Chat:World/2020-07-14

Karl-EerikOtstavel: Hey

TrashCoder22: anyone on?

TrashCoder22: i need help for sre

jacek: good morning

LeBaoHoang8A4: good

Zachary_0722: hey

TheSpiffiest: Just realized it's missing "R" language

TheSpiffiest: I have to learn that for my new job

TheSpiffiest: They are doing 3d graphics with it.

jacek: oO

Astrobytes: 3D data visualisation I hope. Either that they're insane :P

TheSpiffiest: Yep. 3JS and giant data sets

jacek: JS? :scream:

TheSpiffiest: I didn't write it, but I'm tasked with rewriting it.

YCS-Venom: Hello world

ntsd: Hi Community

m.amin: hello

ntsd: I wanna practice English can anyone voice chat with me?

Nghia64582: no

AntiSquid: hi ntsd long time no see

MOSTAFA: Hello

MOSTAFA: How to get other player Code to learn more

MOSTAFA: i am playing

Power of Thor - Episode 1

MOSTAFA: i didn't like my result

MOSTAFA: so i would like like to get other players so i can improve

ntsd: You need to pass it first

MOSTAFA: i did

MOSTAFA: i got 75 %

ntsd: then you'll see other code in same language

ntsd: you need 100%

MOSTAFA: from where

Astrobytes: Have you checked out the 'Hints' button on the left MOSTAFA?

MOSTAFA: no

MOSTAFA: what is Hints

Astrobytes: Press it and see ;)

MOSTAFA: oh i didn't know that Directions work as a string here and i can assign it to a other sting

MOSTAFA: thx man

Astrobytes: No problem

NewfieBullet: What time does the rank reset happen for clash of code?

DaBestest: Right now I'm playing King George's three songs from Hamilton on repeat while I try to finish up a huge project on Scratch I'm making with a bunch of friends for a contest...

jacek: good for you

MSmits: jacek, what happened on Othello?

MSmits: did tric trac and robo submit a better version or ?

Astrobytes: Submitted before coffee by the looks of things :P

MSmits: I fixed up my AVX last night

MSmits: pretty sizable boost to performance

MSmits: not much difference in rating I think though

Astrobytes: Is that your current version?

MSmits: yeah

MSmits: still very simple eval

Astrobytes: Hm yeah, doesn't seem to have improved your position much. Wait til your eval is better, then it might count for more

Astrobytes: Your eval is just mobility + corners?

MSmits: mostly yes

MSmits: now using up to 16 million nodes in turn 1 and then over a million nodes each turn

MSmits: my node pool is 22 million, sooo not much tree-reuse

Washier: :open_mouth:

Astrobytes: Nice

MSmits: it's kinda funny, Robo evals 3k nodes per turn and I eval 1 million. His bot beats mine by far

MSmits: shows how important eval is huh

Astrobytes: Yeah, for sure

MSmits: but i am pretty sure my search type is just fine. No need for minimax

MSmits: not even done fitting the 4 parameters I do ha e

MSmits: have

Astrobytes: Did you get frontiers working?

MSmits: i havent tried again

Astrobytes: It's a strange parameter

MSmits: if its only useful in early game then I don't want to put too much effort into it

Astrobytes: Seems useful throughout, you see a lot of games where it's mostly quiet moves then a big BAM! at the end

MSmits: yeah but how do you know it's frontier ?

Washier: was wondering earlier, how do you think the top othello bots will fare against the greats like Logistello, given the time limit?

MSmits: the reason it is not a fair comparison is that the evals here use hardcoded ntuple-tables and similar. Logistello has a far larger codespace for this

Astrobytes: minimising your frontier discs means a quieter move

MSmits: so we use smaller ntuples and such

Washier: ok

MSmits: not me, but I guess at some point I might

MSmits: what do you mean quieter?

Astrobytes: flipping less frontier discs

Astrobytes: And creating less

MSmits: but isnt that the same as giving your opponent less mobility?

MSmits: I think it might interfere with the mobility eval

Astrobytes: yes, they're related

MSmits: right

MSmits: thats the problem I think

Astrobytes: It's also known as potential mobility

Astrobytes: Finding the balance with these two is tricky

MSmits: yeah I guess so

Astrobytes: For me at least

Astrobytes: Just started working on it again after a break

MSmits: I should probably also look at other eval features like disc stability

Astrobytes: Yeah, I need to do that also. Been lazy

Astrobytes: I conditionally evaluate the C and X squares

Astrobytes: By corner occupancy that is

Washier: i alos tried that, didnt help me.

Washier: but, fark, im still struggling with itterative deepening, so dont listen to me

MSmits: C and X I did make work

MSmits: it's a bit tricky to make sure you dont punish yourself when you have a corner and also the squares next to it

MSmits: X squares are only bad when the corner is still free

Astrobytes: Yeah exactly that

inoryy: why won't you guys write down your progress in blog posts and the likes?

eulerscheZahl: sounds easier to follow than chat activity spread over days if not weeks

Astrobytes: I don't particularly feel worthy of blogging tbh, Smits definitely should

DomiKo: i guess inorry's idea is really cool

inoryy: it would also immensly help with potential career switches down the line

MSmits: it's good advice inoryy, thanks

Astrobytes: Fair point

MSmits: Astrobytes, your bot doesn't have to be the best bot to be worthy of blogging. I think many good blogs just follow a path of improvement. Doesn't matter if it leaves out the very last bit that gets you from rank 10 to rank 3 or something

HeaDTy: https://www.codingame.com/clashofcode/clash/1293621fdd42d4cf6c3e1ec35addaad8aae04ef

eulerscheZahl: no HeaDTy

HeaDTy: Ok sry

DomiKo: if not "shortest mode" i can try

Astrobytes: Yes, I guess that's true MSmits

inoryy: Astrobytes everyone thinks this way; the reality is that a top100 bot with a detailed blog post is more valuable than a #1 with no info

MSmits: very true

inoryy: both for the community and for the blogger

Astrobytes: You do have a point there, yep

MSmits: I'm trying to find a good function to calculate stable disks now

MSmits: i think that could be important

Astrobytes: It seemed like it might be a little 'involved', hence me not doing it yet :P

MSmits: it should be similar to generating and resolving moves

Astrobytes: Yeah, it is, with a few additional bits

Astrobytes: 'bits'

MSmits: yep

inoryy: Astrobytes a real-world example -- Magus CSB blog post has been an invaluable resource for everyone from #1 to #200+ while his own bot basically floats from top50 to top100 based on who submit spams

MSmits: you've convinced everyone inoryy :)

Washier: not sure if it can work here, but on kaggle anybody can share code with the community in the form of a working entry they can submit. as long as people dont share insanely good scripts, everybody wins, but not all people would agree.

Astrobytes: Hm, that's a good example actually. I'll see what I can do.

MSmits: it's a delicate balance between sharing and competition Washier

MSmits: hard to say where the exact line is

eulerscheZahl: also depends on the game and approach to it. "wrote some if-else bot" rarely makes a good post IMO

Washier: yup. and people have fought over it a lot.

MSmits: unless it's an introduction to the game and is followed up by more nuanced evaluations

eulerscheZahl: the CSB bible rebuilds the engine and makes a good blog article that way. i didn't even care about the GA part of it

Astrobytes: Building blocks, yes

eulerscheZahl: then again with public referee the CSB blog post would be less interesting to me. it just tells us what CG missed to state

MSmits: there are 3 types of things that would be interesting to blog it seems to me

MSmits: 1) Strategic stuff pertaining to the game. How do you win? Regardless of the way it's coded

MSmits: 2) Explaining how search types work

eulerscheZahl: 3) your lunch that explains social media

MSmits: 3) Implementation of algorithms

eulerscheZahl: oh

MSmits: or that

MSmits: why would my lunch explain social media

Washier: haha

Washier: i would love to read 2) MSmits

MSmits: it's funny even though i dont get it :P

MSmits: I think I prefer to write specific guides as opposed to writing a blog while I am still figuring things out

dbdr: sounds like a perfectionist mind :0

MSmits: when i am still figuring it out i prefer to chat about it so i get advice

MSmits: well that too

inoryy: you could write down intermediate results

MSmits: thats a good idea

MSmits: thats what people do in post mortems sometimes

jacek: writing blogs could also work as rubber duck debugging

jacek: i wonder how many bugs i would discover if i were to review my code and logic

jacek: and yes, i posted inferior bot, im looking between weight number balance vs their usefulness

Astrobytes: Coming to Wood 1 then DomiKo? :)

eulerscheZahl: sometimes i write my post mortem during the contest. it helps me to reflect what i'm doing, found a few bugs that way

jacek: so its in mortem then

Astrobytes: antemortem

MSmits: nice one eulerscheZahl. Also might be a way to battle losing interest.

MSmits: a bit of variation from staring at code and getting little progress

kovi: yes, but there is a different point of view: it you can't win in, at least you can teach it

Washier: your progress looks pretty good from here.

Astrobytes: DomiKo is absolutely wrecking Wood 2

MSmits: good

kovi: but as msmits wrote sometimes, summarizing what is done/known can help reveal new ideas

MSmits: btw kovi, did you come here for a spefic multi? I don't see you active much outside contests

MSmits: specific that is

eulerscheZahl: kovi knows something we don't know yet

Astrobytes: :)

MSmits: i imagine he knows many things

Washier: sorry MSmits, i meant excellent. (goed in afrikaans is meer komplimenterend)

jacek: can jet fuel melt steel beams?

MSmits: no I meant good that dominik is coming into wood 1 :P

DomiKo: yap :D

Washier: oh crap

Astrobytes: gg dude, great submit

MSmits: domiko sry

DomiKo: but no worries

DomiKo: not right now ;P

DomiKo: WAshier 1:0 for me

Washier: haha. move along, nothing to see here

DomiKo: haha

Washier: smashing me

Washier: oof

Washier: taking al 4 corners, both times

Washier: me prediction, 7-sh

Washier: *7-ish

DomiKo: then it will end i will push better version

DomiKo: when*

Astrobytes: god help us all :)

Washier: at the end

Washier: of submit

vfpx07: Do Python users have an advantage when doing least characters possible?

DomiKo: https://www.codingame.com/replay/476998805

DomiKo: okey

MSmits: I need more c++ to test against, keep going Domiko :)

DomiKo: MSmits

DomiKo: bug?

eulerscheZahl: depends over whom vfpx07

DomiKo: how i won

DomiKo: haha

Washier: wd

MSmits: lemme see that

eulerscheZahl: python wins in shortest vs Java but most likely loses vs bash

MSmits: why bug?

DomiKo: i didn't think that could happend

MSmits: bot detected loss at frame 47

DomiKo: :(

MSmits: I have had wins against people 6 rating up... there's still such a thing as random luck and rps effects :)

DomiKo: 2:0 for me

DomiKo: that's not luck!

DomiKo: haha

MSmits: rps maybe then

Washier: happened to me before believe it or not. once beat LeRenard. freak game

DomiKo: what is rps?

DomiKo: wow

MSmits: rock paper scissors

DomiKo: ohhh

MSmits: you're rock and I'm scissors

MSmits: and everyone else is paper. Well except top 4 :P

DomiKo: haha

DomiKo: gotcha

MSmits: could be that if i change one parameter by a tiny bit, the rock is gone

MSmits: it's like that in many games

MSmits: thats why i always test against multiple players

MSmits: yep, you rock

eulerscheZahl: that's why i hate board games with a static setup

eulerscheZahl: maps need randomness in geenration

MSmits: ye ye we got it, Onitama is great :P

eulerscheZahl: wasn't planning to say or imply that

MSmits: it's still true

Astrobytes: Alright, so we make a version of othello where you randomly swap sides at random intervals in the game just for euler :P

Washier: final push for 7. cmon

DomiKo: RNG lets go!

MSmits: what happened with tomatoes btw

Washier: or opening books

eulerscheZahl: no, random for map generation. not during the game

MSmits: pretty sure his rating was better

Astrobytes: It was

DomiKo: http://cgstats.magusgeek.com/app/multi-othello-1/DomiKo

Astrobytes: Alright, so we have areas of squares in both colours randomly distributed at the start of the game

DomiKo: i have no chance with top7

DomiKo: WOW

DomiKo: Washier

DomiKo: how

Washier: touche

DomiKo: you god

Washier: wut

DomiKo: i end up at 7...

DomiKo: how you guessed that

Washier: wish i had less intuition and more brains

DomiKo: hahah

Washier: its a toit 7 tho. will change

DomiKo: yea

nguyenletienbao: input $8.90

nguyenletienbao: Expected output

nguyenletienbao: 8 3 1 1 0

nguyenletienbao: how to handle this?

fvla: integer division

WINWINWIN: Number of 1 dollar, number of 25 cent, number of 10 cent, number of 5 cent, number of 1 cent

Washier: guess 8 $ bills, 3 quarters etc

nguyenletienbao: can u explain more clearly

fvla: is this reverse engineering

WINWINWIN: Consider a amount that you want to convert into change.

WINWINWIN: The denominations you have are: 1dollar, 25 cent, 10 cent, 5 cent, 1 cent

Jackie_Truong: :grinning:

WINWINWIN: print how many of each you need.

nguyenletienbao: got it:grinning:

AnirudhPatel: There is a puzzle marked with a neural network. has anyone here done it?

sparky: which puzzle?

JulesCesarTCHUISSEU: hello

NinjaCarnivoreAndHisDog_c7bf: hello guys

WINWINWIN: Think he means csb

DaBestest: who's watched Hamilton on Disney+?

TheSpiffiest: Ha. I keep going off on tangets in code clash. I can solve it in a minute, but then I think wait... there's an algorithm that will do this in one go that's more interesting.

TheSpiffiest: regular expressions on the brain

nachonitz: Hello World!