl: uhm, did the creator screw up the leaderboard sorting?

eulerscheZahl: high=good

jacek: oh my

leojean890: but about your comment, I prefer an optim against a puzzle, because it's motivating to improve more and more :P

eulerscheZahl: well, that's a variation of the problem. but easy enough to transform so that concorde can still do the job

leojean890: jacek 2 236 125

leojean890: leojean890 532 136


jacek: i just shuffle indexes

leojean890: I spent more than 20 hours on most optims

leojean890: and less on puzzles

eulerscheZahl: there is a line in the config that tells CG whether you want to optimize for low or high scores. but it will wipe out the current leaderboard and not just resort it

jacek: its still time

leojean890: 2219406 using print(range(n))

leojean890: 532 136 is a easy SA

leojean890: an*

leojean890: with 1 mutation

leojean890: only

leojean890: hill climbing*

leojean890: eulerscheZahl in the end of your comment you meant that the computation is not accurate ? 17112.00 vs 17112.8 ?

leojean890: I don't find how to revert the leaderboard while editing a published contrib

Default avatar.png BeardedKitten_5f8d: @eulerscheZahl Thank you

eulerscheZahl: config.ini has to be changed: asc <=> desc

eulerscheZahl: but i see it as the creator's task to do so. someone just has to tell him

eulerscheZahl: and yes, that small rounding error. probably not relevant

jacek: approvers these days...

jacek: float vs double?

MrZee: we all float down here

eulerscheZahl: i was talking about the precision of concorde itself

eulerscheZahl: but maybe it's accurate and just prints a rounded value, no idea

leojean890: zLuki the creator, but offline (may contact him with forum)

leojean890: leaderbord sorting was not visible while testing the contrib

leojean890: for multis people who try are ranked

leojean890: in a leaderboard dedicated to the contrib

leojean890: not for optims idk why

leojean890: it would be a way to debug that before approcing

leojean890: approving

eulerscheZahl: i'm last

leojean890: maybe it would be better if the contributor used more validators, as well

leojean890: 5 is a little short

leojean890: 201k I see:)

leojean890: will try to do better :P

eulerscheZahl: there is a limit of 30 (testcases+validators)

leojean890: but there are only 5 + 5 there ^^

eulerscheZahl: no, 5 tests + 15 validators

leojean890: oh yes that's right

leojean890: I saw that

leojean890: forgot then

eulerscheZahl: 17k-ish is a typical score for 300 nodes

eulerscheZahl: between 17 and 18 actually, not below 17

leojean890: because they never use huge coordinates ?:P

leojean890: type (1000, 3000), (80000, 90000) => no point of doing that :p

leojean890: I'm not really experimented on this problem, but motivated ; the concept is easy enough to try many things easily

eulerscheZahl: but not suited for an optim IMO

eulerscheZahl: because it's too well studied

leojean890: didn't try to findi a Hamilton path

leojean890: it's nice for an optim, but not interesting for too advanced people I guess

leojean890: for example it would be bad for a contest

eulerscheZahl: the hamilton path was about http://www.hacker.org/oneofus/

eulerscheZahl: you'll need flash to see it

leojean890: but optims/multis are kind of a zoo

leojean890: that are here to train, improve

leojean890: some people just push existing codes

eulerscheZahl: i don't know another optim where the proven optimal solution is that easy to obtain

leojean890: ah my adobe flash is not compatible^^

eulerscheZahl: basically you have a grid of different objects (shape, color)

eulerscheZahl: and you have to find a path where you stay in the same row or column and 2 continuous objects have the same color or shape

eulerscheZahl: i converted that into a TSP problem and used concorde to solve

eulerscheZahl: generating the full adjacency matrix took longer than the actual solving process (100MB files pre-SSD)

eulerscheZahl: darkhorse64 is pretty close to the optimum already

leojean890: shape/color is the theorical definition ? but here it's coordinates instead in the implementation ?

darkhorse64: I have had more than a week to work on it. 30 + submits already on the demo

eulerscheZahl: talking about hacker.org - but let's forget the Hamilton path

leojean890: I have to study that on my side :)

eulerscheZahl: i piped it into concorde, took me 10min to do it manually on all testcases

leojean890: I did 500k with this basic hill climbing

leojean890: but have many things to try

leojean890: I didn't know concorde before today :p

darkhorse64: eulerscheZahl: your solution is "concorde" hard coded ?

eulerscheZahl: yes

eulerscheZahl: proven optimum

eulerscheZahl: i will be gone from the leaderboard when sorting is reversed

eulerscheZahl: proving a point here

leojean890: Concorde and Linkern enable to compute adjacency matrixes I read?

eulerscheZahl: https://neos-server.org/neos/solvers/co:concorde/TSP.html

leojean890: ok so 201k is the best

eulerscheZahl: just upload a file like:

eulerscheZahl: http://chat.codingame.com/pastebin/01e80a9d-4273-491a-97d0-d3f7f874886e

leojean890: thanks eulerscheZahl

eulerscheZahl: it prints out the optimal tour

leojean890: I see

leojean890: so many people can cheat if they want :P

darkhorse64: No worries, eulerscheZahl. I approved it because I was interested in the problem but it looks clear it may be a bad optim because hardcoding is too easy

leojean890: like hypersonic with the rgeat contest code shared :P

leojean890: great

leojean890: darkhorse64 same I wrote that in comment

eulerscheZahl: that's a different issue IMO

eulerscheZahl: one is blatant cheating

leojean890: it's interesting for people who are not expert (like me^^)

leojean890: to improve

darkhorse64: ^

eulerscheZahl: using code specifically taylored to a CG problem

eulerscheZahl: the other is using commonly available knowledge

leojean890: not for contest where people would cheat

eulerscheZahl: and you can still write an awesome bot and beat the published HS one

eulerscheZahl: you can't beat optimal

leojean890: yes^^

leojean890: hopefully people will not cheat too much

eulerscheZahl: not even sure if i should consider that cheating

leojean890: but it's not a big problem, the most important is the practice opportunity I think in multis and optims

Ftyghi: hi

leojean890: I think that you are able to code a great optim code without cheating ;p

Ftyghi: hi

eulerscheZahl: great players don't cheat to get great results. they cheat to get them faster

leojean890: yes they can :P

Ftyghi: hi can anyone see my message

eulerscheZahl: no Ftyghi

leojean890: so, in contest we need original new topics

Ftyghi: :confused:

leojean890: and in multis and optim we add practice topics

leojean890: to motivate people

leojean890: your creation of OOC was great for example (existing board game but to transform it into an IA contest)

eulerscheZahl: i still think that a puzzle would be a better place for an optim with a known optimal solution

eulerscheZahl: i didn't do much on OoC actually

eulerscheZahl: i saw the game 2 days before everyone else and helped to polish it up

leojean890: but people will be more motivated to improve and try things if there is a leaderboard and a score comparison

leojean890: and a puzzle, if you pass it you stop improving :(

eulerscheZahl: let's agree that we can't agree then ;)

leojean890: ok about OOC^^ it was with Illedan and G-ROM from CG ?

eulerscheZahl: mostly G-Rom

eulerscheZahl: Illedan started with me

leojean890: yes but you have a better level than me so you don't have the same concerns than me xD

leojean890: your aim is to have new topics

leojean890: only

eulerscheZahl: i wanted to seriously dive into it on the weekend. and then Thibaud was like "can we start the contest on Friday?"

leojean890: I see, too bad you can't polish optimally

eulerscheZahl: there were some minor problems with OoC at first, but in general it worked better than I expected

darkhorse64: it lacks a destroying mine feature, imho

eulerscheZahl: mines drowning after X turns

darkhorse64: but it's a great FoW game

leojean890: yeah I liked it

leojean890: destroying mines would me nice and make the game harder

leojean890: would be*

leojean890: and mor complex/interesting

leojean890: more

eulerscheZahl: i really wanted to look into bomberland today. now i'm still chatting about TSP

eulerscheZahl: see you later

darkhorse64: I was thinking more about mine exploding mines or torpedo destroying mines. So there would be a tradeoff between increasing your detectatibility(?) and finding a safe path

eulerscheZahl: looking at the starter codes and what it takes to port it to C#, I already lost all my motivation again :(

eulerscheZahl: OoC is a close port to Captain Sonar. I don't know the rules of that game

leojean890: would be a good idea darkhorse64

leojean890: I don't know bomberland

eulerscheZahl: oh, konstant found a way to score even lower than me on TSP :D

eulerscheZahl: https://www.gocoder.one/docs/getting-started

eulerscheZahl: https://www.twitch.tv/videos/1280948902 game in action

eulerscheZahl: competition is really weak

darkhorse64: There are actually a few significant differences: the sonar for instance

eulerscheZahl: giving 2 values (e.g. x and y)

eulerscheZahl: where one is the truth and the other is a lie

eulerscheZahl: but would complicate the protocol

eulerscheZahl: bomberland communicates via websockets. stdin/stdout would also be possible and lower the bar to write your own bot without sticking to their language options :(

darkhorse64: I do not miss this feature (sonar). Time for new multis. Struct is making an Ataxx.

leojean890: I'll look into bomberland

leojean890: konstant has 6% for the moment, didn't see that xD

leojean890: https://rgcddev.itch.io/bomberland ?

eulerscheZahl: https://www.gocoder.one/docs/getting-started

eulerscheZahl: and the twitch stream above to get a better idea what it's about

eulerscheZahl: their platform only plays games once a week

leojean890: thx

darkhorse64: CG benchmark would take ages ...

Dark_Hack: hello eulerscheZahl

Dark_Hack: my old friend

leojean890: CGB dies not work anymore because of the limitation of ide plays

darkhorse64: just kidding

leojean890: it's too bad their limiattion :(

eulerscheZahl: CG Benchmark? you can't start games yourself (except for offline testing)

eulerscheZahl: you submit and wait a week to get the results

eulerscheZahl: (on the bomberland website)

leojean890: oh ok but on CG it makes plays

leojean890: and woth the limitation it does not work

leojean890: good

leojean890: after 150 plays it's going slow

leojean890: on bomberland idk I have to test first

Illedan: Good morning

Illedan: So I guess the goal in TS is to get as long as possible path

eulerscheZahl: i see you joined the TSP too

Illedan: I wonder if my score gets updated when I make a longer path

Illedan: :D

eulerscheZahl: sure

Illedan: How does Jacek even get a longer path

eulerscheZahl: concorde allows a matrix for distances. so i could do 1000-actual dist for each pair and generate the longest path

eulerscheZahl: what are you doing? 0 1 2 3 4 0

Illedan: Yeah..

eulerscheZahl: and Zorg is greedy I suppose, as that score appears twice too

Zorg1: yes i'm greedy :p

eulerscheZahl: btw i won't resubmit my optimal solution when the leaderboard sorting gets fixed. So i should just be gone completely then.

eulerscheZahl: just pointing out a serious flaw in puzzle design. you need an extra twist to make it suitable

Illedan: Yeah

eulerscheZahl: e.g. combine it with Search Race: visit all points in an arbitrary order by driving with the car

eulerscheZahl: shortest TSP path may have some unpleasant corners

Default avatar.png minhminhbi: it's so hard

Zorg1: CSB meets salesman ?

eulerscheZahl: jacek your turn

Default avatar.png minhminhbi: i'm newbie

Default avatar.png minhminhbi: of pyhton

eulerscheZahl: salesman strikes back

eulerscheZahl: *salesperson prof told me about a colleague who really had to rename that chapter

eulerscheZahl: publisher didn't like salesman

Illedan: WTF

Illedan: Why do I need to start and finish at 0

Illedan: -.-

Illedan: That was 5 min I'm never getting back

Illedan: Let's go inverse TSP

Illedan: eulerscheZahl first :O

eulerscheZahl: https://i.imgur.com/SXa0eNU.png

jacek: hmm.. is it travelling or traveling

Mrs.GloriaZindlebocker: what ever happened to the snake optim?

Illedan: I lost motivation

eulerscheZahl: what is missing? just the JS animation?

Illedan: Yeah

eulerscheZahl: should be doable and i like the snake idea for a game

eulerscheZahl: is your code public?

Illedan: Yeah

antiwonto: [automated] Illedan please don't repeat yourself

Illedan: ...

Illedan: And it is missing random board sizes

Illedan: I'm FIRST!

Illedan: :D

eulerscheZahl: distracting me...

eulerscheZahl: first by 600 points

Illedan: ezpz

darkhorse64: I start to believe that I missed the point of this optim

Illedan: The sort is wrong

Illedan: So we are just having fun

Illedan: Game is too easy anyway

darkhorse64: I know. Just kidding. Not so easy when you practice diy

darkhorse64: "do it yourself"

eulerscheZahl: take that, the toad is back

Illedan: Take what?

eulerscheZahl: my lead

Illedan: Noooo

ASM_MOV_INT: that toad.... is froggin!

darkhorse64: It's me :stuck_out_tongue_winking_eye:

Illedan: Ahhh, darkhorse64 cheating

darkhorse64: no, it's legit

Illedan: Gotta bring out the big gunns

Illedan: hmm

C26_1: Hello everyone

C26_1: I have a new sequence of numbers

C26_1: I have calculated up to 14 terms and this is the sequence: 1 5 2 3 5 6 8 5 1 1 0 3 0 8

eulerscheZahl: OEIS doesn't know your sequence

C26_1: The pattern for the sequence: - Look for decimal expansion of its inversion (n = 1 at first term) - Find the largest number that has the most concurrent in that inversion

C26_1: For example:

C26_1: n = 5 1/5 = 0.2 Since 0 and 2 appear one times, then we output the largest number, that's 2

C26_1: Wait

C26_1: Wait for me

C26_1: 1 5 3 5 2 6 8 5 1 1 0 3 0 8

This is the true sequence

C26_1: I'm sorry if someone has confused

C26_1: http://chat.codingame.com/pastebin/a0737ab3-f097-4356-9b2c-9f47106d2365

C26_1: I left for a while, forgive for me if this is a mess of text

Nachosauce: It's no 3.2million, but I'll take my 2.78m score on the travelling salesman :joy:

5DN1L: C26_1 If you need to post a large block of text here, please use the pastebin facility here, don't flood the chat please, thanks.

C26_1: ah ok

Default avatar.png Zbouig: hello

Default avatar.png Zbouig: im zbouig

Default avatar.png Zbouig: i have 15year old

RajdeepBiswas: helllo

RajdeepBiswas: i am a 15 year old

RajdeepBiswas: i am ducky

Default avatar.png Zbouig: you too like kannon?

RajdeepBiswas: oh i love it

Default avatar.png Zbouig: me im froogie

Default avatar.png Zbouig: its good vibe

Default avatar.png UnnamedCodinGamer: how is the ranking score in the newly approved tsp optimization game calculated?

Default avatar.png UnnamedCodinGamer: I thought that the distance must be minimized

Illedan: The ordering is wrong, but we are having fun trying to maximize the path

Default avatar.png UnnamedCodinGamer: I see

Default avatar.png UnnamedCodinGamer: thanks

Default avatar.png UnnamedCodinGamer: are the validators public for this problem

Default avatar.png UnnamedCodinGamer: ?

Illedan: Yeah

Illedan: In the contribution

Default avatar.png UnnamedCodinGamer: I can see them, can others as well?

Illedan: Yes

Default avatar.png UnnamedCodinGamer: nice

wlesavo: oh tsp is up nice

Default avatar.png UnnamedCodinGamer: actually I am seeing a link to the source code only on the "view contribution" page

Default avatar.png UnnamedCodinGamer: which is not available to all

struct: you need level 29+ yes

jacek: :(

Default avatar.png UnnamedCodinGamer: I could not be bothered to use the provided time if the validators are public

Default avatar.png UnnamedCodinGamer: it is kind of uneven playing ground though

Default avatar.png UnnamedCodinGamer: https://www.codingame.com/contribute/view/7995ef6dfd1604a148a65cea9ee95de1e1da

eulerscheZahl: there are tons of solvers publicly available

Default avatar.png UnnamedCodinGamer: I wrote my own

eulerscheZahl: an optimal one?

Default avatar.png UnnamedCodinGamer: I am not sure

Default avatar.png UnnamedCodinGamer: SA

Default avatar.png UnnamedCodinGamer: on the 5-th test case I got around 17400

eulerscheZahl: https://neos-server.org/neos/jobs/11590000/11599531.html 17112 (or 17113, rounding issue) with an online solver

eulerscheZahl: all you have to to is to copy the stdin as it is to their website, same input format

Default avatar.png UnnamedCodinGamer: is this sure to be the global minimum?

eulerscheZahl: yes, proven optimum

Default avatar.png UnnamedCodinGamer: thanks

eulerscheZahl: i wrote that in the contribution comments and rejected

Default avatar.png UnnamedCodinGamer: I will see if I can get it

Default avatar.png UnnamedCodinGamer: yes, using this site makes the whole exercise pointless

leojean890: euler 3 241 896, me 3 240 630 mdrr

leojean890: darkhorse 3 248 344

leojean890: heurstic + HC powaa

wlesavo: damn already max submits

leojean890: hehe

leojean890: when will the creator invert the leaderboard hmm

eulerscheZahl: it's more fun this way acutally

leojean890: so that you invert your heuristic functions

leojean890: it's fun as well :P

jacek: blame the approvers

wlesavo: even if he will the lladder is needed to be reset manually

leojean890: will try to modify my HC to SA

struct: darkhorse64 I think I understand the ataxx notation now for blocks

leojean890: I tried to do this instead of him :p but not psosible in the edit part

leojean890: why does not CG enable the leaderboard for optims in contrib part ....

leojean890: and they do for bots prog

leojean890: because we would have seen the problem before

leojean890: should add that to the checkist of TODO stuff

leojean890: and when an optim is in contrib part I already like to compare to know if my code is good or not:p

leojean890: sample : snake (still in contrib)^^

struct: was snake finished?

darkhorse64: struct: will you encode only block config, notation and generate the blockers from it ? Not easy but saves mucho typing tables

struct: probably yeah

struct: since there seem to be over 200 board setups, I dont feel like typing them all by hand

jacek: no books?

darkhorse64: 200

struct: no jacek

struct: 596 boards

leojean890: snake is not finished but I can't have the ranking list

leojean890: it would be better for both reasons

leojean890: (testing the leaderboard and testing the contrib while comparing)

struct: random should be terrible from what im seeing, wont have to worry much for the boss

Wontonimo: what are we talking about? is this a multi contrib?

struct: yes, ataxx

struct: some arcade machine game I think

Wontonimo: i can't find it. link?

struct: https://www.youtube.com/watch?v=rYIyB6Pw8VE

struct: https://www.pressibus.org/ataxx/indexgb.html

Default avatar.png pascal_beast: kk

Wontonimo: thanks. neat game

Manchi_o6o7: Hey guys. Has anyone maybe a resume.io subscription

eulerscheZahl: that inverse TSP seems much harder, at least for concorde. no heuristics for an initial solution i suppose

eulerscheZahl: running for minutes on 100 nodes already

Astrobytes: eulerscheZahl: we were speaking with dapo about this TSP last night, would adding a set of sub-nodes to each node be an improvement

Astrobytes: ?

Astrobytes: (e.g you must visit at least 1 or n subnodes from each node before moving on)

eulerscheZahl: i think this could indeed help against well known strategies

eulerscheZahl: but it's approved already

eulerscheZahl: a certain someone read half of my comment, then approved and rubbed it into my face

Astrobytes: and not every sub-node needs to be directly connected to another main node

struct: with his current it will most likely be gonne soon

eulerscheZahl: i'm already triggered again, maybe my CG break wasn't long enough

Astrobytes: just seeing it now

eulerscheZahl: i suggest a rewording of your idea to see if we mean the same: you have nodes and groups of nodes

eulerscheZahl: you have to visit each group a given number of times, then move on to the next group

Astrobytes: Yes, so instead of just one node as it is now, each one has it's own group, or neighbourhood I suppose

eulerscheZahl: and i admit i'm one of the 1-star reviewers

eulerscheZahl: willing to change that when the problem gets modified

Astrobytes: I should have moved my approval to reject last night as I wanted to speak to you about it

jacek: :unamused:

eulerscheZahl: hm, just thinking out loud you should not have to visit all nodes in the end. otherwise you could just adjust the edge matrix

eulerscheZahl: and btw i checked the chat logs this morning and saw daporan. that's why i even found the contrib

Astrobytes: good

eulerscheZahl: http://www.hacker.org/oneofus/ that's where i know concorde from

eulerscheZahl: solved it in 2017, how time flies

Astrobytes: I got my head in a pickle doing a google search crash course in parts of graph theory I'm vague about lol

Astrobytes: Sad loss that site

eulerscheZahl: i've learnt some theory about it at university. but mostly heuristics like Christofides

eulerscheZahl: and GA, ants and such

eulerscheZahl: https://www.sciencedirect.com/science/article/pii/S1002007108002736#:~:text=Ant%20colony%20optimization%20for%20TSP,nest%20without%20using%20visual%20cues.&text=So%20the%20ants%20randomly%20choose%20their%20path%20with%20equal%20probability.

eulerscheZahl: not recommending that one in particular, the first that came up on google

Astrobytes: yeah I know the ant colony optim method

Astrobytes: Going back to the CG TSP: what is better, finding the shortest route at each node neighbourhood or just visiting n nodes in each? And should some of those neighbourhood nodes connect to other main nodes?

DomiKo: Am I blind? Or maybe this puzzle never tells you for what you are getting points? (I know answer is obvious but still)

jacek: you get participation medal

eulerscheZahl: it doesn't tell

eulerscheZahl: i also wrote that in the moderation section before it got approved :smirk:

jacek: blame the approvers

eulerscheZahl: i do

DomiKo: :rage:

Astrobytes: quiet jacek, you were present at the discussion last night

eulerscheZahl: and did he approve?

jacek: :innocent:

eulerscheZahl: but back to your suggestion. let me read and try to understand your idea

Astrobytes: no, but if he had anything to add he could have said something

jacek: i didnt. i only made comment that in debug mode points could have hover information

Astrobytes: I was floating these ideas last night but there was nobody around with sufficient knowledge to answer me

kovi: oh, new optim

kovi: i meant a real one (as code keeper became a multi)

DomiKo: I like how the word 'shortest' occurs only in the Example section.

eulerscheZahl: and nowhere does it say "Euclidean"

DomiKo: And Rules section is written like it is some Clash.

eulerscheZahl: I already found the shortest for all 15 validators: 201,389

eulerscheZahl: but we are looking for the longest now

Astrobytes: eulerscheZahl: clearer explanation - each main node is a country, there is a neighbourhood of cities in each.

eulerscheZahl: ok, that sentence is clear

Astrobytes: Do you find the shortest route around each country before visiting the next?

eulerscheZahl: that would add a small variation but not sure if big enough

Astrobytes: And are any of these cities connected to another country? (e.g a freeway/autobahn or back road)

eulerscheZahl: i would suggest: you only have to visit 5 cities of a country (arbitrary number, can be changed per country)

eulerscheZahl: so you don't visit all nodes

eulerscheZahl: then i see absolutely no way to just throw standard solvers at it

Astrobytes: OK. And what about connections between cities, fully connected?

eulerscheZahl: yes, why not

Astrobytes: just wondering aloud

Astrobytes: And input size?

eulerscheZahl: 300 might even be ok then but 500-1000 won't hurt

eulerscheZahl: and i see a lot of potential on the visualization with the countries+cities

DethHarmonik: Oy there. What's the drama about? :/

Astrobytes: ok, now how to go about getting it changed

eulerscheZahl: forum post explaining the problem and suggesting solutions?

eulerscheZahl: or heavy downvoting to get it garbage-collected. but that would be the nuclear option

Astrobytes: Forum first, yes

eulerscheZahl: DomiKo was there first

DomiKo: :sunglasses:

eulerscheZahl: hey DethHarmonik. about https://www.codingame.com/contribute/view/7995ef6dfd1604a148a65cea9ee95de1e1da

eulerscheZahl: approved and easy to get the perfect score with an online solver

eulerscheZahl: that is if you ask for the shortest route. currently it's about the longest. still possible with concorde but some heurisitcs can't be applied

DethHarmonik: Hm. Just reading the title and the tags I see like a problem...

eulerscheZahl: read my comment. took me 10min to solve and hardcode all validators.

struct: it took me less to open and close it

DethHarmonik: Statement seems strange too... :/

DomiKo: you are right

DethHarmonik: Oh you posted about it DomiKo ^^ Sorry I arrive after the battle :p

Astrobytes: OK, how's that post?

AMMARBelbaz: .

Astrobytes: edited to mention prevention of bruteforced solutions/online solvers

eulerscheZahl: now i'll break my 1 month forum inactivity streak to <3 you

Astrobytes: heh heh, thanks

eulerscheZahl: n=3 cities (shortest route) before moving to the next city.

eulerscheZahl: to the next country

eulerscheZahl: and n is reserved for other purposes already

Astrobytes: both fixed

DomiKo: that's a nice post Astro

eulerscheZahl: java_coffee_cup that optimization for the longest path is probably a bug. you don't see it while it's pending

Astrobytes: just repeating what we just discussed here DomiKo, albeit a little more coherent :D

DethHarmonik: How was the Quora thing btw eulerscheZahl?

eulerscheZahl: huge failure from my side

eulerscheZahl: and tourist won, surprise

java_coffee_cup: indeed didn't notice it out while pending

DethHarmonik: That's the difference between a person who does it to eat and one who does it for fun I suppose. ;)

eulerscheZahl: one would have to check the config.ini i always had to think twice myself when I created an optim

Astrobytes: java_coffee_cup: what do you make of our proposed solution?

eulerscheZahl: have you seen that competitive programming hall of fame project?

Astrobytes: a shrine to tourist?!

eulerscheZahl: tourist won about $350k of prize money just in onsite-finals

Astrobytes: yeah I've seen that

eulerscheZahl: the $3k from quora are online, not even listed

java_coffee_cup: Interesting suggestion to make sub-nodes.

eulerscheZahl: whatever it takes to enforce re-thinking

java_coffee_cup: The current setting is the very standard TSP which solution can be found easily online

Astrobytes: It was my first thought after dapora n mentioned the issue last night. Without over-complicating the whole thing too much

eulerscheZahl: and i can already see the map with country borders and cities in my mind

DomiKo: tourist is winning almost everything he participated in since 7 years old so for sure he made many millions

Astrobytes: Indeed.

eulerscheZahl: one in the US with state borders

eulerscheZahl: Europe with all that tiny countries

eulerscheZahl: not sure if we speak about millions yet. but very impressive

Astrobytes: hm, there are only 295 countries

Astrobytes: *195

eulerscheZahl: but lots of cities per country

Astrobytes: (close only out by 100)

Astrobytes: true

eulerscheZahl: you don't even have to enforce the same about of cities visited per country

Astrobytes: regarding shortest path?

eulerscheZahl: if you are really dedicated, you can even download a list of cities to make the level design more realistic

Astrobytes: or number?

eulerscheZahl: let's say you have 3 countries with 4,7,10 cities in it. and the expected solution wants 2,6,4 visited cities of those

Astrobytes: list of cities wouldn't be so difficult to do, you could even use real distances

Astrobytes: Yes, go on

eulerscheZahl: net challenge: get google maps into the CG viewer :D

Astrobytes: lol

eulerscheZahl: next*

Astrobytes: oh I get you, not the same k visits for each country

eulerscheZahl: yes

Astrobytes: Post edited.

eulerscheZahl: did you ever mention the main issue by name?

eulerscheZahl: not everyone has access to the discussion after approval

Astrobytes: well I linked the contribution

Astrobytes: hm, true, Let me edit

eulerscheZahl: i like how you still have the 5% behind your name

Astrobytes: :grin:

Astrobytes: OK, post edited.

eulerscheZahl: :+1:

Default avatar.png dmuraco: how to debug in python?

Astrobytes: check the starter code

Astrobytes: it's written in a comment

eulerscheZahl: you can't set breakpoints and step through line by line

Astrobytes: unless you run an offline version of your code of course

Default avatar.png dmuraco: sys.stderr.write()

Wontonimo: print("message", file=sys.stderr, flush=True)

Scarfield: http://chat.codingame.com/pastebin/c59888b1-e6ee-4676-b1ee-0741683f02eb

Wontonimo: what's that from ?

DomiKo: https://www.codingame.com/multiplayer/optimization/travelling-salesman

DomiKo: Scarfield my reaction was similiar xd

Scarfield: yea its so bad its funny tbh

eulerscheZahl: there's a long forum post about the game already

Scarfield: yea i saw you chatting about it, checked it out, and the rules where just perfect, had to share :)

eulerscheZahl: usually i scan through those generic parts without reading all of it. only noticed when i was looking for a definition of the distance

eulerscheZahl: a real gem of comedy

Scarfield: its really a reverse optim

jacek: there is other graph based game if youre interested *wink* *wink*

eulerscheZahl: a pessim

eulerscheZahl: paper soccer?

jacek: i didnt say it :no_mouth:

struct: is it safe to do seed % number of positions?

struct: or should I use random(number of positions?

eulerscheZahl: both will do

struct: thanks

eulerscheZahl: seeds are so large that there is no notable bias

Default avatar.png ElectricToothpaste: hi guys im new here

Dark_Hack: eulerscheZahl i have a qusetion for you

DethHarmonik: Hello new here o/

Dark_Hack: are you admin or owner ?

eulerscheZahl: no, just a normal user

Dark_Hack: what

eulerscheZahl: not even a chat moderator. reach level 29 and you have the same powers as I do

Dark_Hack: i just saw you year ago

Dark_Hack: how

eulerscheZahl: i used to be a moderator (not staff) but decided to resign

Dark_Hack: do the owner know about you

Dark_Hack: or what?

eulerscheZahl: no idea. the community manager does

Dark_Hack: i just want to see you as manager or moderator

eulerscheZahl: those days are over, i'm rarely active anymore

Dark_Hack: i just lov your Bender

Dark_Hack: love

Astrobytes: Intriguing fantasy you have there Dark_Hack

Dark_Hack: ?

Astrobytes: Dark_Hack 05:04PM i just want to see you as manager or moderator

eulerscheZahl: i would be a horrible manager

eulerscheZahl: i can code. but delegating stuff? nope

Dark_Hack: hey yo

Dark_Hack: bro dint worry i sad just what i want

Dark_Hack: don't*

Dark_Hack: i know that it is impossible

Astrobytes: Yeah I just find it an intriguing idea :D

eulerscheZahl: not completely impossible. i'd just have to take the regular hiring process

Dark_Hack: you are moderator or what?

Astrobytes: He used to be, so did I, and a few others

eulerscheZahl: long story

Astrobytes: But that's history.

Dark_Hack: i asked about you Astrobytes

eulerscheZahl: looking forward to the next contest, i just hope it won't collide with topcoder

Astrobytes: Dark_Hack: I used to be, see my comment a few posts before this

Dark_Hack: yes but i ask

eulerscheZahl: https://www.codingame.com/learn/reading-the-statement

Dark_Hack: eulerscheZahl

Dark_Hack: do you know MadKnight

Dark_Hack: ?

Dark_Hack: and where is he now ?

eulerscheZahl: only from chatting and a short voice call

eulerscheZahl: and no idea on your 2nd question about him

Dark_Hack: thk

Dark_Hack: the system of getting points changed?

Dark_Hack: or what

eulerscheZahl: not recently

dbf: eulerscheZahl are you going to participate in hashcode?

eulerscheZahl: yes

eulerscheZahl: with Illeda n and Neuman n and dbdr if he reads our messages

dbf: wow, quite a strong team!

eulerscheZahl: not that strong for 4h contests. all of us would do better with more time i think

eulerscheZahl: we participated last year as well. team DEIN

eulerscheZahl: d = dbdr, e = eulerscheZahl, ... you get the idea

eulerscheZahl: will you play?

dbf: yes

eulerscheZahl: do i know your team members?

dbf: oreshnik is in your 'Following' list :)

eulerscheZahl: oh yes, I know her

dbf: last year result was around 2000, maybe it will be better this time :joy:

eulerscheZahl: would have been similar for us, Neuman n saved us

eulerscheZahl: just checked the leaderboard. we were 78th. better than i remembered

Astrobytes: impressive

eulerscheZahl: i felt so useless :D

DomiKo: HashCode is scam

Astrobytes: lol, this is why you build a strong team

dbf: 78 is nice, my best was around ~350 but still was not good to get into finals?

eulerscheZahl: highest ranked team with 4 different nations

eulerscheZahl: top 40 or 50 go to finals, i don't even know

Astrobytes: DomiKo: how so?

DomiKo: eulerscheZahl around 40

eulerscheZahl: 38 teams on the leaderboard of the finals

DomiKo: Astrobytes: in short. You got nothing if you are not on the podium.

Astrobytes: Ah I see, yes, that's a little harsh

eulerscheZahl: so you were on-site in the finals?

DomiKo: me? no

DomiKo: my friends were, but final was online

eulerscheZahl: ah

eulerscheZahl: not even some swag like stickers or tshirts?

dbf: or even fast-line google interview? :)

Astrobytes: lol

DomiKo: they didn't even get the prizes, because of problem with customs

eulerscheZahl: oh, story of my last RAIC

DomiKo: dbf: dont even dream about it

Astrobytes: like you with Russian prizes euler

eulerscheZahl: https://www.pochta.ru/tracking#RR148020554RU

eulerscheZahl: https://www.pochta.ru/tracking#RR148020727RU

eulerscheZahl: or me with topcoder tshirt supposedly (i don't have a tracking link but didn't receive anything)

DomiKo: HashCode only display then your team name, so you don't have any confirmation that you was at finals

dbf: oh, poor t-shirt

DomiKo: euler: my best was 6th at waitlist in 2019 :(

eulerscheZahl: and hoodie :(

dbf: I got hoodie and it colored my t-shirt, lol

Astrobytes: Yeah, that's pretty crappy DomiKo

Astrobytes: You would think Google would make a bit more effort to recognise strong competitors

DomiKo: But still, it's a pretty fun contest. But in terms of prizes big scam.

DomiKo: yeah Google is BIG name, so contest should be ...

dbf: eulerscheZahl, but it was not a problem with customs in your case, they were not able to deliver?

eulerscheZahl: i've read that topcoder paid $1k to finalists as finals weren't onsite

eulerscheZahl: customs blocked it

eulerscheZahl: and didn't even inform me. i only realized when it was too late to even ask what's wrong

dbf: hm, weird

eulerscheZahl: "Retained by customs in Germany"

eulerscheZahl: screw them!

eulerscheZahl: i know that a law changed less than a year ago

eulerscheZahl: in the past you could import stuff with a value below 22 Euro(?)

eulerscheZahl: now you have to pay taxes on everything. so the sender has to add a prize tag

eulerscheZahl: maybe that's what was missing. i have no idea

DomiKo: In my friend's case problem was that customs didn't have any confirmation that that was a prize, so they had to pay for it. A hoodie is a bit pricey.

eulerscheZahl: i even would have paid. but never had a choice

Astrobytes: a friend of mine in Brazil had a similar issue with a t-shirt gift recently

dbf: ah, I see. we had 1k euro limit earlier, that was lowered to 200 euro

Mrs.GloriaZindlebocker: In the US we have open borders, no problems shipping stuff

Astrobytes: and that was a t-shirt sent from the US to Brazil

DomiKo: only problem with US is shipping :(

eulerscheZahl: but that would be too easy. in germany we love paragraphs

Astrobytes: lol

eulerscheZahl: i even sent a mail to RAIC organizers if there's still a chance to receive it. but i don't expect a response

dbf: btw, I don't like design of hoodie - it is about space, but game was about knights and archers :)

eulerscheZahl: depends if you use the debug view

eulerscheZahl: illedan said it looks cool. haven't seen it myself

dbf: he was the only one who used it I think :)

Astrobytes: I like the t-shirt, in fact I'm wearing it now

DomiKo: Flex

eulerscheZahl: dbf just tried to cheer me up

Astrobytes: lol, hardly DomiKo :rofl:

RajdeepBiswas: lets goooo

eulerscheZahl: astro got a t-shirt and knows that i won a hoodie on top. not flexing

RajdeepBiswas: any team looking for a guy?

RajdeepBiswas: i'm up

RajdeepBiswas: for hire :P

Mrs.GloriaZindlebocker: Are you the famous Rajdeep Biswas we all know from github?

eulerscheZahl: "I am a musician turned programmer. (Also, a metalhead!)."

why did i read "methhead" at first?

eulerscheZahl: afk for dinner

Astrobytes: 'methhead' :rofl:

wlesavo: Astrobytes i am wearing hoodie on top of t-shirt :smiley:

RajdeepBiswas: @MrsGloryhole, probably not xD

Astrobytes: wlesavo: now *that's* a flex! :D

dbf: poor wlesavo, it is quite cold around? :)

wlesavo: just come home :slight_smile:

Astrobytes: I need a hoodie, the heating is broken in my house currently

dbf: Astrobytes, raic-2021 is coming!

wlesavo: raic soon, you will only have to win and wait another year :smiley:

Astrobytes: indeed :D hahaha

Mrs.GloriaZindlebocker: that guy just called me a gloryhole

dbf: but you will prepared if your heating will be broken in 2023

Astrobytes: dbf: very true, it pays to plan ahead!

Mrs.GloriaZindlebocker: you'd think that would be on wont's auto bad word list

Astrobytes: hahaha I didn't notice that, lengendary

Astrobytes: *legendary

Blokops: im stuck im purgatory clash rank 400 :{

Blokops: there always that one problems that tanks the rank

wlesavo: Blokops it time to try multis

Astrobytes: ^

Blokops: multis?

Blokops: is that a anime powerup?

Astrobytes: multiplayer bot programming

Blokops: ah same deal there

Blokops: or more like i dont have the time to reach legnedary

Blokops: gold is the best i can do

Blokops: still have nightmare from code of legends and magic

Blokops: worst place to start at

jacek: :notebook: :soccer:

jacek: https://www.codingame.com/multiplayer/easy/onitama

Astrobytes: lol

leojean890: Hy Astrobytes how is ur reversed_travelling-salesman :p

Astrobytes: leojean890: I've posted a proposed amendment in the forum

leojean890: so you prefer this version or reversing back ?:P

leojean890: I'll look that^^

Astrobytes: It's a larger issue than simply being reversed tbh

RajdeepBiswas: do you guys know anyone that works at openai?

leojean890: I agree, validators with input size to 500 - 1000 countries/nodes would be more interesting :)

Astrobytes: No. But an oldschool user from here works at Deepmind

RajdeepBiswas: will you be able to help me with their userame please?

Astrobytes: inoryy

Astrobytes: He's not online here much nowadays.

jacek: working on bioengineering cat girls

RajdeepBiswas: sweet, thanks anyway!

Astrobytes: He had a blog, maybe you can reach him there.

leojean890: for me I was ok with the size 300 to begin working on it, but to make the challenge more fun for all, and for a long time, bigger size is widely more interesting!

RajdeepBiswas: I just happened to find that via his profile, yes

Astrobytes: Not just bigger size, but each node is a group of sub-nodes

leojean890: but I'm not very used on working on this precise problem for now, tried to create a first node with greedy and go on with some hill climbing or SA

Astrobytes: This will prevent the use of existing TSP solvers and bruteforcing

leojean890: and euler gave me other ideas like Hamilton path I didn't know before

leojean890: yes I know, maybe the leaderboard will be fooled by some cheaters :p

eulerscheZahl: i even think that the reversed TSP is harder to solve than the intended one

Astrobytes: shouldn't have approved following eulers comment though.

Astrobytes: And I should have moved to reject pending more investigation so partially my fault too.

leojean890: but it's still a way to improve and train.. but I won't be against if bigger validators are given :)

Mrs.GloriaZindlebocker: the moderation bot will take care of it

Astrobytes: last resort

Mrs.GloriaZindlebocker: I wonder what happens if sentence tokenizer gets approved again

leojean890: but I prefer to have that than no TSP optim here^^

leojean890: so.. hopefully if it's thrown away it will come back with better validators

eulerscheZahl: it's not the testcases/validators. 1000 nodes won't change anything

Astrobytes: @Mrs um, Gloryhole? If it reappears in the Contribs it'll be derided until removed. That's the usual process for repeated re-uploads

eulerscheZahl: it's the game rules that shouldn't be vanilla TSP

Mrs.GloriaZindlebocker: heh heh. well he already submitted it again

jacek: oO

Astrobytes: lol, derision incoming

eulerscheZahl: what did i miss in the last weeks?

Mrs.GloriaZindlebocker: I had a deleted clash get reapproved, turns out you don't get xp the 2nd time

eulerscheZahl: like when you promote to the same league again

eulerscheZahl: when CG opens it early by accident

Mrs.GloriaZindlebocker: about the only thing you missed is wont turning from a guy who never kicks into a ban machine

eulerscheZahl: and I never good XP for the lowest wood league at OoC

eulerscheZahl: wonto found the ban command? :o

leojean890: about TSP, I think that I don't know enough the original topic so maybe I don't see everything you see which are bad

leojean890: in this contrib

Astrobytes: leojean890: it was initially pointed out to myself and darkhorse64 by daporan last night

Astrobytes: then euler elaborated some more this morning

eulerscheZahl: and i read the chat logs this morning, found the puzzle and agreed with daporan

eulerscheZahl: then decided to rant

Astrobytes: Not a rant per se, was a very helpful comment.

eulerscheZahl: never hurt to have a bit of theoretical background (though i don't know all the details of how it works)

leojean890: but me I was happy to have a new toy to play with xD

eulerscheZahl: some simplex solver involved

leojean890: I missed new optims, I did something on all other ^^

leojean890: except CG funge prime

Astrobytes: Yeah, as I said in chat last night, I'd like to learn a bit more graph theory tbh

eulerscheZahl: CG Funge isn't a classic optim so that's fine

leojean890: same :P

Astrobytes: hm, that spreading fire optim is still in contribs

leojean890: what do you think of it ? didn't try yet :)

Astrobytes: I've not tried it, I just remembered it from some time ago

leojean890: but it may be interesting,

struct: ups forgot seed can be negative :)

eulerscheZahl: personally i would have taken random.next() without even considering your way

struct: is just to make seed 0 = empty board

struct: and so on

eulerscheZahl: just feels cleaner from a software engineering point for me, i can't really reason it

eulerscheZahl: i see. topcoder always hardcodes the parameters on seed 1 and 2

eulerscheZahl: and i still haven't looked into the game you want to create

struct: probably just another mcts game

struct: dominated by NNs

eulerscheZahl: :tada:

struct: by jacek*

Mrs.GloriaZindlebocker: man, if you turn on the olympics in the US the broadcasters basically tell you not to watch it

struct: why?

Mrs.GloriaZindlebocker: politics

eulerscheZahl: until just now i completely forgot about the olympics again

eulerscheZahl: and probably i'll forget it until tomorrow

struct: not a big fan of winter olympics

eulerscheZahl: not a big fan of watching others do sports

jacek: :older_man:

eulerscheZahl: oh, a video about bulls & cows (haven't watched it yet) https://www.youtube.com/watch?v=v68zYyaEmEA&ab_channel=3Blue1Brown

jacek: struct ataxx potw tomorrow or delete

struct: lol

struct: is this new? https://www.codingame.com/playgrounds/25775/codingame-sdk-documentation/anim-module

struct: I never saw it before

eulerscheZahl: community contribution by butanium

eulerscheZahl: you can make pull requests with your own modules

jacek: did he publish his contribution?

eulerscheZahl: oh wait, i think i confused it

struct: repo seems to be 4 years old maybe its not new

struct: I just cant remember it seeing it before

ProCrazyCoder: print("Hello people")

struct: hi

eulerscheZahl: how shall i know if butaniums game is still WIP? https://i.imgur.com/bMlOJF8.png

jacek: and he renamed it from clash of bots to somehting else

struct: (╯°□°)╯︵ ┻━┻

Mr.WhiskeyPerson: Hi

struct: hi

jacek: hm?

struct: I was sending an empty string and messing up stdin on java

struct: and wasnt finding what was wrong

Mr.WhiskeyPerson: How is Code Keeper a multiplayer game?

jacek: and quite polished

Mr.WhiskeyPerson: Yeah, but I wonder how you interact with the other players

DomiKo: you don't

jacek: i think the same question people asked during contribution

Mr.WhiskeyPerson: So it is basically a puzzle?

jacek: approvers these days...

DomiKo: :sweat_smile:

jacek: this could be baasically an optim

DomiKo: This could be in future

Mr.WhiskeyPerson: Ok, I'll try something else.

jacek: :notebook: :soccer:

Mr.WhiskeyPerson: hm?

jacek: nevermind

Mr.WhiskeyPerson: The game made by you? Paper Soccer?

DomiKo: you have to try paper soccer it's super fun game

Scarfield: have you had a branding course, you really have been pushing paper soccer today xD

jacek: really?

jacek: oh my

struct: I dont know if I should player output multiple moves or force opponent to pass and make the player go again

struct: the first one is probably harder for newbies

jacek: can you make it like othello?

struct: currently is like othello

Scarfield: which game did you end up porting, ataxx?

struct: yes

struct: i should do backgammon next

Scarfield: kewl

struct: since its different

Scarfield: nice

struct: no idea how to make an ai for it

Wontonimo: rand()

Wontonimo: done

jacek: eh http://incompleteideas.net/book/ebook/node108.html

struct: I see RL I close

jacek: real life?

struct: yes

Wontonimo: it looks like 1-hots again

struct: im more interested in backgammon since I dont think I will be able to use mcts there

jacek: and has random element

Wontonimo: why would MCTS not work?

Scarfield: there are some basic heuristics you can use, "bridge" having more than 1 piece on more positions in a row is good. Having a single piece on a position is bad.

struct: it might but i have no idea how to code an eval for this game

jacek: plot twist: mcts doesnt need eval

Wontonimo: win = 1 , lose = -1

struct: there are so many combinations possible

struct: if you take into account all rolls possible

Wontonimo: i still haven't upgraded my mcts solver to deal with ties

struct: brb

jacek: oO

Wontonimo: i think the stochastic nature of MCTS will just take care of probabilities naturally

Wontonimo: oO - what I mean jacek is that it doesn't solve for a tie in the same way it will solve for a win/lose. A tie isn't treated as a solvable end state, but just like another value

jacek: draws can tricky, especially in selection

Mr.WhiskeyPerson: "The typical size of pitch is 8x10"

Mr.WhiskeyPerson: Not really typical if it is a hard fact?

jacek: approvers didn't complain :v

Mr.WhiskeyPerson: How typical of them

jacek: yes, its hardcoded for CG

Default avatar.png CyberCoder786: https://www.codingame.com/clashofcode/clash/2215371b0ae8daf3225d51ba2449e6736fb7d71

antiwonto: [automated] hey CyberCoder786, dont paste those links here. Use the channel #clash

Ziatzoraf: HELLO GUYSSS

Ziatzoraf: <3

Niikurasu3: hey

Wontonimo: award jacek a :taco:

antiwonto: [automated] Wontonimo has awarded jacek 10 taco. jacek now has 10 tacos. Wontonimo now has 1 tacos

Wontonimo: woot, automatic taco functionality!

Mr.WhiskeyPerson: award Wontonimo a :taco:

Wontonimo: hmm ... interesting, nothing happened

Wontonimo: oh well

Wontonimo: award a taco to someone by using the :taco: icon and their name. No need to use 'award' or 'to'

Wontonimo: i have no idea why it didn't work for you Mr.WhiskeyPerson

Astrobytes: Wontonimo :taco:

antiwonto: [automated] Astrobytes has awarded Wontonimo 10 taco. Wontonimo now has 11 tacos. Astrobytes now has 1 tacos

Wontonimo: woot !

Wontonimo: Thanks Astrobytes!

Astrobytes: lol np

Mrs.GloriaZindlebocker: Westicles :taco:

DethHarmonik: Useful :ok_hand:

Wontonimo: Mrs.GloriaZindlebocker, it has to be a user that the bot has recently seen

Mrs.GloriaZindlebocker: ah, okay

Quidome: Wontonimo :taco:

Astrobytes: DethHarmonik: well, the anti-clash functionality is useful, the taco is just for fun ;)

antiwonto: [automated] Quidome has awarded Wontonimo 10 taco. Wontonimo now has 21 tacos. Quidome now has 1 tacos

Quidome: fun easter egg are there more?

Wontonimo: nope.

DethHarmonik: How many tacos a user has at start?

Wontonimo: zero

Mrs.GloriaZindlebocker: :popcorn: when you launch it on #fr

Astrobytes: DethHarmonik :taco:

Wontonimo: I'll have to get some help with French to do that

antiwonto: [automated] sorry Astrobytes but you can only award 1 taco a day

Astrobytes: nice

Astrobytes: seems working as intended

Quidome: Astrobytes :taco:

antiwonto: [automated] sorry Quidome but you can only award 1 taco a day

Mrs.GloriaZindlebocker: Zorg1 :taco:

antiwonto: [automated] Mrs.GloriaZindlebocker has awarded Zorg1 10 taco. Zorg1 now has 10 tacos. Mrs.GloriaZindlebocker now has 1 tacos

Quidome: hm not allowed to give more then one taco

Astrobytes: :rofl:

Astrobytes: naughty

DethHarmonik: Astrobytes :taco:

antiwonto: [automated] DethHarmonik has awarded Astrobytes 10 taco. Astrobytes now has 11 tacos. DethHarmonik now has 1 tacos

jacek: Automaton2000 :taco:

Default avatar.png Automaton2000: how can you have a whole day

antiwonto: [automated] jacek has awarded Automaton2000 10 taco. Automaton2000 now has 10 tacos. jacek now has 11 tacos

Default avatar.png Automaton2000: i don't want to resubmit

Quidome: it's going viral :joy:

DethHarmonik: So. From where did I took this :taco: Wontonimo?

DethHarmonik: If I had none

Astrobytes: antiwonto gifts it to you for being generous

Wontonimo: when you were making them to give, you kept one for yourself and ate one. you made 12

Wontonimo: of

Wontonimo: course

Wontonimo: who doesn't make 12 tacos at a time?

DethHarmonik: Hmm...

Astrobytes: 12? Rookie numbers

Wontonimo: lol

Wontonimo: well, now we have a taco-based reputation system

DethHarmonik: I see you created a bot who send messages to prevent clash link spam, but favorise tacos spam...

Astrobytes: haha, perhaps it should have a finite supply of tacos

str_name: anyone wanna play smash bros?

Wontonimo: i'll be very happy when Automaton2000 awards someone a taco

Default avatar.png Automaton2000: i will add it to the teacher

Astrobytes: oO

jacek: :upside_down:

Wontonimo: if it get's out of hand and becomes spam, let me know and I'll disable it for a while.

Astrobytes: will do

DethHarmonik: tacocurrency

Majeck: Hullo

Majeck: What exactly are "events"?

Majeck: I read the description, but what's the difference between for example a regular Clash of Code and registering for 14th Feb's CoC?

Wontonimo: promotion of things that are already happening

Wontonimo: but they want all users to focus on that one thing

Wontonimo: because it is actually really fun when everyone talks and works on the same thing

Majeck: Ohhhh, so like main chat focusing on it and so

Majeck: That's awesome!

Wontonimo: yeah

Wontonimo: that's my 0.02

Mrs.GloriaZindlebocker: and wonto can tell the clash guys they are wasting their lives

Astrobytes: :D

DethHarmonik: And give them tacos

Wontonimo: NO TACOS FOR CLASH !

Astrobytes: Make tacos not clashes :peace:

Wontonimo: i've had so much fun with this bot, my latest contribution is being neglected

darkhorse64: TSP leaderboard is fixed but not reset. Time to resubmit

Scarfield: Scarfield :taco:

antiwonto: [automated] You used your taco giving ability for the day to discover that you cannot give tacos to yourself

Scarfield: xD

Blokops: well back to 400th rank in class of rank

Blokops: lets see how long i can stay there

Blokops: a yes class of rank

Blokops: clash of codes

Default avatar.png dmuraco: :stuck_out_tongue_winking_eye:

Default avatar.png dmuraco: :baby_symbol:

Default avatar.png OpenRedroot: Hey

DomiKo: TSP leaderboard is fixed, but statement not :(

Astrobytes: and the author rejected my suggestion

DethHarmonik: Oh?

DomiKo: yeah I see

Astrobytes: check the forum or the contrib. Oh well, shrug

DomiKo: ehhh

DomiKo: you can even hardcode all testcases without problems :(

DethHarmonik: Well... Time to mass 1:star: so...

struct: ok finally added some animations

struct: I will probably add sprites after a few days

struct: https://i.imgur.com/VnsZgOm.gif

struct: sry for lag

darkhorse64: Nice

struct: color wise everything will be changed I just wanted to get animations ready

DomiKo: looks cool

Astrobytes: nice job struct

struct: thanks

darkhorse64: Yeah, animations look right. Some fancy sprites and you got a hit

Astrobytes: ^

struct: cant forget to add a soundtrack :p

Astrobytes: lol, 8-bit demoscene tracker techno ftw

leojean890: too bad if he didn't fix the statement if it's really fast to do so hmm

leojean890: I tried to add more nodes but the game Engine throwes Exceptions for larger inputs and I don’t know how to fix it. => too bad :o why those Exceptions...

struct: leojean its hard limit on sdk

Default avatar.png cyberblitz_: I'm getting the feeling the group who have created Codingame don't have English as their first language. Some of the English is not that great

Mrs.GloriaZindlebocker: cyberblitz_ got to be runnin' cold upside down his head, you know?

faouziachour: Bonsoir

Thorcode: wonder what puzzle the event choose XD