Chat:World/2021-03-28

From CG community
Revision as of 12:03, 15 June 2021 by Chat Log (talk | contribs) (Created page with "File:Default_avatar.png Edzz: What's up <img src=/a/60779622006508> DanniBoy: yoo <img src=/a/60764386369579> WesCraterMadman: hey I'm working on The Labyrinth puzzle an...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Default avatar.png Edzz: What's up

DanniBoy: yoo

WesCraterMadman: hey I'm working on The Labyrinth puzzle and I think I'm missing some key insight. Anyone here have experience with that one?

WesCraterMadman: aight good taljk

davilla: haven't done that one, sorry

Default avatar.png masterbhai: \hrw.agrw

Default avatar.png masterbhai: eghiruagiruw

Default avatar.png masterbhai: fgouwgiruw

KiwiTae: WesCraterMadman yaaa

KiwiTae: did it

jacek: :no_mouth:

ZuhairAbid: hey

ZuhairAbid: can anyone help me?

jacek: ask specific question

ZuhairAbid: How can I find which problems I have solved

jacek: puzzles?

ZuhairAbid: no problems

ZuhairAbid: algorithm problems

Doju: argh i hate this language so much

Doju: how can anyone enjoy writing c++

ZuhairAbid: dont write in C++ then

Doju: i need the speed though :/

ZuhairAbid: what about javascript?

Doju: i'd very much like to write in python but its so slow

jacek: in puzzles page at the bottom you have puzzles and/or problems you solved. for clashes there is not such thing

Doju: ZuhairAbid for competitions i kinda need c++

Doju: or at least a compiled language

jacek: c is also fast :v

ZuhairAbid: like competetive programming?

jacek: multiplayer bot programming

Doju: ZuhairAbid the challenges on this site

MSmits: Doju learn rust

Doju: like the upcoming spring challenge

MSmits: it's just as fast

jacek: :scream:

MSmits: and much more enjoyable, apparently

Doju: MSmits thanks for the recommendation

Doju: I'll have a look

MSmits: we have some experts here, try dbdr if he is on

jacek: for spring challenge you wont likely need that much speed. java or c# would suffice

jacek: mostly likely it will be heuristics battle

MSmits: in most contests you will do fine with C# and java yes.

MSmits: the simpler the game,the more performance will matter

Default avatar.png PrivateDidgeridoo_374e: can anyone help me that how can I make strtol alike function?

Doju: Yeah

MSmits: and most of the time these games are not simple

Doju: I did decently well in the pac one with a heuristics bot written in python

Doju: in the fall challenge i did terribly with a search written in python

MSmits: thats actually a game where performance helps more than average, (the pacman one)

dbdr: jacek why do you think speed won't matter?

MSmits: Doju the rust guy just arrived

ZuhairAbid: @ jacek wheres the puzzle page

dbdr: :)

Doju: Oo

ZuhairAbid: wheres the puzzle page?

jacek: dbdr for challenges at least in the beginning its heuristic battle

jacek: https://www.codingame.com/training

MSmits: i skip that, i spend 3 days writing a search bot and then shoot up to silver :P

Doju: Hey dbdr any tips on learning rust?

dbdr: yes, but at the end perf often matters

Doju: And do you think I should learn rust instead of c++ for the challenges here?

dbdr: Doju what languages do you already know?

Doju: python and javascript

AntiSquid: what's up with the yellow dots on "YOUR LAST ACTIVITIES" ?

AntiSquid: https://imgur.com/a/aGcuFq0

dbdr: ok. personally I would recommend rust over cpp for challenges. however if you want to get a job with it too you have much more options with cpp

AntiSquid: not all have it, but it's there

AntiSquid: ontopic: rust syntax is a bit confusing to me

jacek: AntiSquid i have tick there

MSmits: he already said he hates writing cpp though, so i doubt he would want a job coding it. He would hate his job

dbdr: ok, missed that part

Doju: Well, I think i'd enjoy the language a bit more if i understood it

Doju: but at least for now it's pure pain

Doju: to do anything

MSmits: well it does have a steep learning curve

MSmits: I am at the bottom of it still and no incentive to climb up

dbdr: AntiSquid: that's fair enough, but syntax does not stay a problem if you keep using a language, it's just the early days

AntiSquid: so what's the speed difference rust vs c++ ?

dbdr: very little

MSmits: I dont think there is one AntiSquid, from what i have read. It depends on the operation you are doing

MSmits: it's in the fuzzy area

AntiSquid: didn't even look into setting it up locally to use for other stuff, but this would matter to me, along with what other use cases it has apart from CG

AntiSquid: for me personally i mean

dbdr: are you guys doing any multi these days?

jacek: what else should we do? clashes?

MSmits: i dont have much time atm, but I am going to try machine learning once i do. I converted the xor example to c++ last night

AntiSquid: i did, i know it's hard to notice when my rank / cp doesn't change much dbdr :P

MSmits: then will try on oware

dbdr: I just started TryAngle, was easy to climb up, so I guess not many have tried hard

AntiSquid: there are too many multis and too little time for me

AntiSquid: i want to have a look at all of them, so i just made a "ranking system" to see what i prioritise based on personal preferences :D

dbdr: I have a physical oware now by the way :)

jacek: with seeds?

MSmits: nice

AntiSquid: does it have real biodegradable beans?

dbdr: no, stones

MSmits: you mean you kept an egg carton instead of throwing it away dbdr?

dbdr: lol

AntiSquid: you only need half a carton

MSmits: mmh dont you need 12 holes?

AntiSquid: use the other half to simulate future moves while making your opponent wait

MSmits: they package eggs weirdly in GB?

AntiSquid: i was thinking of 12 egg boxes

MSmits: oh right

AntiSquid: there's 6 egg, 12 egg, 15 and even 30

jacek: huh? there are only 10 and 30 here

dbdr: unknown guy is #1 in TryAngle

MSmits: btw, i apologize now for changing the subject to eggs

AntiSquid: anything higher than that you get in boxes of stacked up 30 egg trays

dbdr: kovi #2

kovi: maniak is moderately strong topcoder

jacek: never heard of him

dbdr: oh ok, just not active here

AntiSquid: lots of unknown guys climbing high, in 1 or 2 games, they clearly don't have much time either

kovi: its a nice heuristic game, would have been a greet early2021 challenge

AntiSquid: are you doing top coder jacek ?

MSmits: I'm ok with that, as long as each game has some reasonable spread of opponents strengths

jacek: no

dbdr: kovi, I don't even take opponent into account :D

AntiSquid: having an AutomatonNN moment then ?

AutomatonNN: why really cant do this?

dbdr: and ignore most of the rules

MSmits: like muzero

MSmits: thats pro

kovi: after bandas i might get back to tryangle

AntiSquid: i expect another sim favoring game in next contest

MSmits: what does that mean

MSmits: do you mean a search favoring game?

AntiSquid: like fall2020

MSmits: i mean any good bot in any game will use sim

MSmits: even heuristic bots use sim

MSmits: i think you mean search

MSmits: like, search to some depth

MSmits: if you find a path with pacman to find the closest thingy to eat, you're simming, because BFS is a simulation

AntiSquid: i say sim, because for heuristic you don't need high focus on them

dbdr: you can have a pure heuri without sim. just if then else spagetthi :)

MSmits: usually when people say they only use if else, they still sim. Some of those conditons will usuall contain some simulation element

AntiSquid: it's more about knowing what are best moves for given states, you don't need to always predict how it affects the game

kovi: in my view heuristic = simplified model can include sim/search


MSmits: yeah

MSmits: thats what I meant

AntiSquid: with heuristics i mean something that resembles doing NN with handcoded / submit spam derived weights :P

MSmits: ohh... real AI :)

AntiSquid: lots of hype recently about real AI in papers and videos

AntiSquid: but then when you read the fine print ... you see it's more speculation that we are closer

MSmits: AI is a super vague term anyway

MSmits: people used to call a BFS algorithm AI

MSmits: because it looks intelligent

MSmits: ohh, it goes around obstacles...

AntiSquid: overly used and misused ?

MSmits: hard to say it is misused when it is poorly defined

MSmits: but overly used, yeah

MSmits: it also annoys me that when people discuss AI it is automatically assumed to be some form of machine learning

MSmits: I feel AI is about acting intelligently, no matter how this was achieved

MSmits: and AI has a whole history that predates machine learning

Passifi: well aren't we like peak hype cycle for AI in general period ? I really wonder whether it will truly serve me my big mac in the next 10 years ...

MSmits: it will not

MSmits: it will try to convince you to eat healthier

Passifi: fair chance

Passifi: so it will serve me mc donalds salad the future is truly a darl place

Passifi: *dark

MSmits: it will probably sabotage all mcdonalds to save the human race

AntiSquid: well back your SIM + BFS analogy, is that AI or just an algo to you ?

KiwiTae: it will feed u the momery of eating a bigmac after you ate your salad

MSmits: well you're asking me what the definition of AI is, but I am not the person who decides what words mean.

KiwiTae: in the industry they call AI anything that automates tasks..

MSmits: BFS causes intelligent behavior though

jacek: imo AI is broader term. id say its AI if it acts smart, even if its just some ifology

KiwiTae: brings investors in cause its a cool word

MSmits: and a simulation is needed to predict the future, also a part of intelligence

AntiSquid: you need to research the MCdonalds salad dressing Passifi, on the topic of eating healthy :)

MSmits: i tend to agree with jacek

Passifi: oh I

MSmits: in my ai course i started with pathfinding

Passifi: d never eat their salad

MSmits: then some optimization, minimax, q learning and NNs

jacek: https://img-9gag-fun.9cache.com/photo/aOYA1mE_700bwp.webp

Passifi: I think they got rid of it in the US because they thought it actually hurt business ^

AntiSquid: ya Kiwi, do they call an automated script AI too ?

Passifi: You go to a fast food place to greasen up your arteries so that the blood flows properly again. Everyone knows that

jacek: i heard they serve salad in macdonalds even if people dont eat it, because it makes them less guilty of eating unhealthy

KiwiTae: AntiSquid yea...

MSmits: makes the food more colourful

KiwiTae: some game AI are just scripted... go straight until hitbox then jump... repeat

MSmits: I remember eating at a Chinese restaurant with my family and they made these pretty flowers out of carrots

MSmits: I ate everyone's flowercarrot and then my family said i shouldnt because they were just decoration

jacek: and made of plastic

KiwiTae: they serve cornsoup in mc donalds in taipei so good

MSmits: lol no

AntiSquid: is it proper US GMO cornsoup ?

MSmits: aren't dogs basically GMD's ?

MSmits: even though we didnt do it in a lab

MSmits: or GMW's

AntiSquid: something like this (carrots) or were they spiralled together ? https://i.pinimg.com/originals/95/bf/cb/95bfcb0a8d0dbbe1130cf1ea5c5fefff.jpg

MSmits: we made species that have noses too short to breathe properly or with skulls to small for their brain cuz we thought they looked cute

AntiSquid: i know what you mean MSmits

MSmits: AntiSquid these look better, they were more spiralled i think

Passifi: I always thought this was a weird simily breeding selectivly isn't the same as GMing something. There is plenty of healthy genetic chaos in the mix which clearly smoothens out the human influence also where we overdid it. It did lead to severe problems for the dogs nevertheless...

MSmits: it's not exactly the same, but you're still removing natural selection from the equation and selecting by some other preference

dbdr: it's another form of natural selection

MSmits: as evidenced by the difference between breeds of dogs, this goes very very far

AntiSquid: controlled natural selection

MSmits: dbdr sure but that's taking a very wide definition of "natural"

dbdr: humans are natural animals :)

MSmits: right

AntiSquid: well when you use viruses, agro bacteria or chemicals to modify something then it seems unnatural to me

Default avatar.png armalitedev: pog

MSmits: it happens naturally also

AntiSquid: and they are playing basically with this stuff, the science isn't that well researched yet, i mean there wasn't enough time

MSmits: yeah there's risk

MSmits: but not the risk people think it has

AntiSquid: obviously since it's funded, "someone" wants to see results and earn profits from the investments

MSmits: people think the stuff is not safe to eat, but thats nonsense

MSmits: it's mostly that the stuff might become a plague and drown out all other species

Default avatar.png armalitedev: amogus

Passifi: there is a huge difference in speed and concentration and that can lead to pretty dire outcomes later on think of chinese super babies that just drop dead by the age of 20

Default avatar.png armalitedev: sus

MSmits: Passifi huh?

MSmits: do you mean because they ate corn?

Passifi: also there was artificial food we considered to be safe and sound and then later on we found it to be harmful like trans fats :)

Passifi: oh I was thinking gm in general

Passifi: although corn syrup for example is pretty bad for your health ;)

MSmits: you have to be careful about comparing things like those transfats and gm corn though

MSmits: it's still corn, it just does what it does more efficiently

AntiSquid: workers that drop dead when they reach their expiry date so they won't be a burden on the pension fund :thinking:

MSmits: well if they made it so they wouldnt expire at all, that would be better

AntiSquid: sure everything is possible, but then would we have highly restrictive lockdowns when a virus could supposedly get rid of all the pensioneers ?

MSmits: no i meant, just work for eternity, who said anything about pensioneers

MSmits: these guys dont age

AntiSquid: talking to passifi :P i am still reading as i type

MSmits: lol

leojean890: when I took AI course at school, I only learnt "minmax", "backtracking" ^^

leojean890: (2012)

MSmits: yeah that makes sense

Passifi: but with additional experience they might get ideas and then revolt better to have them drop dead I say

leojean890: I suppose now it'sonly ML

MSmits: dont think so

MSmits: i think they still start with algorithms such as bfs

leojean890: many people think that ML == AI

MSmits: CS 50 at harvard starts with BFS

leojean890: yes, BFS is important to understand before learning the rest

MSmits: when it's discussed by popular science things, like news/magazines yeah leojean890

AntiSquid: Passifi, not a conspiracy, it's something china considers for their soldiers, helmets that detonate when needed, example soldier gets captured and might be interogated

leojean890: sometimes they make an "advanced algo" course learning BFS

Passifi: the courses I've taken all start with teaching me how to forget all the calculus I've learned and replace it with a different opimization method ;)

MSmits: I would just find a good old soup bowl

leojean890: and then an AI course ^^

AntiSquid: there was a video about ML just being a phase and after we make the process of automating the learning better we can focus on AI that derives new ideas from what it learns on it's own

AntiSquid: wait i will link the video that better explains this :P

MSmits: i should get back to work on my studies... deadlines approaching

MSmits: ttyl guys

Passifi: as long as we can make the ai follow the three laws we'll be fine :)

AntiSquid: nvm can't find it, i think it was something Francois Chollet wrote and not a video actually, worth looking up this author

AntiSquid: Passifi law 0 most important one, you don't want grown up babies that are taken care of by AI

Passifi: Well what makes me the most worried about the future is the amount of people who think being a grown up baby being pampered by AI is a Utopian World ;)

Passifi: i've seen what a life without purpose does to people. It aint pretty, and once they got used to it its hard to get out of it.

Dekku: baro yleebo

Dekku: <3

dbdr: Passifi on what kind of people have you seen that?

Passifi: handicapped people I took care of back when I worked in that area. Retired elderly too. Now one would think its their deseases but nah those who got something to do( and not just busy work) often do very well :)

Passifi: But those who don't... well it's just a sad sight to behold

dbdr: yeah, feeling useful is super important

AntiSquid: i used to work in a care home too, the time of the day when i went around to write down what each resident wants to eat was pretty much one of the highlight of their day, they'd open up and talk a lot ...

davilla: man, how much XP do I need to reach level 20?

davilla: I mean, are all the levels the same?

Yassine_EL_HOSNI: nop

davilla: this is going to take a while

Yassine_EL_HOSNI: ok you are level 14.. and i am 13.. i need 523xp in total.. to get to level 14.. what about you?

Yassine_EL_HOSNI: it should be +523

davilla: it says Level 14 (i) ... 372/580 XP

Yassine_EL_HOSNI: 523->580->...

Yassine_EL_HOSNI: you need to find the function they are using to calculate XP growth

davilla: but how can I be on level 14 if I only have 372?

Yassine_EL_HOSNI: i am guessing it is more than just i+1-i

davilla: oh, that's 372 just in level 14

Yassine_EL_HOSNI: it means eche level strats form 0 to it max xp

davilla: right

Yassine_EL_HOSNI: go do all the easy puzzles and you might get level 20 in few days

davilla: the easy puzzles are only +50 XP

Yassine_EL_HOSNI: the easy puzzles can sometimes be hard.. at least from my experience hhhhh

Yassine_EL_HOSNI: anyways have fun

davilla: so how many XP is level 1?

davilla: working backwards it's easy to get 55, which I think I remember seeing

AntiSquid: davilla https://www.codingame.com/training/medium/the-experience-for-creating-puzzles

WOLFRAHH: hii AntiSquid

davilla: ah, thanks

AntiSquid: you trying to get banned again WOLFRAHH ?

davilla: in the form of a puzzle :-P

WOLFRAHH: why

AntiSquid: you tell me why

WOLFRAHH: ohh for saying hii

WOLFRAHH: sorry

AntiSquid: lol no, that's just the last nail in the coffin ...

WOLFRAHH: i don't understand

AntiSquid: consider the previous bans and consider what i told you before

WOLFRAHH: yeah but i don't do anything

davilla: OK so I just could solve 6 very hard puzzles or 60+ easy puzzles to get there.

davilla: ooh how do you get collaboration?

AntiSquid: ya classic ones only, newer ones only give 50xp

davilla: oh the classic are worth more? good to know

AntiSquid: what do you mean? collaboration

davilla: WolfR has Coding Speed: Gold level, Collaboration: Silver level

jacek: this mean he upvoted/approved or contributed in contributions

AntiSquid: ah that's part of the quest map on the "front" page

AntiSquid: https://www.codingame.com/home scroll down davilla, just do those tasks and you can get that too

davilla: that's just my quest map, I don't see any tasks

icecream24: I just realized there is VIM mode for the editor. I am so happy :'D

jacek: now you cant quit it

AntiSquid: some of the quests on the quest map unlock those "certifications"

davilla: oh OK thanks

davilla: I guess that must be down the Clash of Code path

davilla: since that's the main one he seems to follow

jacek: :scream:

WOLFRAHH: hey can anybody tell me how i can be in 1000 position in clash of code it is my task in quest map

jacek: do some clashes and dont be last

davilla: write a program that minifies your programs

WOLFRAHH: davilla are you saying to me

WOLFRAHH: or saying to jacek

davilla: you, it applies to your question

davilla: but I was joking

AntiSquid: each path has its own certifications

WOLFRAHH: yeah my coding speed is gold level

AntiSquid: you need to keep doing clash until you are top 1000 basically

davilla: and then you need to stop doing clash

WOLFRAHH: ohh why

davilla: lest you fall out of top 1000

WOLFRAHH: davilla

davilla: no I'm joking again

WOLFRAHH: ohh comon davilla

WOLFRAHH: but i am solving CSB to wood 1 to bronze or my code beat the boss

AntiSquid: my last cert for collaboration path: i need to make a contribution, either quick clash puzzle (i won't), puzzle or game, still overplanning the game

davilla: now that's fun, I love the creative side

davilla: made a few puzzles on codewars

AntiSquid: you just need 1 and no creativity is required :D

davilla: so do you want to bounce ideas or you have a plan already?

AntiSquid: i have a plan, for a long time now, just not implemented

davilla: ah OK I'm curious if you want to share

AntiSquid: multiple ideas, but no don't want to share, sorry

jacek: :mad:

jacek: :rage:

davilla: hey no problem, but I guess you don't get that collaboration badge then

davilla: :-P

davilla: good luck, anyway

AntiSquid: i do, that's the weird part about "collaboration", you can do it alone

davilla: lol

davilla: is there any real collaboration that goes on here?

Yassine_EL_HOSNI: hhhh

jacek: fireworks game :v

davilla: I notices on codewars sometimes people got together to solve

Yassine_EL_HOSNI: let us collaborate in solving my puzzles

davilla: what language? or is that a joke?

Yassine_EL_HOSNI: just a joke hhhh

jacek: oO

davilla: OK cooperative game, there's a twise

davilla: twist*

WOLFRAHH: hey AntiSquid why u don't created any contribution to complete task on quest map

WOLFRAHH: Or you have achievements I'am a creator

AntiSquid: that's different

jacek: hes moderator

AntiSquid: again different

jacek: not a squid?

WOLFRAHH: he is moderator this not make a sense jacek

jacek: i agree

GooMooc: hi all, someone know why a code work in my local python interpreter and return a TypeError in the Codingame interpretor?? How can i solve it ?

davilla: what version are you running locally?

WOLFRAHH: yeah same queston GooMooc

GooMooc: TypeError: sort() takes no positional argument

WOLFRAHH: we are saying which version of python interpreter you have

WOLFRAHH: GooMooc

davilla: how are you using sort? are you giving it a key?

davilla: or reversing the order?

GooMooc: python 3.6

WOLFRAHH: ohh

GooMooc: i use ide geany and the command line are the same in local and codingame :

WOLFRAHH: i think codingame is using 3.8

GooMooc: sfp.sort(sfp.sort(key=lambda x : x[0]), key=lambda y : y[1])


GooMooc: it work locally and return sfp sort by 2 criteria

davilla: sort doesn't return anything, does it?

GooMooc: it change my sfp list order

GooMooc: sfp is a tuple list with coordinate x and y

davilla: that's a bit odd, I would write it in 2 lines

davilla: or there's an easy way to combine it into a single sort

WOLFRAHH: i agree with davilla

GooMooc: ok, i know it could be written better or differently, but what it work in local and not in codingame ... the difference of version is the explanation???

davilla: anyway I believe the reason you're getting the error is because it's giving None as an argument, since the inner sort doesn't return anything

GooMooc: None as an argument? i dont understand

davilla: myList.sort(None) gives the error you stated

davilla: in 3.8

GooMooc: def Limit2Center(xo,yo,center,sfp): http://chat.codingame.com/pastebin/16ad69ce-b5bc-4497-9358-c98cb889ccbd

davilla: I don't have 3.6 on this computer but my guess is something has changed, even tho the docs are about the same

GooMooc: ok so ill try to update my python and retry the function

GooMooc: def Limit2Center(xo,yo,center,sfp): http://chat.codingame.com/pastebin/545c2099-02ef-49b0-abcf-5bb4c434dc2a

davilla: the key insight here, if you're keen on learning, is that sort doesn't return anything

davilla: a = [1, 3, 2] b = a.sort() print(b) # None

davilla: print(a) # [1, 2, 3]

GooMooc: You surely know more than me Davilla and thank for your answer but it work locally and return the list sorted as i want ... so i ve to update

davilla: OK is it clear how to update?

GooMooc: sudo apt get update

davilla: oh you mean update the version of python, yeah that would probably be helpful

Uljahn: also you don't need to sort to get the first element, you could use min/max with lambda key

davilla: good point, that would surely be faster

GooMooc: yes, there are other way... but i want to sort with multicriteria

davilla: in that case I would do it in 2 lines

davilla: first the one, then the other

Uljahn: you can use a tuple of criterias in lambda function

Uljahn: key=lambda x : (x[0], x[1])

davilla: I love python because it lets you do stuff like that

GooMooc: yes i ve beggining writing in two lines but only one seeming better

davilla: well there are a couple of ways to write it on one line

GooMooc: ty uljahn iv try this code but was not working locally

davilla: one way is as mentioned with a single more complex key another way is to use sorted() instead of sort()

GooMooc: i think it was returning None, so ive search an other way locally... and now it dont work in CG... lol

GooMooc: yes i ve seen sorted but each loop my list change and i dont need to save it so sort was indicate

davilla: cool

davilla: well the complexity doesn't always follow the number of lines, I think is a key point

davilla: min() is the fastest way for a single element, then sort() with a complex key, then sorted(), then 2 sort() on different lines, then 2 sorted() on one line

davilla: anyway you have options

GooMooc: I m learning coding, first i determinate what it needed to obtain a result and after that i ve to search command that can giveme the result... i m awake that there a lot of possibility to obtain the same result but i cant explore all of them in once.... but ok for min and max, ive use it in a precedent puzzle, now i sort.... ty all for your answer.

davilla: great, glad we could help

GooMooc: yes, ty for your help... updating my python will permit me to reproduce some step of CGcode

Queuebee: Hey, I just accidentally got 100% on something with a wrong solution (ran tests to check and found that in 10% -- depending on parameters of the cases my code failed)

Queuebee: Where can I send feedback?

Queuebee: it's this puzzle https://www.codingame.com/ide/puzzle/paper-folding-curve

Default avatar.png LifeSword: Hello. Can anyone tell me how to view the debug output? I'm using C+=, and I'm output to cerr, but I can't see that output anywhere in the editor.

jacek: its left down corner, where you can see the standard output as well

jacek: Queuebee you can tell it on forum

Default avatar.png LifeSword: I was doing a clash, and I all could see in the bottom left was the cout output.

jacek: are you sure the cerr line was invoked then?

jacek: and had endl at the end

Yassine_EL_HOSNI: hello everyone, i have a question, to accept a contribution, do i have to solve it in all languages ??

AntiSquid: no

jacek: only english and french

Yassine_EL_HOSNI: lol

Yassine_EL_HOSNI: i misread one of the check boxes

jacek: oO

Default avatar.png idfkJustMe: guys can u help to find slow start ?

BlaiseEbuth: "find slow start" ?

Default avatar.png idfkJustMe: yes

BlaiseEbuth: Don't understand...

Default avatar.png idfkJustMe: ok, try to explain i am new in this platform so i dont no how to use it. someone say to go "Slow start" but cant find ))

BlaiseEbuth: There's no such section. So it's normal you don't find.

Default avatar.png idfkJustMe: ok thx , is this platform for beginners ?

Default avatar.png idfkJustMe: cuz its so hard for me to do smth.

BlaiseEbuth: Depend of what you mean by beginner.

BlaiseEbuth: If you have some basis so yes.

Default avatar.png idfkJustMe: i know basics

BlaiseEbuth: So it's ok.

Default avatar.png idfkJustMe: ok thx a lot good luck.

BlaiseEbuth: You can start here: https://www.codingame.com/training/easy

Default avatar.png idfkJustMe: yep.

MasterCoderxD: heyguys

MasterCoderxD: help

MasterCoderxD: My Java code to check if a string is a palindrome or not isn't working

MasterCoderxD: I just started Java, I am more of a python (and C++) guy

MasterCoderxD: My code;

MasterCoderxD: http://chat.codingame.com/pastebin/de34261c-5b24-4cd8-b52e-be3595864aba

MasterCoderxD: The issue is with the s==k part. When I print k, I see that k equals s, but still it prints false.

MasterCoderxD: I think it might me some whitespace issue...

DaNinja: s1.equals(s2)

Default avatar.png gorsat: I'm in the same boat as a java n00b. Kinda hate it. Anyway, you need to do s.equals(k) rather than s==k (which checks to see if they refer to the same string object)

MasterCoderxD: ok thnx

MasterCoderxD: gorsat use equals() method like DaNinja says

Default avatar.png gorsat: not sure what the red text is. I obviously was not asking a question.

orangesnowfox: gorsat: red text is "you were mentioned" (via someone typing your username)

Default avatar.png gorsat: Thanks orangesnowfox. :slight_smile:

planck6: how are you simge

planck6: cakir

Default avatar.png moyindavid: https://www.codingame.com/clashofcode/clash/1662435f2acfeb76009c6253ffdce447d0dfec7 c ++ reverse

jacek: ++c

Default avatar.png moyindavid: any lang reverse https://www.codingame.com/clashofcode/clash/1662438cf7e7d0c7745d3ae0cd6d872f13777cf

GooMooc: python3.8 installed and finally running with my ide, now the issue is the same in local and CG ide... Problem Solved.

GooMooc: if xo <= center and yo <= center :

       sfp.sort(key=lambda x : (x[0], x[1]))

GooMooc: is the correct syntax in 3.8

jacek: :tada:

Kellthazar: Urgh, this conway sequence puzzle...

Kellthazar: My code can't pass test 4 and 6...

Default avatar.png Revenantus: Is it possible to play Clash of Code without entering the 'shortest code' challenges? I dislike those with the language variations.

jacek: only in private clash

orangesnowfox: ... Why does my search race bot keep failing 1-2 (random) test cases?

jacek: timeout?

orangesnowfox: I mean, probably, but I also have it set to use 25 out of 50 ms ._.

jacek: in optimizations there is lower margin of error. for 50ms try lower time to 35ms or less

Default avatar.png Revenantus: Thanks, Jacek. I'll just not worry too much about squeezing all the whitespace out of my solutions during those, then!

jacek: oh hmm

jacek: because its rust :v

orangesnowfox: Nah, this is a copy of my CSB bot, it's C++

orangesnowfox: jacek: It also randomly times out when I do the web ide tests... Despite, well having literally no reason to

jacek: segfaults then?

orangesnowfox: Wouldn't it give a message about segfaults?

jacek: dunno

jacek: what message do you get now

orangesnowfox: "Timeout"

jacek: oO

BlaiseEbuth: No stack trace in case of crash in C++ on CG.

orangesnowfox: ah, right, because O3

jacek: too much ozone

BlaiseEbuth: Numa numa yeah...

orangesnowfox: ... I swapped out my clock impl and everything was fine

orangesnowfox: note to self `ctime`'s clock is *not* what I want to be using, apparently

orangesnowfox: That sounds highly unlikely, but, whatever

jacek: use chrono [solved]

orangesnowfox: yeah, basically.

MACKEYTH: Is there any word on when the 2020 Fall contest is going to make its way into puzzles and competition?

jacek: it is bot programming

jacek: https://www.codingame.com/multiplayer/bot-programming/fall-challenge-2020

MACKEYTH: Oh, neat! Thx!

djsapsan: hi guys, how to add friend? it doesnt work

jacek: just follow someone and let that someone follow you

Default avatar.png Sepehr.khashei: dirin dirin mord

Default avatar.png Sepehr.khashei: komak

Default avatar.png Sepehr.khashei: ...

Default avatar.png Sepehr.khashei: abu-alhasani

orangesnowfox: me in search race: "haha, I have saved 10 moves by... Making my code faster"

MadKnight: a pretty normal thing.... what do u mean it sounds weird?

KiwiTae: truue

jacek: eeyup*

C_breeze: please can someone approve my puzzle? https://www.codingame.com/contribute/view/5911d10658339fb867510c109b982a7fc304?comment=42321

Default avatar.png JBM: why yours before the oters?

C_breeze: sorry to bother

AntiSquid: ya we have an unwritten rule here to approve the ones linked in chat last, sorry

MadKnight: ..... unless it's like 1 confirm away and is running out of time

AntiSquid: then we frown it away

MadKnight: but is it their fault ?

C_breeze: didn't realise, it has a day left to validate so can wait I guess

AntiSquid: always

AntiSquid: ah well is it clash ?

C_breeze: not clash

Kapanion: https://www.codingame.com/clashofcode/clash/166270818a0281463438ef32100ff0698189b12

AntiSquid: oh it's not clash, here's a suggestion then, because i hate clash: make it work in progress and then make it live again since that way you avoid it being taken down, but in your case it won't matter you don't have any approvers

AntiSquid: weird you seem to have good comments, no approvals though

AntiSquid: ioane_kapanadze use #clash channel to post those links please

Kapanion: ok

jacek: hmm why would guy put minimax tag for atari go but not mcts

Default avatar.png Kaquino: hey guys i have a bug

Default avatar.png Kaquino: i can’t concat on php

AntiSquid: https://www.w3schools.in/php/operators/concatenation-operators/

AntiSquid: lol .in duckduckgo result @_@

Default avatar.png Sekseos: Can you still edit past clash game code that you didnt get a chance to finish?

jacek: no

Yassine_EL_HOSNI: hello everyone, how difficult is the Atari Go bot programming??

Default avatar.png normrm2191: it looked like it would take a while so I honestly skipped it

therealbeef: Do the optimization games have an issue with timeouts? I'm getting random validator failures on Search Race unless I dial down my timeout to 30ms

Default avatar.png normrm2191: yep, thats how they determine if ur code is optimized enough

therealbeef: I mean, for SR you get 50ms per turn but if i stay close to that 50ms I get random fails (but never in IDE tests)

therealbeef: That kind of think never happened to me in multis

orangesnowfox: therealbeef: yeah, optimization puzzles are more strict with their timers

AntiSquid: therealbeef i swear there's something wrong with search race validators or the entire timeout thing, i get very random timeouts, always on different validator

AntiSquid: and it's just that opti puzzle that has this issue for me

orangesnowfox: AntiSquid: same though

AntiSquid: think there's a bug in there, 100% there is mean

orangesnowfox: Well, actually, it's all

AntiSquid: because i played all test cases multiple times and checked them carefully, i don't timeout in any way and also i complete all tracks in a good time

AntiSquid: i am really annoyed actually about it

orangesnowfox: AntiSquid: Yeah, I normally have to submit at least twice

orangesnowfox: and I only use 36 ms

AntiSquid: it times out at 42 and 30 ms too . that doesn't make any sense

orangesnowfox: yeah... It times out *less* at 30

orangesnowfox: but... "sighs"

AntiSquid: i get 98% most of the time and as i said, different validator each time.

orangesnowfox: mhm

orangesnowfox: yeah, same

orangesnowfox: AntiSquid: What do you use for time?

AntiSquid: i think i tried both high precision and steady clock, let me check @_@

AntiSquid: my if-else JS solution doesn't timeout, but that's not a surprise i guess ?

orangesnowfox: Ah, so std::chrono, well, there's that suggestion gone :v

AntiSquid: steady clock resubbed just now and ... waiting ... oh 76% with 40 ms timeout hahaha

orangesnowfox: hahaha... Tuning is a pain

therealbeef: 35ms is ok for me, but 40ms i always get 1 or 2. at 48 i get between 1 and 5, never 0

orangesnowfox: Same on 2048 tho

AntiSquid: i have no timeout issues on that one but i use fixed depth

AntiSquid: no time check

orangesnowfox: multiple answers per turn, still use timer

orangesnowfox: ... I'm trying to decide how much time I want to devote to random sims, how much to depth, and how much to ga *sigh*

AntiSquid: it's a really weird one imo (2048) some cases lower depth yields better score

AntiSquid: also keeping too many branches can decrease score too :D

orangesnowfox: hahaha... Yeah...

orangesnowfox: I use 0 game specific knowledge in my 2048, it's a bit of a pain to tune

Rag: is there a way to only queue for specific clash of code types?

AntiSquid: i tried "monotonous sequence" check and the "snake sequence" check both can be negative

AntiSquid: see if any of this helps you orangesnowfox https://stackoverflow.com/questions/22342854/what-is-the-optimal-algorithm-for-the-game-2048

orangesnowfox: AntiSquid: But you have access to the rng...

AntiSquid: read through it

AntiSquid: you can ignore the rng part

AntiSquid: it has some good heuristics, at least good source of inspiration

AntiSquid: did something along those lines, but not entirely

orangesnowfox: I'll look at it then

AntiSquid: also my code isn't optimized

AntiSquid: well you're at 12mil, i am currently at 3.2 mil, probably can reach 4.5 mil with a bit of optimization, but need to do massive changes for better performance

AntiSquid: did you use bitmasks orangesnowfox ?

orangesnowfox: bitmasks for *what* AntiSquif

orangesnowfox: *AntiSquid

AntiSquid: speed

orangesnowfox: No... "for which thing"

AntiSquid: for the table itself

orangesnowfox: No... I store the power of 2 instead of the value itself though

orangesnowfox: I could maybe store a side-table of zeros, but I'm not sure how I'd use that

AntiSquid: right, how do you sort best branches without heuristics ?

orangesnowfox: "score"

AntiSquid: that's still a heuristic :P

orangesnowfox: Technically a heuristic

orangesnowfox: But not a game specific one

orangesnowfox: :p

AntiSquid: well here the score is gamespecific :P what values you combine

orangesnowfox: hush you :p

AntiSquid: score alone gives me a bad result

orangesnowfox: Code must be slow :wink:

AntiSquid: i get worse result

orangesnowfox: Yeah, but also, how many moves per turn?

orangesnowfox: Also, do you use move 1 fully? That helped a lot (saved about 30 turns worth of moves for me)

orangesnowfox: My typical problem is running out of time... I have depth = 30

AntiSquid: well if i use the just score heuristic then there's less time involved selecting best ones

AntiSquid: that would be the only diff in the code so theoretically i should get better score if that's the better heuristic

orangesnowfox: Yeah... And, it isn't lol

AntiSquid: true it is not switched back to my own heuristic, did few submits, gained +200k points from my previous best +1 rank, marginal improvement but shows which heuristic is better

AntiSquid: orangesnowfox are you going to try the heuristics i linked see if you climb higher?

orangesnowfox: Probably lol

AntiSquid: the "Monotonicity" 100% helps

orangesnowfox: Yeah... I'm not sure how to quantify that lol

AntiSquid: you pick any corner you prefer and check from there how many numbers are sequentially smaller

orangesnowfox: hmm, but what about transpositions?

orangesnowfox: and "sequentially smaller" in which direction?

Rival819: im taking a free C# class on Codecademy.....then coming back here hehe will understand a lot better its already awespme what I have achieved learning

Rival819: im bronze 33 in CSB

Rival819: from no knowlede

Rival819: knowledge, so now will be back in force

Rival819: hehe

Rival819: im learning concatenations and interpolations of text atm

Rival819: :)

Rival819: well almost done that part actually

orangesnowfox: AntiSquid... I suddenly am really tempted to re-implement my code in c++ to see if it goes faster

orangesnowfox: But I'm *tired* right now, so I have to convince myself I really don't want to

Default avatar.png DragonNinjaA: what are puts

GilbertGaming27: Gilbet

ANONYMOUS42: i would like to join your cult

Default avatar.png evaporatetranslator: hi

ANONYMOUS42: hi

Nuggs: GIBLERT