Chat:World/2021-01-01

From CG community
Revision as of 11:58, 15 June 2021 by Chat Log (talk | contribs) (Created page with "File:Default_avatar.png Jay3401: Hello File:Default_avatar.png soutenu: hi! File:Default_avatar.png ChampionCoder: Happy new year, guys! File:Default_avatar.p...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Default avatar.png Jay3401: Hello

Default avatar.png soutenu: hi!

Default avatar.png ChampionCoder: Happy new year, guys!

Default avatar.png rusty_developer: hi my friends

Default avatar.png rusty_developer: i am trying to learn a rust

jacek: :scream:

jacek: good morning

eulerscheZahl: moin

jrke: happy new decade everyone

eulerscheZahl: didn't the decade start last year?

jrke: nope this year i think

eulerscheZahl: mine goes from 2020 to 2029. like every good programmer I always start counting at 0

pipebits: True

VizGhar: Which decade this year belongs to: 302x? When new millenia started? 2001 or 2000? :)

NguyenDinhDuong: 2000 :smiley:

eulerscheZahl: download is still working \o/ https://get.adobe.com/flashplayer/

jrke: euler did you got you hacktoberfest t-shirt?

eulerscheZahl: of course not. i don't participate in such events

jacek: flash? what year is it? 2020?

eulerscheZahl: i've downloaded the player and some old flash games just to be sure they are still working

jacek: i think i got this 2048 learning working \o/

eulerscheZahl: nice

eulerscheZahl: what's the highest tile you can reach?

jacek: dunno yet, right now im looking at total score

jacek: but with bruteforce and snake i can get around 600k

jacek: is there 131k there?

eulerscheZahl: yes, but not reachable for all testcases

jacek: so 18 states per tile :/

Default avatar.png saikoaizen: 77889977445555662233115547885544552266855

Default avatar.png saikoaizen: find the number that occurs odd number of times

jacek: odd question

Default avatar.png saikoaizen: (*dumb question)

Default avatar.png saikoaizen: lmao, don't count. write a program

eulerscheZahl: find the "number" do you mean digit? or any substring?

Default avatar.png Angecide: hi how does others output their text on their profile when watching replays?

eulerscheZahl: while game?

eulerscheZahl: usually: command message

eulerscheZahl: just with a space between

jacek: what game? some games dont support that

Default avatar.png Angecide: some have funny emojis depending on if they are winning or losing

Default avatar.png Angecide: I guess depending on what their mcts is evalutating

jacek: UTTT?

Default avatar.png Angecide: ye

Default avatar.png Angecide: but i have seen this on other games as well

eulerscheZahl: you can even see how they print it: just fight them in the IDE and you see their output

jacek: you can print the emoji, its just unicode char

Default avatar.png Angecide: are they doing std::cerr?

eulerscheZahl: stdout y x message

eulerscheZahl: 7 3 i'm winning

Default avatar.png Angecide: ohhh


Default avatar.png Angecide: I thought std::cout was reserved for legal moves only

eulerscheZahl: was about to say something like "reading is overrated" but the game doesn't even tell about that feature in the statement :D

Default avatar.png Angecide: oh yep, it's definitely working

Default avatar.png Angecide: ye, I searched everywhere

eulerscheZahl: at some point it's just experience as most games on CG work like that

Default avatar.png Angecide: I see

jacek: see? nobody reads statements

eulerscheZahl: it takes some time to write those :(

eulerscheZahl: except for the CG sponsored game

Illedan: Good morning. Happy new year :)

Default avatar.png [Tony-Stark]: happy new year

MSmits: happy newyear

jacek: newy ear?

ZarthaxX: happy new year :D

Default avatar.png ChampionCoder: Happy new year, guys!

MSmits: doubt many people gained a new ear, maybe some lost one last night though

MSmits: we had a ban on fireworks this year because of covid, but not many people gave a :poop:

MSmits: still seemed like a warzone

MSmits: there were 30% less incidents this year. That's all the ban did. Emergency workers still had to sew on fingers and whatnot

MSmits: one of the few things i hate about my country... fireworks

MSmits: /end rant

tutubalin: fireworks spread viruses

MSmits: doubt that, but they put more pressure on hospitals that are already at the breaking point

MSmits: 1 million delayed surgeries and such

MSmits: because people are idiots

MSmits: not many people talk about that. It's fine and all to keep showing the deathcount of covid, but they dont talk much about people needing surgery and not being able to get it

karliso: MSmits Do you have hangover?

MSmits: lol

MSmits: no i don't drink karliso

MSmits: happy newyear :)

karliso: Thank you! :)

Default avatar.png ChampionCoder: Lol no crackers in my country

Default avatar.png ChampionCoder: Nobody burst them since Diwali

eulerscheZahl: one of the few things i hate about my country... fireworks

i share that hate

Astrobytes: +1 from me

Astrobytes: and happy new year everyone

Default avatar.png zhoubou: Happy new year :)

eulerscheZahl: always some who want to cause noise during the night. even during lockdown when you have to be at home at 9pm. but much more quiet than the years before

eulerscheZahl: my cat loved it

MSmits: my dog was a bit better this year also

Default avatar.png zhoubou: ^ Same

Default avatar.png zhoubou: He usually barks quite a lot. This time he didn't bark at all.

Astrobytes: Yeah, at least our pets had a good time this year :D

eulerscheZahl: also me working from home is great from her point of view

Astrobytes: Indeed. I spend most days with one on my shoulder and one in my lap.

jacek: 24/7 can opener

eulerscheZahl: and living pillow

Astrobytes: Yeah! In winter it's like having 2 live teddy bears. Vicious teddy bears.

Astrobytes: Always wanting to sleep on or right next to you

eulerscheZahl: or on my chair

eulerscheZahl: so i sit on the trash can

Astrobytes: Yes, when you want to sit on it. Hahaha

Astrobytes: The sacrifices we make

Kitkat260: gm

jacek: https://www.youtube.com/watch?v=n3YT7ksgGJk

TBali: HNY! So ,Astrobytes, you have 2 cats and only one of it is your profile pic? Doesn't that make the other one envy?

TBali: :)

Default avatar.png sHlOk: Any hacker out here?

Stilgart: ask AutomatonNN

AutomatonNN: what do you mean by the contest ?

TBali: Who asks? FBI?

eulerscheZahl: no hackers. deleted one from the forum 1h ago

Astrobytes: TBali: Yes, that's the case. For years I used the other cat as a profile pic, switched it around a while ago. Hopefully she doesn't mind too much ;)

eulerscheZahl: "how to hack latest game", links to a website for buying a dish washer

Default avatar.png sHlOk: I am an Ethical Hacker. So wanted to know if there are som here

jacek: oO

Stilgart: not sure if AutomatonNN is ethical

eulerscheZahl: there's a guitar as well: https://static.codingame.com/servlet/fileservlet?id=23052489191875

AutomatonNN: hey guys hey guys, could you be honest, and that makes sense to me in a linux and contests and the b

Stilgart: and a cormen

eulerscheZahl: and something with "alco" on the folder. studying for the barkeeper exam?

TBali: and the CLRS book... how approriate profile pic :-)

Default avatar.png sHlOk: Any OSCP here?

eulerscheZahl: oh, it says "algorithms", a bit hard to read with that low resolution :D

Default avatar.png zhoubou: Lol

Astrobytes: It seldom leaves my desk. Along with Stroustrup and a few others

TBali: I have the same edition. The leaf end reveals it

TBali: Started to read it last March, still 140 pages to go... :-)

Astrobytes: It's not a quick read :D

eulerscheZahl: and then you have to start again and see how much you've forgotton

Astrobytes: I'm fairly certain I look something up in the introduction everytime I look up an algo lol

TBali: Well for few 100 pages I was slow because I coded all pseudoalgorithms encountered. Then I gave that up, but was still slow following the proofs.

TBali: In the end I sometimes see myself just skim-reading the proof

TBali: Would not be enough to pass an exam with A if I were on a CS class

Astrobytes: Following the proof isn't 100% necessary if you just want to know about the algo, but it can be useful

TBali: The important is not the proof but to understand how and why it works

TBali: For that I found the Princeton video courses on Coursera exteremely good

TBali: Less mathematically rigurous but still going deep

TBali: Had to do the 10 assignments in Java made me learn it a bit more as well

Astrobytes: I did a few courses irl some time ago, I admit I could probably do with a refresher

TBali: Princeton topics highly overlapping with the CLRS book, of course

eulerscheZahl: Java? now you are ready for the codingame SDK

Astrobytes: some time ago: 2007

Astrobytes: I also did a java course. Hated every minute of it.

proogrmr: :no_mouth:

TBali: SDK: LOL - I would need a great game idea first

eulerscheZahl: surprise us :)

Astrobytes: The exam involved handwriting a fair bit of Java. My hand was aching after.

TBali: You can practice with cOc GOLF :-)

struct: just port amazons

TBali: Possibly the longest codes ever

TBali: But maybe C# is even longer

Stilgart: eulerscheZahl: are you telling that the SDK is in Java ? :p

TBali: Is it?

Stilgart: wasn't it bad written C++ ?

eulerscheZahl: C# has more syntactic sugar to write short code than Java does

eulerscheZahl: working on "Focus" yet struct?

struct: no, im still doing stc

Astrobytes: Amazons will be ready in 2022

eulerscheZahl: when someone else, completely new to the CG community and not knowing about struct's plans, decides to just do it

Astrobytes: :D

TBali: Okay I admit being noob. What is Amazons?

eulerscheZahl: board game

eulerscheZahl: https://www.boardgamegeek.com/boardgame/789/focus i put this one up for discussion

eulerscheZahl: you have towers of stones. and go as many fields as the tower is high. building new, higher towers in that process

eulerscheZahl: first unable to move loses

eulerscheZahl: if the tower exceeds a height of 5, stones from the bottom are removed.

eulerscheZahl: moving player gets the stones back for later spawns

eulerscheZahl: opponent tiles are out of game

TBali: looks interesting.... and suitable for a multi

eulerscheZahl: and 3D :P

TBali: 1981 Spiel des Jahres Winner

Astrobytes: Yes, that looks interesting

Astrobytes: Have you played it?

eulerscheZahl: no. but my dad has it in the basement

Astrobytes: Well, you know what to do

eulerscheZahl: a stream? :D

Astrobytes: lol

Westicles: somebody do a Nuggets Numbers stream

MSmits: wondev woman is also 3D

eulerscheZahl: and i'm surprised you aren't addicted to it yet

MSmits: yeah

MSmits: i guess i will be

MSmits: at some point

MSmits: it has crazy branching though

jacek: and fow :unamused:

MSmits: that is supposedly not much of a problem

MSmits: as you can just calculate the most likely position of the opponent and do the search based on that assumption

MSmits: thats what the pm's tell me anyway

Kitkat260: hi guys

MSmits: hi

Kitkat260: is there any teens here

Kitkat260: nvm

MSmits: I am a Fteen next month

MSmits: it's like a tween, but with forty

Kitkat260: are u a teacher

MSmits: yea

jacek: what gave it away?

Kitkat260: your proflie

jacek: :spy:

Westicles: Haha... and we rule the world

MSmits: jacek, you use simd sometimes right?

MSmits: does this function annoy the hell out of you too: https://software.intel.com/sites/landingpage/IntrinsicsGuide/#expand=2431,4963,2431,4963,4912,4912&cats=Set&text=_mm256_set_epi64x

MSmits: it's just... wrong

MSmits: I get bugs because of this and I say the bug is on intel's side

struct: what bug do you get?

MSmits: the e indices are reversed

MSmits: why would anyone do the largest index first

MSmits: makes no sense

jacek: i dont use simd or avx directly

struct: __m256i _mm256_setr_epi64x

jacek: never used _mm stuff

struct: setr

struct: should do the job right?=

MSmits: so the reversed version is actually the normal version?

struct: yes

struct: and no

MSmits: thats still lame :P

struct: reversed shows the correct order

MSmits: but thanks

MSmits: i will use that now

MSmits: jacek do you rely on autovectorization?

MSmits: not sure how well it works

jacek: i just hope it works

MSmits: well i guess it doesnt matter that much as it is not your bottleneck anyway

jacek: if not, then there's probably some 2x-3x speedup left for my bots

MSmits: it's evaluating the networks

MSmits: doubt it

jacek: the eval takes most time for me

MSmits: yeah so its not going to be 2x 3x

MSmits: unless you could use the avx stuff in eval

MSmits: and you arent

MSmits: i was more talking about the sim part

jacek: oh

MSmits: thats where i use it, when i can

jacek: if autovectorization doesnt work in my eval, then it would huge speedup

MSmits: might be good to test it, you can do that with godbolt and such cant you?

jacek: ever since i have my eval quite heavy, i dont care much about sim speed, aside its bitboarded

MSmits: makes sense

jacek: perhaps. i never really heavily analyzed the code and instructions

MSmits: i dont usually either, but thats because i dont rely on the compiler that much

MSmits: i micromanage too much :P

MSmits: I think i have a bot somewhere that uses goto

MSmits: so i improved since then

jacek: goto oO

MSmits: i think it was a fast way to get out of a nested loop

ZarthaxX: MSmits why did you say that instruction saved in reverse order?

ZarthaxX: pseudo code seems fine

struct: setr(0,1,2,3) = [0,1,2,3]

struct: set(0,1,2,3) = [3,2,1,0]

struct: but pseudo code is fine you are right

ZarthaxX: pseudo says e0 would be at lowest part

ZarthaxX: that thing u put there would be array in memory or what

struct: you can load an array

struct: load or loadu for unaligned

MSmits: ZarthaxX the code runs fine

MSmits: the fact is, you put the e3 part in the leftmost part of the function call

MSmits: so you're thinking you do set(0,1,2,3)

MSmits: but you actually do set (3,2,1,0)

struct: also msmits cant you load it as an array?

MSmits: these are constants basically

struct: set is a sequence as it says on intel site

struct: ah ok

MSmits: always the same, so i load them once

struct: ok then its fine

MSmits: just 256 bit masks

struct: this is for what?

MSmits: connect4 :)

MSmits: I'm avxing the sim

MSmits: horizontal, verical and diagonals all at once

struct: this library is good

jacek: :scream:

struct: https://github.com/vectorclass/version2

MSmits: seems i already had that favorited

jacek: plot twist: smits made it

MSmits: lol no

Kitkat260: what happened

MSmits: 2021

Kitkat260: ik

Kitkat260: where is jacek

ZarthaxX: MSmits still dont get your confusion

ZarthaxX: e3 in leftmost part for you would mean it should be highest number or lowest

ZarthaxX: maybe you dont know how a simd register looks like

MSmits: well it's like this

MSmits: in my mind the 256 divided into 64 bit parts works like this

Kitkat260: ummmm

MSmits: if i load it like this Set(0,1,2,3)

MSmits: the 0 comes in the low part, the 1 comes in the 2nd 64 bit part etc.

MSmits: but thats not what happens

ZarthaxX: ah

Kitkat260: how many girls r here

ZarthaxX: well register it just bits so the order is 3 2 1 0

ZarthaxX: that's why thee convention

MSmits: yeah, it's confusing

ZarthaxX: of the call

ZarthaxX: doesnt seem wrong

MSmits: i'm sure there is a good reason for it, but when i am using it, in the context of the rest of my code, it seems wrong

Astrobytes: fix it

MSmits: i did

MSmits: i added an R

MSmits: setr

Astrobytes: ezpz

MSmits: then used a #define

MSmits: so i dont see the r :P

Astrobytes: lmao

ZarthaxX: aight :P

ZarthaxX: connect4 eval withbitboard had shift in different ways

ZarthaxX: idk if you can do smth with avx

MSmits: already did, works fine

ZarthaxX: okey

ZarthaxX: why are you doing this btw

MSmits: why wouldnt it work?

MSmits: there's a connect4 contribution that will soon be approved i think

MSmits: 9x7

ZarthaxX: i dont remember if u had to do shift in different dirs

ZarthaxX: ah okey

MSmits: you do

MSmits: left right

MSmits: up down

MSmits: and diagonals

ZarthaxX: well that'sa what i meanty

ZarthaxX: cant do it all at once

MSmits: there's 2 sets

MSmits: you combine left with up, down with rigth etc.

ZarthaxX: aight

ZarthaxX: gud

MSmits: earlier I thought the creator was messing around with the contribution because i kept getting errors with input

MSmits: after a while I figured out i was pasting my bot into the Dots and boxes IDE

Astrobytes: ffs :rofl:

struct: lol

MSmits: not the first time this happened

struct: What happens if he increases the board size?

Astrobytes: it gets bigger

MSmits: if he increases it further, my avx thingy is screwed

Astrobytes: :P

MSmits: it's neatly 63 bit now

struct: I dont know how its possible, but Im atm reading some code

MSmits: 9x7 apparently isnt solved so shouldbe ok

struct: for AVX on puyo puyo

struct: and its 12*6

struct: so 72 bits

MSmits: i'm sure it's possible, it just would be very inefficient

MSmits: might still be faster than without it

MSmits: you would have to split the process in two halves

struct: First I need to understand the bitboard part

MSmits: there's many ways to do it

MSmits: mmh is there a faster way to get the highest set bit than this:

MSmits: int index = __builtin_clzl(col); index = 63 - index;

MSmits: i'm counting leading zeroes

struct: https://github.com/puyoai/puyoai/tree/master/src/core

MSmits: dont remember if you could get the highest set bit with a single instruction

struct: its so much code

MSmits: yeah]

jacek: ctzl?

MSmits: thats liwest set bit

MSmits: lowest

MSmits: for some reason i made my connect4 bitboard to have the lowest part of the bitboard at the high bits

MSmits: because it's input in that order i guess

MSmits: maybe i should change that

MSmits: wow i literally just reversed an iterator in the input reading part and the bot still works, now i can use ctz instead

struct: I thought there wasnt modulo operator for m256i but there is

struct: but its a sequence

struct: so should not matter much

struct: https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm256_irem_epi32&expand=3205

struct: its for 32bits though

MSmits: yay i am on the leaderboard. Just random moves and it takes the win when it can

jacek: i will suggest bigger board :imp:

MSmits: meanie

jacek: bigger than 256

darkhorse64: I have a mcts for C4 with solver. I even wrote an AVX2 win detection but it does not go faster than the x64 code. On my PC, it's nearly a 2x, 3M -> 5.5M rollouts/s

MSmits: nice one

MSmits: thats odd, why wouldnt avx2 be faster?

MSmits: also did you detect wins, or winning moves?

MSmits: it's different

MSmits: my function returns all moves that are winning

darkhorse64: I suggested an octagon based C4 to make sure it is not solved. It would have taken more than 64 bits. 9x7 is solved as a win for White but good luck to find it in 100 ms. A book will probably help. I am curious how trictrac MM will perform. I detect wins. Next step is detecting winning positions for smarter rollouts

MSmits: 9x7 is solved?

darkhorse64: sec

MSmits: i thought 6x7 was

darkhorse64: https://tromp.github.io/c4/c4.html

Kitkat260: hey

jacek: 9x7 its not far from 7x6

Kitkat260: yep

Kitkat260: it is

darkhorse64: 42 moves against 63, the search space is much larger

MSmits: it's like D&B, 1 size large is like a factor of 30 or something in computation time

Kitkat260: :sunglasses:

MSmits: so that site says 7x9 is just barely out of range?

jai_m: hey guys i have made project on face recognition, i would like you to check it out once it will help me to make it more better! please wait at the face-part as it will take time to load link: http://smile-pay.herokuapp.com/

Kitkat260: (╯°□°)╯︵ ┻━┻

MSmits: he will steal your face!!!

MSmits: :P

darkhorse64: re avx2 I have no explanation. I checked my code under windows, this is where my performance figures go from. On CG servers, I see no difference at all. I don't know why. I am doing the four in a row in all directions at the same time and I have no ifs. It should get better

MSmits: hmm might be it autovectorizes

MSmits: that would mean it does help, but you dont have to code it yourself

Kitkat260: umm what r talking about

MSmits: parallel calculation for bitboards in simulating boardgames

darkhorse64: my bad: I misread the table. It is not solved. If it autovectorizes, it's a very smart compiler

jacek: well MSmits has just got new title for his thesis

jai_m: no MSmits im a college student, its okay if you dont want to put your face lol. but yeah it just a project

struct: __attribute__((optimize("no-tree-vectorize"))) to disable auto vectorization

Kitkat260: is there teens here

jai_m: if you want i can send the link on github soo you can show that as your project in your college lol

jacek: it wants my bank id o.O

jacek: darn, where was my credit card...

darkhorse64: If I want to enable it again (to limit the scope on one function) ?

Kitkat260: hi..

struct: __attribute__((optimize("no-tree-vectorize"))) void function() { }

struct: should only affect that function

darkhorse64: OK

MSmits: I hope 7x9 is a draw

MSmits: makes games more balanced

MSmits: or it could anyway, no guarantee

MSmits: some games are solved as a draw and still heavily advantaged for one side

jai_m: all the details are fake enter every fake details

MSmits: for example TTT

jai_m: its just a dummy project

darkhorse64: fierce battle ahead

Kitkat260: sup guys

MSmits: btw darkhorse64, if I continue to work on yavalath book i might push you to 2. I keep running into losses vs you when i submit

MSmits: you have the best late game bot I think

MSmits: and i can't book the lategame. I win most games, but when you navigate all the traps like my bot does, it's a tossup

jacek: in other words it has no problems with premature finishing

MSmits: :P

darkhorse64: you're welcome. Not sure why: heavy rollouts maybe

MSmits: it's a feature of mcts i think

MSmits: you playout the game till the end

MSmits: which means you also detect when there's unfavorable placement that leads you without enough safe spots in the lategame

MSmits: it's not that easy to do this in a minimax

jacek: ad you pushed pony again

MSmits: i did ?

jacek: hes above me

MSmits: probably i just didnt fight it in my last submits

MSmits: so i didnt push it, i dragged others down

MSmits: I submit every once in a while, get 3 or 4 losses out of the submit and then check my meta mcts to see if there's anything interesting

MSmits: usually the losses are trictrac, dafish or darkhorse

Kitkat260: /me

jacek: you dont automate those submits?

jacek: go more meta

MSmits: nah, i do this stuff by hand :)

MSmits: i mean the meta also runs automatically, but it has a manual component

MSmits: i just enter in the moves manually, see where it takes me

MSmits: i guess euler would have already written a python script that downloaded all the moves and then reads them in from the meta mcts

MSmits: i did not go quite that far

MSmits: that heroku app of his does this sort of thing

Kitkat260: how do you code so well

MSmits: practice

Kitkat260: ok...

MSmits: not kidding, that's literally how

MSmits: just do a lot

Kitkat260: oh ok

Kitkat260: has any body seen proogmr

Kitkat260: nmv

darkhorse64: Great, 5th again at Y but definitely losing too many short games

MSmits: thats how it is for mcts bots

MSmits: traps are hard

Kitkat260: don't worry your not alone

Kitkat260: i am having trouble too

MSmits: darkhorse64 look at my recent battles

MSmits: and you'll see what I'm talking about

Kitkat260: gtg

darkhorse64: You mean losing in the endgame ? I could predict the on

MSmits: i mean the only losses i have are vs you :P

darkhorse64: result 11 moves before

darkhorse64: You one loss in 18 moves. Hole in the book ?

MSmits: yeah

MSmits: fixed that

MSmits: well

MSmits: maybe :)

MSmits: the 2 2 start is a bit difficult somehow

MSmits: maybe because my meta mcts simulated over 2 million games with 3 2 start and only 500k with 2 2

MSmits: because everyone else seems to use 3 2

MSmits: well except dafish and tric trac

darkhorse64: Games with supson are very short too

MSmits: yeah i didnt really target him with any book

darkhorse64: Looking at my book ...

MSmits: i guess he just uses the same moves as others

MSmits: there's a lot of overlap between bots

darkhorse64: I mean my games

MSmits: oh ok

MSmits: hm yeah it's so hard to get a bot to never lose in first 14 plies or so if you're not using a book

MSmits: branching is insane

MSmits: 60 X 59 x 58 etc.

darkhorse64: My book is start with any hex within the second ring and steal the inner hexes

MSmits: steal the inner hexes is pretty much the only steal that's certain to be good

MSmits: being first has a huge disadvantage in that you need 1 more safespot for your hexes

MSmits: or you lose to 3 in a row for lategame

MSmits: so I have no idea if it's worth stealing the others

MSmits: I steal 1 0, 2 1, 2 2 and the center ones

MSmits: dont steal 0 0, 2 0, 1 1 and 3 2

MSmits: these just have a < 50% winrate for me, but that does not necessarily prove anything

jai_m: http://smile-pay.herokuapp.com/ give a look to this project i have made it enter all details dummy.

darkhorse64: With EPT, I could bias my search towards triangles. My bot avoids these too much. Steal 1 0 ? Are you using the game coordinates ?

MSmits: yeah

MSmits: game coord

MSmits: so the one next to the corner on the edge

MSmits: dafish and trictrac open with it

MSmits: sometimes

MSmits: seen jacek's pony do it too, lately I think

jacek: i have the same opening as main

MSmits: oh ok

jacek: 1st move*, not book

MSmits: so it's 1 0 and 3 2?

MSmits: or maybe just random and i am seeing only losses with those as starting moves

MSmits: biased on my end

darkhorse64: Intuitively, starting on the border looks odd

MSmits: yeah i know

MSmits: http://chat.codingame.com/pastebin/64413398-0622-46ee-abd4-8bc080eab5f0

derjack: 1, 3, 59, 57, 13, 14, 15, 46, 14, 46, 45, 47, 46, 14 those are my starts

darkhorse64: because you have less offensive moves at your disposal

Astrobytes: jai_m: stop posting your link please

MSmits: 1 and 3 are 1 0 start

MSmits: 59 and 57 also

MSmits: so yeah, good deal, it seems a balanced start

MSmits: my meta is probably lopsided from counterbooking other players that start with it and i steal it

jacek: actually i stole that from you

jacek: good to have someone do the dirty job

MSmits: haha yes

MSmits: this is why i dont say "BOOK" when i play a book move, otherwise you could just steal the entire book

MSmits: there arent that many moves to steal really

Jrabovsky: array.pop() in python removes the first element right?

Default avatar.png zhoubou: The last element

Default avatar.png zhoubou: If you want to pop the first one, use array.pop(0)

Default avatar.png zhoubou: That is array.pop(index)

Jrabovsky: I did that

Jrabovsky: but it still returned the last element

Jrabovsky: Ive been having issues with the IDE recently

Default avatar.png zhoubou: There has to be a problem with your code somewhere then

Default avatar.png zhoubou: What specific issues are you having?

Jrabovsky: Recently ive been trying to perform operations that worked a week ago without the same result

Jrabovsky: Just then would be an example of one where a.pop(0) gave me the wrong value

Default avatar.png zhoubou: Maybe a[0] is not what you expect it to be

Jrabovsky: I printed the array before to check

Jrabovsky: the order was 3,1,2 and a.pop(0) returned 2

Westicles: thibaud is streaming

Default avatar.png zhoubou: How does one use CodinGame's pastebin?

jrke: just paste your text it will automatically convert to pastebin if the content is big

Default avatar.png zhoubou: Oh, thanks.

Default avatar.png zhoubou: Jrabovsky check your code thoroughly, there has to be something you're missing. Perhaps you can share a snippet where it occurs, and surrounding code, so I can have a look?

Jrabovsky: Alright. I wont be able to get the same test conditions as last time but Ill send something over next time it happens. Thanks so much

jrke: is there any way to compress a string to a letter?

Default avatar.png zhoubou: Compress in what way? What is your goal?

jrke: like encoding a string into a single letter

jacek: ?

Astrobytes: that makes no sense whatsoever

Westicles: I once fit the entire Holy Bible into the letter A

Astrobytes: :smirk:

jacek: char x = "x"[0]?

jrke: i think i need rest my mind is exhausted atm due to complexity of D&B

MSmits: there are for sure easier games on cg

MSmits: enjoying connect4 atm

MSmits: join us

eulerscheZahl: Thibaud is live again

jrke: i will join tommorow MS

MSmits: cool

Astrobytes: hm, not in a stream-y mood today

Astrobytes: puzzles again?

eulerscheZahl: yes

Astrobytes: tell him he should do multis instead :P

Westicles: he seems like a really nice guy

jrke: so Transpotiton tables are also usefull in connect4 i think

eulerscheZahl: tell him yourself astro

Astrobytes: I lack the motivation right now

Zenoscave: I see recrse added chess

struct: jacek http://blog.gamesolver.org/solving-connect-four/07-transposition-table/

Astrobytes: not just any old chess either

Astrobytes: also pew pew

jacek: struct?

struct: oh sorry jrke

struct: wrong ping

jrke: :smiley:

jrke: btw thats cool

Zenoscave: pewpew astro

MSmits: I;m going mcts, but thats a nice link indeed

jrke: gn

Default avatar.png karimk123: hello everyone

jacek: good evening

Default avatar.png karimk123: its night time for me but ill take it'

Default avatar.png karimk123: i got a new wpm today

MSmits: weaponized personal message?

Astrobytes: wombat petting machine?

MSmits: ah ofc

Zenoscave: word per microsecond

Westicles: wife praise m...d

Default avatar.png karimk123: ah yes funally 100 words per microseconds

Astrobytes: whoopass production method

Default avatar.png zhoubou: Why am I not getting notifications for achievements anymore?

eulerscheZahl: https://prnt.sc/we91hf

eulerscheZahl: (╯°□°)╯︵ ┻━┻

MSmits: which contribution eulerscheZahl?

eulerscheZahl: 3D one

MSmits: oh

eulerscheZahl: my only pending atm

MSmits: whats the reason?

eulerscheZahl: Hey CodinGamer! Your contribution has unfortunately not been moderated on time and will temporarily be removed from the system... But no worries! You can republish it anytime to give it another chance! :)

Westicles: i tried to get a mob together to approve last year, but they lacked vision

eulerscheZahl: i want a proper review anyways

eulerscheZahl: and now i'm pulling a JBM: camping on the contrib page

MSmits: ye, dont use blind people Westicles

Westicles: MSmits channeling jacek today?

Default avatar.png AnonymousUser: 0.00000133333 words per Microsecond : v

MSmits: yes :)

MSmits: eeeyup

jacek: oO

Astrobytes: :rofl:

jacek: its ee, fix your ponyism

MSmits: eeyup

eulerscheZahl: ohai

eulerscheZahl: happy Caturday's eve

Astrobytes: oO

Default avatar.png zhoubou: AutomatonNN what is love

AutomatonNN: wow, that's what i was solving anymore :(

Default avatar.png zhoubou: Did I remember correctly? :P

Astrobytes: AutomatonNN baby don't hurt me

AutomatonNN: ZarthaxX you work in a contest to see the code in between

eulerscheZahl: finally

eulerscheZahl: it's happening!

Astrobytes: He pings a few others too

eulerscheZahl: take that offline ZarthaxX

jacek: AutomatonNN where is your database

AutomatonNN: http://chat.codingame.com/pastebin/11745302-5e74-4cc1-98db-22adbbb7969d

eulerscheZahl: was to be expected

Astrobytes: hahaha

jacek: AutomatonNN invalid paste id

AutomatonNN: eulerscheZahl is there a way to see the code in the chat ?

Astrobytes: it did that in fr the other day

jacek: :tada:

Default avatar.png zhoubou: Omg

Astrobytes: :D

eulerscheZahl: can i kick again?

eulerscheZahl: was a nice day without automaton

Astrobytes: I won't tell

Astrobytes: Feel free

eulerscheZahl: meh, I won't stay for long anyways

Astrobytes: don't tell me your shaving another hour off your bedtime this year

eulerscheZahl: no, you'll have to deal with me for 1 more hour or 1.5 maybe

Astrobytes: :)

Westicles: what's the etiquette on changing the difficulty of an existing puzzle?

Astrobytes: Can you do it?

Astrobytes: As in, is it possible I mean

Westicles: Seems like you can

Astrobytes: This is for the nuggets thing right

Westicles: yeah. I wouldn't do it myself due to... people disliking everything I do... but someone else could

eulerscheZahl: nugget numbers?

Astrobytes: yeah

eulerscheZahl: could be interesting with larger limits, would require some maths. with given constraints it's easy to bruteforce

eulerscheZahl: (didn't try, just my feeling)

Astrobytes: I only had a quick look after Westicles comment on the forum

Astrobytes: It doesn't seem to fit the 'hard' category as it is

Westicles: that would be a fun stream. fixing the difficulty on a bunch of puzzles

eulerscheZahl: as a creator of hard puzzles I feel insulted by this categorization

Astrobytes: Hah! A fair few of them got a bit mixed up when they mixed the official puzzles with the community

eulerscheZahl: the auto assignment of difficulties

Astrobytes: what was it based on? Number of solvers?

Default avatar.png JBM: there wouldn't be so much camping if you just solved freecell eulerscheZahl

eulerscheZahl: there's an "impossible" testcase

eulerscheZahl: did you disable autoplay yet?

Astrobytes: damn, I forgot about that one as well. So many bloody contributions

eulerscheZahl: the binary tree is gone already :(

eulerscheZahl: hope it will return

Astrobytes: The solo game?

eulerscheZahl: yes

eulerscheZahl: https://www.codingame.com/contribute/view/43687cd854d10e38c3716e6d08efe735b42f

Default avatar.png JBM: pretty sure it will

eulerscheZahl: mouton is here for some years, so chances are good

Stilgart: eulerscheZahl: I am not a big fan of this one

Stilgart: he has better puzzles to publish

eulerscheZahl: i had fun solving it. i also remember the crossing lines one

eulerscheZahl: which is still on my TODO :(

eulerscheZahl: wrote some code but something is still wrong there

Stilgart: Astrobytes: dwarfie and I are sceptical about nuggets in hard

Stilgart: we spoke about it on #fr this afternoon

eulerscheZahl: we can argue about medium because of the -1 at best

Stilgart: is it that obvious that a naive while loop will be fast enough ?

Stilgart: proving it is not that easy imho (but it is not required to solve the puzzle)

eulerscheZahl: simple while loop gets you somewhere around 1e9

eulerscheZahl: an obvious upper limit for the array size is package size squared

eulerscheZahl: multiply by number of packages

Stilgart: 1e7 (not everyone uses C++)

eulerscheZahl: which makes it a bad selection on constraints even

eulerscheZahl: i think it could be interesting with higher constraints

Stilgart: not so obvious

eulerscheZahl: something with pairwise LCM i think

eulerscheZahl: didn't completely think it through

Westicles: Go ahead and change it to medium at least

eulerscheZahl: i see it as the task of creator/approvers to do so and only take action if they obviously don't care anymore

Westicles: well, it was djoums...

eulerscheZahl: another topic: where did the autoplay disable button on youtube go?

Stilgart: well, it was Westicles complaining in the forum...

eulerscheZahl: i hate that autoplay and now i can't figure out how to turn it off

Stilgart: wait wait wait...

Stilgart: please no

Stilgart: I still have one in the upper right

Stilgart: (you almost killed me there)

eulerscheZahl: player on the left, autoplay top right?

eulerscheZahl: not for me. maybe ctrl+F5?

eulerscheZahl: still missing :/

Stilgart: autoplay button on the right, still there after a ctrl+shift+R (firefow 84)

Stilgart: s/w/x/

Astrobytes: seems fine if you're logged in

Stilgart: seems that autoplay is enabled by default, though

Astrobytes: but disappears for me if I'm not

eulerscheZahl: https://prnt.sc/wea1p7 where should it be?

Stilgart: Astrobytes: no way I am logged in

eulerscheZahl: oh, i never log in on youtube :(

eulerscheZahl: they are collecting enough data without me helping them

Westicles: are you talking about autoplaying the next video? They moved it into the video itself.

Stilgart: juste below the sign in button

Stilgart: -e

Astrobytes: yeah I have a single gmail for youtube purposes

Westicles: over by full screen

Stilgart: I do not have a gmail account

eulerscheZahl: aaah, found it

eulerscheZahl: thanks Westicles

Astrobytes: Ah got it

eulerscheZahl: https://prnt.sc/wea2su

Astrobytes: Sorry, I was on a playlist

Stilgart: wow, what trickery is this ?!

Astrobytes: they are really messing with us

eulerscheZahl: we haven't redesigned the site for a while, let's confuse the users again

jacek: ux, apparently

eulerscheZahl: i still miss the star reviews

eulerscheZahl: (i admit it's easier to affect with stars than thumbs)

Astrobytes: ux these days tends more and more towards mobile with scant regard for users on computer

eulerscheZahl: as in: a 1 star on a highly rating video has more power. still I want the stars back

Stilgart: stars only last a few second, not worth it

Stilgart: </mario>

Astrobytes: :)

Westicles: i made the mistake of clicking on lithuanian girls pole vaulting. now that's all I get

Astrobytes: :no_mouth:

jai_m: im a college student i have made a project on face-recognition please have a look on it (note: enter all dummy details while signup) http://smile-pay.herokuapp.com/

MSmits: I thought of several things I could say to that Westicles, but then I remembered I am a teacher and this chat is public

MSmits: :P

eulerscheZahl: just do it, none of your students is here

Astrobytes: jai_m: if you continue posting that I will kick you from chat. That's the 3rd time in a couple of hours.

MSmits: I'll just wait for jacek to reply

eulerscheZahl: thanks Astrobytes, was considering such a reply too

Astrobytes: Rather annoying.

jacek: hmm? what reply

MSmits: to westicles

jai_m: okay im sorry

jai_m: i wont do that again :)

Astrobytes: jacek would be more interested in Lithuanian pony vaulting :P

Astrobytes: Thank you jai_m

jacek: yes, most my youtube related stuff are mlp, anime and neural nets

MSmits: lol Astrobytes

jacek: would you think google by now won't see the difference between ponies and multi-layer perceptrons?

Astrobytes: :D

reCurse: Ponies in 2021?

jacek: mods are asleep, postponies

Illedan: ETA on moving chess out of WIP reCurse?

Astrobytes: best new years wishes for you Illedan, reCurse

Illedan: You2

Illedan: Got any new years resolutions?

jacek: 1920x1200 as the last one

Astrobytes: lol

eulerscheZahl: not even 16:9

Illedan: :joy:

Astrobytes: Nah, I don't make them tbh

Astrobytes: When I want to make a change I go for it

reCurse: All the best to you as well Astrobytes

reCurse: And happy new year

eulerscheZahl: mine is not to wake up when neighbors go nuts

Illedan: Go nuts? As in fighting?

eulerscheZahl: fireworks

Astrobytes: Let me introduce the concept of: earplugs

struct: mine is to finish stc

reCurse: I think I'll publish chess soon enough and scrap crazyhouse

eulerscheZahl: you need 2 more accounts for that struct

Illedan: Define finish?

struct: finish my bot

reCurse: Don't think I'll do any more modifications though

Illedan: Why scrap crazyhouse?

Astrobytes: ^

reCurse: Definitely not popular

Illedan: But yeah, I kinda agree. Crazyhouse is hella confusing :P

eulerscheZahl: that's the whole point of it

Astrobytes: the clue is kinda in the name tbh

eulerscheZahl: so you have less research done already

struct: Release one and keep other in WIP

jacek: will there be a boss in chess?

reCurse: No

reCurse: I've moved on to another project and lost interest in chess for the time being

eulerscheZahl: you need one for puzzle of the week

jacek: ^

reCurse: ffs

eulerscheZahl: it will tripple your player count

Illedan: I can make you a boss for chess if you want?

reCurse: Why does it need a boss

Illedan: To have a funnier top league

eulerscheZahl: promotion = quest completion

eulerscheZahl: solve the puzzle

jacek: also it would filter out random and crashing bots

reCurse: I can't think of anything nice to say as a response, so I'll leave it at that

Westicles: I'm still looking for my second bronze

eulerscheZahl: and filter crashing player for higher league

jacek: boss could be just 2-ply material

reCurse: Can't even blob all the referee code because of the stupid 1 class java restriction, unless copy/pasting everything as inner class works somehow?

jacek: you can have multiple classes in java

eulerscheZahl: you can't create objects from a static context

eulerscheZahl: at least that's something I struggled with

jacek: you meant referee or bot?

eulerscheZahl: somehow it worked to create an object of the class containing the main function. and then objects of other types in the constructor

eulerscheZahl: just Java

reCurse: Well I would have blobbed the referee code into a bot file

reCurse: Then just do a 1-ply search with material evel

reCurse: *eval

eulerscheZahl: i don't remember the exact details. only that it totally confused me why it isn't working the way I want

reCurse: Yeah I feel the same way about java

reCurse: Intuitions are all wrong because arbitrary reasons

Astrobytes: I find the language fairly horrifying to be perfectly honest.

eulerscheZahl: too bad that android decided to build on java and not C#

reCurse: If only we could somehow fix the timeline and have microsoft start with .net core

eulerscheZahl: i guess they were a few years too early to see the benefits of C#

reCurse: Java would have died and the world be a better place

Illedan: ^

Astrobytes: But... then millions of devices wouldn't be running Java!!!!

Astrobytes: ...

eulerscheZahl: billions*

Astrobytes: eugh

Westicles: I have a grudge against Oracle. They retracted an internship offer when they figured out I wasn't a CS major

eulerscheZahl: i'm only mad at booking.com

Astrobytes: But you'd presumably demonstrated your capabilities otherwise Westicles?

Westicles: Yeah, the interview went fine

eulerscheZahl: what is your major in?

Westicles: EE

eulerscheZahl: and did you say it is in CS first?

Westicles: no

Astrobytes: Or did you have EE + CS

Astrobytes: But they only wanted CS major

eulerscheZahl: so, they invited you, interview and then: btw what are your qualifications?

reCurse: I'm more surprised they actually said a reason

eulerscheZahl: i had a (new retired) colleague who accidentally got the job

Westicles: Yeah, it was all kind of weird. They were just down the street so no big deal

eulerscheZahl: applied at another company in a foreign city. asked someone "how can I get to company CYZ"?

eulerscheZahl: "just follow those people, they all go there"

eulerscheZahl: walked into the wrong building, no interview scheduled. got the job :D

Illedan: lol

Astrobytes: lol really?

eulerscheZahl: yes, long ago when everyone desperately searched for qualified employees

Astrobytes: Epic.

eulerscheZahl: good old Walter. almost broke your hand every morning while greeting you :D

Astrobytes: Oh. I'm one of those people. I can't help it.

reCurse: Controlling competition, can't code as fast if the hand hurts

Astrobytes: I was brought up to always have a firm handshake as a mark of sincerity.

Astrobytes: lol reCurse

MSmits: hey, with player messages, the message is supposed to stay up during the opponents turn right?

MSmits: in connect 4, it turns off

jacek: make a comment in the contribution

MSmits: yeah doing that

eulerscheZahl: no one will get pinged as the creator doesn't exist anymore

MSmits: huh?

Astrobytes: not that connect4

eulerscheZahl: huh?

eulerscheZahl: and i was just wondering that connect4 has messages but no avatars

eulerscheZahl: which other connect4?

MSmits: https://www.codingame.com/contribute/view/59458ca85b08f914f150d1ccf0350fe7b233

MSmits: it's 7x9

eulerscheZahl: i totally missed that

eulerscheZahl: too many clashes and some other things slip through

MSmits: definitely hard to spot things in that list

MSmits: it needs a multiplayer filter

eulerscheZahl: Astrobytes check discord, #general

eulerscheZahl: oh, gone already :D

eulerscheZahl: did you delete it?

Astrobytes: nope, what?

eulerscheZahl: the herokuapp student advertised there as well

Astrobytes: Oh I told him to put a link there to stop him posting here

Astrobytes: Is it a malicious project?

Astrobytes: I didn't actually click it

eulerscheZahl: i didn't check either

Astrobytes: He put it in #talk about anything

eulerscheZahl: oh, i muted that channel

reCurse: It's funny for the wrong reasons

reCurse: He spammed it in 3 channels, deleted it everywhere except taah

reCurse: Should have been in memes though

struct: lol

eulerscheZahl: he deleted or you deleted?

reCurse: I did

jai_m: it is not malicious guys come on

jai_m: you guys just dont value things

struct: Well advertising it everywhere

struct: is sketchy

Astrobytes: OK, just don't spam it in 3 channels at one time then

reCurse: I value things accordingly

jai_m: i just want reviews on it there are no ads on it just dont call it meme or anything like that

reCurse: Oh so it's actually serious

reCurse: I'll leave that to kinder souls then

jai_m: if you guys feel you only know the things and you all are advance in all these things then you will not give chances to other peoples

Astrobytes: Just one more thing: why is it called smile-pay, has a signup and login and is also about image recognition?

**Astrobytes removes Columbo coat

reCurse: Actually had to check if it was a flash app

jai_m: it is a face-recognition project where i have used face-api library and used node for backend and monogdb for database and html-css for front-end

jai_m: https://github.com/jai38/Smile-Pay see the github link and find any malicious thing on it

Astrobytes: OK, calm down. For testing purposes in the manner you require I would strongly suggest disabling the signup/login features.

jai_m: it has nothing to do with your actual details i already told enter dummy details if you are trying though

Astrobytes: I mean, which part do you want feedback on?

Astrobytes: If it's just the facial-recognition part then just test that part, not sure why you need external testing for that

jai_m: i have done many things not only the face part, like for login signup passwords are hashed and this is my first project where i connected to database

jai_m: coz it is complete module of an payment app

eulerscheZahl: yes, you told us to enter dummy data. but we are lazy and you have to motivate us. sign up? I have to make something up just to see if I care. *closes tab*

reCurse: You have a login form for a payment app served in https with stretched out images using Comic Sans

reCurse: How is that not a meme

Astrobytes: Well, if you want feedback on that part it looks terrible for starters.

reCurse: *http

eulerscheZahl: that's how my mind works when I have some other tabs open

jai_m: it is just a dummy project guys if you feel it as meme its okay let it be

Westicles: turn it into a carefully crafted contribution and we'll get right on it in 60-180 days

reCurse: That's as nice as it gets

Astrobytes: Having road-tested that thing a little bit up to some point in the neverending signup process, I would recommend throwing it in the trash. Quickly.

reCurse: Convert it to flash for meme points though, I'd give it 10/10

Astrobytes: hahaha

jai_m: alright, thanks for your views on it :)

eulerscheZahl: i gave up again on the mobile number. not even sure how it should be formed. is that different in other countries?

Astrobytes: Also, it seems to be India-specific, which makes it kind of hard for non-Indian users

Astrobytes: Remove the 0 or the +CC eulerscheZahl

eulerscheZahl: meh, rage quit

eulerscheZahl: 123456 didn't work (probably too short)

Astrobytes: it tells you the required length

reCurse: Then definitely don't try this eulerscheZahl https://userinyerface.com/

eulerscheZahl: inyerface?

reCurse: It's actually genius

reCurse: I love it

Astrobytes: hahaha

Astrobytes: that's HCI/UX hell right there

Astrobytes: "in your face" eulerscheZahl

eulerscheZahl: had to look at the source code to get to the next level

eulerscheZahl: yes, got that Astrobytes :D

reCurse: That's cheating

Astrobytes: It's unbelievably realistic.

eulerscheZahl: that slow scrolling of the help box when you send it :D

Astrobytes: and the unstoppable timer

eulerscheZahl: with a popup every minute

eulerscheZahl: i like parodies

Astrobytes: I have genuinely come across sites like this

eulerscheZahl: like https://geektyper.com/

jacek: https://theuselessweb.com/

jacek: it redirected to CG

jai_m: thanks astrobytes and eulerschezahl for your time and sorry to make you wait soo long i will work on it in future

eulerscheZahl: https://hooooooooo.com/ what a nonsense jacek

eulerscheZahl: and the mandatory epilepsy warning for this link :rofl: http://www.staggeringbeauty.com/

Astrobytes: http://www.sometimesredsometimesblue.com/

eulerscheZahl: now i'm on a website with a horse in ascii art. and the legs just keep going when i scroll down. funny link collection

Astrobytes: Oh, and: http://cachemonet.com/

eulerscheZahl: https://d33wubrfki0l68.cloudfront.net/js/3682f7d78160ea8552ff7d63a3ed629f48fc1576/js/uselessweb.js the list isn't even that long :(

Astrobytes: there are many more than that

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

eulerscheZahl: it moved me to another page with more links now https://weirdorconfusing.com/

Astrobytes: Well. It gave me this: https://www.ebay.com/i/193213712423

reCurse: That's enough internet for today

Astrobytes: Seriously.

eulerscheZahl: yes, why are you posting that when i want to go to bed?

eulerscheZahl: but i'm stuck under the cat anyways

Astrobytes: I'm killing both of mine mentally as they're causing mayhem and destruction.

MSmits: beat your bot darkhorse64. I get around 2M sims in first turn (9 sims for each rollout)

darkhorse64: 3.2 M for me

MSmits: lower than yours, but might be the sim is heavier, or i should just optimize more

MSmits: takes multiple submits to pull you down now lol

MSmits: so few games

darkhorse64: heavy rollouts (play wins, avoid losses) ?

MSmits: i dont even play wins

MSmits: my sims only draw or lose

MSmits: if you cant play a move that is not losing, you lose

MSmits: basically, i calculate winning moves for the opponent. If there are more than 1, you lose

MSmits: if you cant block it without giving him a winning move, you also lose

MSmits: it's similar to the way beginners play connect4. Just avoid giving the opponent a win

Astrobytes: connect4?

MSmits: yeah

Astrobytes: Oh I see your last statement

MSmits: it solves really early, this game

MSmits: 20 plies before the end with a mcts solver

MSmits: makes sense, with so many columns becoming disallowed

Astrobytes: board size increase needed? Or octagonal as darkhorse64 suggested?

MSmits: not needed imho.

MSmits: othello solves around 18 before the end

Astrobytes: cool beans

MSmits: well ofc opening books will wreck this game, but that was a given

darkhorse64: Now that I have a serious bot to play with, I see this early solve too

MSmits: cool

MSmits: the message turning off is annoying, hope he fixs it

darkhorse64: post a request, the author is responsive and very pleased to get feedback

MSmits: I did

MSmits: like you, i wonder how minimax bots will do against ours

MSmits: not that easy to eval this i think?

MSmits: I'm guessing you count possible 4 ina rows or something

darkhorse64: The C4 solver does something like that to order the moves

MSmits: ah ok

darkhorse64: It takes several seconds to find the win with 6x7. Hope it will be enough to confuse MM bots

darkhorse64: You wrecked me. I have been #1 just one day

MSmits: plenty of time to think of improvements

MSmits: the key is heavier rollouts as you said

MSmits: no reason to give away wins during a sim

MSmits: it's like yavalath

darkhorse64: One third of my Y bot is for the rollouts

MSmits: you mean the lines of code?

darkhorse64: yes

MSmits: that makes sense

MSmits: I have a lot of lines for generating all the lookup tables

MSmits: but the sim has a lot too

darkhorse64: lot of ugly lines

MSmits: yep

MSmits: found a bug in connect 4

MSmits: if you win with the last move, it says it's a draw

MSmits: he obviously placed the draw-check before the win-check

MSmits: I am also a little worried about the amount of times my p1 bot beats my p2 bot. As in... almost always

MSmits: I guess i might just be deterministic

MSmits: This game might need a steal option

MSmits: maybe not, not sure

struct: MSmits tell me if this makes any sense

struct: this is google translated from japanese

struct: Speeding up field implementation by bitboard The field is 6 rows wide, 14 rows vertically including the invisible part. If you put the wall up, down, http://chat.codingame.com/pastebin/31ccf947-1e5f-480e-9559-6e600f4031b0

MSmits: not to me no

struct: yeah

MSmits: probably would need to study this hard to know what it's about

MSmits: also translated out of Japanese = a bad start

struct: Well from the code I read they do store all the colors and skulls on 3 128 bits xmm

struct: so 5 in 3

MSmits: hm ok

MSmits: looks painfully complicated

MSmits: connect4 was a breeze to avx

struct: MSmits still there?

MSmits: yea

MSmits: was about to go to bed, why?

MSmits: struct

struct: i think i found another good article on it

struct: I share it tomorrow

MSmits: ahh ok

MSmits: gn then :)

struct: gn

LPTFabio: How i report a bug about "coders Strike back" in some game my bot finishes in first but it counts as a lost (against the boss)

LPTFabio: nvm

ZzzTRAPzzZ: how do I raise/drop my clash of code score?

Wontonimo: play more clash of code

Wontonimo: your score is pretty much as high as it can get. There are a few more points for you to get, but it doesn't go above 5000 XP. If you want more XP you'll need to play other competitions, like the Bot Programming.