Chat:World/2021-01-29

From CG community
Jump to navigation Jump to search

Default avatar.png popomafu: lol

VizGhar: Any contributor here? I have 2 sets of images for sprite animation. one is idle animation, other is attack animation. Problem is, that attack animation has wider images and it looks like the API scales images to match some square

VizGhar: so my character is much smaller while in "attack" mode

Hedayat_Farahi: wasssup beautiful people

Hedayat_Farahi: fuck ya'll

VizGhar: Solved my problem :| assets were incorrect

Default avatar.png Carsonfanboy: hey guys

Default avatar.png Carsonfanboy: i am normaly only used to python and now im trying to do smth in c++

Default avatar.png Carsonfanboy: can smbd tell my how to define or change an variable?

Default avatar.png hoistbypetard: you have to declare it before you can do anything with it

Default avatar.png hoistbypetard: that means you either tell the compiler its type or tell the compiler to guess the type using the auto keyword

Default avatar.png hoistbypetard: so if you want foo to be zero,

Default avatar.png hoistbypetard: int foo = 0;

Default avatar.png hoistbypetard: or

Default avatar.png hoistbypetard: int foo;

Default avatar.png hoistbypetard: (later)

Default avatar.png hoistbypetard: foo = 0;

Default avatar.png Carsonfanboy: oh ok i see

Default avatar.png Carsonfanboy: so i can change it at any time?

Default avatar.png hoistbypetard: as long as its in scope

Default avatar.png hoistbypetard: and you didn't declare it const

Default avatar.png hoistbypetard: running through some of the examples here: https://runestone.academy/runestone/books/published/cpp4python/index.html might help you

Default avatar.png Carsonfanboy: kk thank u

Default avatar.png Carsonfanboy: one more question:

Default avatar.png Carsonfanboy: if i do: http://chat.codingame.com/pastebin/3fb313d1-aba9-4d6f-a71f-eaec6cb88e83

Default avatar.png Carsonfanboy: than i do: http://chat.codingame.com/pastebin/ed96550b-7dae-410d-b5c4-f2c1daca77c6

Default avatar.png Carsonfanboy: is this right?

Default avatar.png Carsonfanboy: nvmnd

Default avatar.png Carsonfanboy: i solved it

Default avatar.png Angecide: is it possible, with some clever tricks, to extract opponents and play against them offline or can this only be done inside CG?

Default avatar.png Angecide: in order to validate my bot in fast succession

kovi: https://github.com/s-vivien/CGBenchmark

Default avatar.png Angecide: thanks

kovi: (even though i havent used it yet)

philRG: thanks

Default avatar.png Brandon: @kovi Thannnnnks

tutubalin: "Given a backward quote from Yoda, rewrite it in an understandable way."

Banned this CoC puzzle should be

DaNinja: <(°.°)>

depthzer0: In our hearts Yoda because

depthzer0: He is grandpa of Cheburashka

depthzer0: https://tlum.ru/uploads/106b23d858b5b2c8b95add8eb4b37e3884f644060f4cdace95d42a8c9fcb7fc4.jpeg

tutubalin: oh, i misunderstood the statement :(

tutubalin: it was not that bad

NobodyGuy: python code is so... short! :)

Default avatar.png Rockstar555: hii

tutubalin: NobodyGuy Ruby is shorter

tutubalin: GolfScript is shorter than Ruby

Andriamanitra: depends on what you're doing

Andriamanitra: if you want to reverse a string "string"[::-1] is shorter than "string".reverse

tutubalin: Jelly is shorter than GolfScript

Default avatar.png zizo1337: reverse is faster to type, to me atleast

Default avatar.png Rockstar555: yeah

Default avatar.png Rockstar555: i think

Default avatar.png Rockstar555: shortcut is better

Default avatar.png Rockstar555: i don't like shortest mode

Satzmann666: Shortest is the best

Default avatar.png npo: Hello there

Gebes: Sad java noises @Satzmann666

Default avatar.png vincccc: :rage:

Tobu123: hello

Default avatar.png ArturSworski: econd

Default avatar.png normalkid: dudes

Default avatar.png normalkid: sup

Default avatar.png sayedm: Hello

Default avatar.png PaulCaron: za warudo

Default avatar.png PaulCaron: guys why you are toxic ???

Default avatar.png PaulCaron: no answer

Default avatar.png PaulCaron: nop answer dead

Default avatar.png PaulCaron: var mountainH = 0;// hauteur de la montagne var altMax = 0;// altitude maximalevar mountainIndex = 0; // index de la montagne à detruire// game loop while (true){ for (var i = 0; i < 8; i++){mountainH = parseInt(readline()); //Lecture de la hauteur de la montagne d’index i// complétez ici}print(mountainIndex); // The index of the mountain to fire on.}

Default avatar.png PaulCaron: :computer:

struct: o..o

eulerscheZahl: calm down satan

BlaiseEbuth: Already kicked him on #fr

eulerscheZahl: i know

BlaiseEbuth: And so ?

1400179: Are checkers endgame with equal number of kings on each side proven to be forced draws?

1400179: If both players play perfectly

eulerscheZahl: MSmits that's for you

struct: if no king can be taken on the next move I think its a draw yes

1400179: Sure seems like it but I don't know theory

AntiSquid: PaulCaron learn to meme more eloquently!

1400179: Much more interesting game AI wise than I thought though

1400179: Definitely better than othello

1400179: How to throw a game: https://www.codingame.com/replay/525331262

AntiSquid: You've gotta love crab.

1400179: F.

BlaiseEbuth: :crab:

1400179: Dealing with endgame shuffling is pretty challenging

1400179: Definitely a good step before chess

1400179: Also is the game bugged? I thought jump + promotion + capture backward was legal...

1400179: Did I play this game wrong all along when I was a kid?

eulerscheZahl: there are different versions of the game

1400179: :scream:

1400179: Where's the FIDD when you need it

Cybersick: Hello Everyone

1400179: Do I need to throw away 10 hours of training by fixing it though :thinking:

1400179: Hi

jacek: FIDD?

1400179: FIDE for checkers? Random guess

jacek: as for 3-repetition, i was thinking to add it to checkers and marking it as something slightly-better-than-lose so i would force to look for more

1400179: Yeah I'm surprised that's not standard either

1400179: It feels like it should be easy to solve endgames with some domain knowledge though

jacek: and in variation when you can jump backwards, in this case you wont get promoted

1400179: That's weird

jacek: as a kid, did you also played with optional jumps? :v

1400179: Not that I remember

1400179: Sounds like a lame variation though

jacek: oh right, you can choose which jump. in international variation you must take jump with most pieces

1400179: So many variations!

jacek: eeyup

jacek: i still need to fix my bot. i copy pasted some bitboard implementation, but it errors with cycle jumps

1400179: :frown:

jacek: move generation is hard here :c

struct: Cant really find anything on same number of kings

1400179: Then MSmits is my last hope

PatrickMcGinnisII: I can't believe my minimax on chess is timing out on depth 3

1400179: Exponentials are a b

struct: This version doesnt have the 7 move draw rule on 2 kings vs 1 king right?

1400179: Don't think so

1400179: Call me judgemental but expectations are a bit low when the color given at init is not even correct :p

cooldude420: Hello there!

1400179: Hi

cooldude420: Can you please assist me with my code? I would very much appreciate it!

Default avatar.png DJNinjaSpoon_37b8: :boy_tone5:

1400179: Maybe someone has time

struct: ask the question cooldude420

cooldude420: If someone does, it would be very helpful! Thank you!

cooldude420: Woah how do you get red text? That is very cool.

struct: What problem are you having, what are you solving?

struct: Just type the person username

Default avatar.png Manuel-Martin-Jukisz: Hi guys, i will have an interview in few days. Do you guys recomend anything that related to data manipulation?

cooldude420: Ohh I see, struct. I am currently working on Mars Lander in episode 2, using C++.

struct: I haven't solved that one sorry

1400179: Define data manipulation

cooldude420: It is alright, I appreciate the help!

cooldude420: I hope you all have a good one.

1400179: Cheers

Default avatar.png Manuel-Martin-Jukisz: reCurse, Anything really, i just want to practice general things. The company works with the financial market and they refine and display the data into an app.

jacek: you can ask away here, maybe someone solved it and will help you

1400179: Did they give any hint about the tools and techs they use? That would be a start

Default avatar.png Manuel-Martin-Jukisz: Nothing really in deep, they are useing React JS

1400179: Is this a frontend or backend position?

Default avatar.png Manuel-Martin-Jukisz: for this coding interview will be forntend

Default avatar.png Manuel-Martin-Jukisz: frontend

1400179: Ok

1400179: Then maybe lookup tutorials and stuff on react?

1400179: Look into simple CRUD apps

1400179: It's hard to tell otherwise what kind of interview they'll have, there's a lot of kinds

1400179: So go with what you know

Default avatar.png jetlow345: Hi guys, I am new to programming !!

jacek: hi, new to programming

Default avatar.png Manuel-Martin-Jukisz: reCurse, than you, i already ask for the kind of interview they will make.

Default avatar.png jetlow345: Normally how you guys practice coding?

BlaiseEbuth: I don't. I'm just here for trolling.

Default avatar.png ErrorRazor: How do I have one puzzle solved in Go, never touched that language

PatrickMcGinnisII: reCurse could you add a + to the UCI notation in chess for check condition for the moves sent? it would help wood2 from having so many draws.

TBali: ErrorRazor - solving Onboarding in Go should not be a problem, hint is provided in IDE

Default avatar.png ErrorRazor: I'm not saying I want to solve one, I'm saying that the site thinks I have although I haven't

PatrickMcGinnisII: glhf life calls

TBali: strange,

TBali: Here you can which puzzle you solved in shich language (at least what CG thinks :-) )

TBali: https://chadok.info/codingame/

TBali: *check

TBali: **which

Default avatar.png ErrorRazor: Oh yeah, Onboarding was the tutorial and I did it in Go. Totally forgot lol

jacek: now do it in Chess

Default avatar.png ErrorRazor: Bruh

Default avatar.png ErrorRazor: I'm gonna do it in DnD

jacek: so far no one did puzzle of the week

eulerscheZahl: no time, busy with topcoder

TBali: Pluralsight's skill assessment is seriously flawed

Wontonimo: well, my first contribution is about to expire in 50 min ... if anyone here is feeling generous in time, please consider reviewing it https://www.codingame.com/contribute/view/5996057237509a508857acda4426be02fc01

eulerscheZahl: meh, you can submit it again when time is over

eulerscheZahl: (story of my less popular puzzles)

Wontonimo: good to know, thanks

MSmits: reCurse, you still here?

MSmits: are you looking to write an endgame book for in your live bot or for offline to train with?

MSmits: and i assume it's checkers, didn't read the entire convo

1400179: Not necessarily, I was more curious about endgame theory

eulerscheZahl: checkers, yes

MSmits: ah

MSmits: well there's two kinds

MSmits: WLD and DTM

1400179: Have the intuition that equal number of kings means forced draw

1400179: With perfect play

MSmits: win/loss/draw is most common, but it's useless on CG

eulerscheZahl: isn't DTM a racing league?

MSmits: Depth to mate, means the number of turns to win, for the given state

MSmits: this is what you need for CG checkers

1400179: Are you talking about some kind of tablebase?

MSmits: yes

1400179: Ah ok, I was more interested into the theory itself

1400179: But that's fine too I suppose

1400179: I just don't think it fits CG constraints either way

MSmits: well the theory i know is how to generate the db

1400179: Ok

1400179: But you haven't examined it per se

MSmits: you can fit a 3 piece db

MSmits: I generated parts of it offline

1400179: How big is a 3 piece db

MSmits: hmm, not sure exactly but it fits. Every piece you add makes it 10-20 times larger

MSmits: roughly

MSmits: 4 is way too big

1400179: Well I'm kind of using some space already eh

MSmits: it could be that the step from 3 to 4 makes it go from negligible to way too mcuh

1400179: At least it could start providing an answer to my original question

1400179: If 2vs2 is forced draw

1400179: That's something

MSmits: 2 piece vs 2 piece?

MSmits: it's not

1400179: 2k vs 2k

MSmits: it depends on the state

1400179: Quiet state

MSmits: hmm

MSmits: you should probably just generate the db offline to be sure

MSmits: you can easily check it after that

MSmits: you need retrograde analysis

MSmits: you know how this works?

1400179: The big lines

1400179: Never coded it

1400179: Feels a lot like DP

MSmits: let me get my code, sec

1400179: This game is a lot more fun than I thought

1400179: Should have done it earlier

eulerscheZahl: do you want to write a classic non-NN bot now?

1400179: No not at all lol

1400179: But the number of fixes and improvements I made from it is monstruous

MSmits: crap this is complicated, i saw binomial coefficients just now

1400179: I could probably make my other NN bots much stronger now

eulerscheZahl: you are already the unbeaten #1 in breakthrough

1400179: Your point?

eulerscheZahl: calm down, give us a chance

1400179: I couldn't be bothered much investigating othello losses

1400179: But checkers losses now that's fun

eulerscheZahl: you see what's going on

1400179: Yeah

MSmits: yeah checkers is a much more fun game to watch

eulerscheZahl: othello: *flip* - ooh

MSmits: I'm currently running my 3 piece endbook in release mode in VS, probably should have compiled it first, it's slow

MSmits: so yeah, you can definitely not generate the book in first 1s

MSmits: like you can in oware

1400179: Would have been surprising

struct: not even 2 pieces db?

1400179: Isn't that useless?

struct: yes

Wontonimo: even better, 1 piece db

1400179: I'm surprised at how balanced checkers seems to be in general, and spare me the solved game crap :)

MSmits: the difficult part of endgame db for checkers (it's worse in chess) is that you need to enumerate combinations of types in the right order. So All kings first, then slowly move in steps to all "men"

MSmits: king states come after men states, so they are a more advanced endgame

MSmits: so you need to calculate them first

1400179: I think you can do all kings db

1400179: Seems like more useful

MSmits: sure, if you have to choose it is

MSmits: but you could have 1 m piece 2 kings vs 2 kings

MSmits: might still be a draw

1400179: The tradeoff between space/time and usefulness seems a no brainer

Wontonimo: with 100ms response cap, I don't think it is "solved" anymore

1400179: Yeah but maybe your search would avoid those situations by relying on the king db

1400179: Going straight for the juicy win

jacek: or just mark 3-fold-repetition as almost loss

1400179: That's orthogonal

jacek: what does it means

jacek: sounds mathy

1400179: It is

jacek: :Scream:

1400179: Means unrelated independent concern

1400179: In this case

MSmits: reCurse, the end game timer means you need to not just know win/draw/loss, but pick the state that wins the quickest. some endgame sequences are 100+ turns

MSmits: some wins might be draws on CG

jacek: or just download some 7p endgames and train on them

MSmits: thats why i meant DTM db instead of WLD

1400179: Wonder if the tradeoff is worth it

1400179: Gamble on the move timer for bigger table

MSmits: this is the best source on checkers btw:

MSmits: http://www.fierz.ch/download.php

MSmits: also information on generating endgame db

1400179: Thanks

jacek: depends how often those 100+ wins occurs

MSmits: not dtm i think, they explain WLd, but you can figure out dtm from that

1400179: Uh? They have 5 piece but not less?

1400179: tbh I'd gamble on the move timer

jacek: maybe its too trivial

MSmits: possibly 5 piece contains all lesser ones too?

1400179: Could store a 4x bigger table that way? Easily

MSmits: remember, it's a factor of 10-20 each time

jacek: btw. draught folks just discovered nnue and try it for their draughts engines

1400179: Might just make the difference between one more piece, or it fitting at all

struct: Its a shame that lidraughts doesnt have tablebase on analysis board

MSmits: biggest WLD db are 10 pieces for checkers

MSmits: DTM is 7, but not publicly available

1400179: And on top of that

1400179: I can use those big dbs for training

1400179: Yummy

MSmits: just be careful about the timer. The greater the db, the longer the longest possible game from each state is

MSmits: if i remember correctly 7 piece already had a 250 turn endgame in there somewhere

1400179: The training will pick that up

1400179: All good

MSmits: as long as you put the turn timer into the training it's fine

MSmits: people seem to ignore that for oware

MSmits: i guess its different because you can win on a non-ended game there

1400179: Sorry for how it sounds like, but crap I hope I don't end up 100% before doing that lol

1400179: Motivation goes down after

MSmits: what do you mean 100%?

jacek: like bt? :unamused:

1400179: 100% win

jacek: though i won once!

eulerscheZahl: 100% winrate

MSmits: do you mean rank 1?

MSmits: oh

jacek: and i dont have timer for oware :o

MSmits: well... I could... nvm :P

1400179: :unamused:

jacek: dont keep it to yourself

jacek: open yourself to us

MSmits: hehe no need

MSmits: I'm working on game of drones now

1400179: Sorry for my harsh reaction the other day

1400179: Still think it's a D move

MSmits: need to get gold, my students are silver

MSmits: it's allright

1400179: But should have been less of a D myself

jacek: its all because of D lang

MSmits: don't worry about it reCurse

MSmits: I was not intending to stick it to you, but i understand why it would seem that way.

struct: astro or trictrac here?

1400179: ?

trictrac: yes

1400179: Oh

trictrac: hello struct

struct: hi tric trac, I was wondering about the starting positions

struct: Should I allow for starts on the middle?

struct: For a piece to start

struct: Currently its always symetric and I avoid corners

trictrac: yes fully random only symetrie for the 2 players

struct: Ok, currently is not fully random


struct: I decided to make a player have a piece on each 4x4 space

trictrac: thats good too

trictrac: good idea you can keep that

struct: Astro suggested to not place pieces next to each other

MSmits: Do you mean forced starts, or random starts?

MSmits: it's different

jacek: random

struct: random

MSmits: ah

jacek: random, the biggest enemy of MSmits

1400179: Forced starts too, as long as they're numerous

MSmits: nah, it saves me from my obsession of using a particular avenue to win. I prefer not having the option

MSmits: it's cleaner, i do a game till i run out of ideas, then i can move on

1400179: Is amazon interesting to analyze? Didn't seem like it to me at first glance

trictrac: so it's not an option to have random for MSmits tranquillity :slight_smile:

MSmits: no, please do have random for my tranquility :P

MSmits: btw, i blame miarem and reCurse for starting this on uttt :P That was the gateway drug

1400179: I only started it because I realized (at least) Nagrarok was doing it

MSmits: he must have been doing it poorly though, his book didnt go deep at all

1400179: Was still good enough at the time

MSmits: ah ok

MSmits: i never noticed it

1400179: Then I went to the dark side and built a pipeline to continuously batch games against top 10 and build a gigantic counterbook

1400179: I reformed since

MSmits: oh, you went much further than me then, I never got around to automating from replays

MSmits: i just did it manually

MSmits: well the meta mcts was automated, manual wasnt a big deal because it takes hours to counter a single move

MSmits: mmh i guess with those 10 cores you used though....

1400179: 12

MSmits: even better

1400179: Ran mcts for a million iteration on each losing move

MSmits: ohh so it's not actually running full games?

MSmits: thats different

MSmits: I run thousands of games to counter a move

MSmits: depending on the game, between 10 and 100 ms per turn

1400179: Worked well enough :P

1400179: And it timed nicely with the rate of games from batch

MSmits: yeah there's many ways to do it i guess

MSmits: miarem ran games with extremely long turns

MSmits: I thought you mostly did this to obtain information to train a NN

1400179: No I did this out of spite

MSmits: ahh

MSmits: well i guess it would seem annoying when others do it to you

MSmits: to me it's 2 things really. On some games it's just exciting to me, to see what the best moves are. Add to that increased winrate on the leaderboard

MSmits: but on other games that first part is nonsense

MSmits: for example othello

1400179: Othello is like a mini uttt

jacek: or oware

MSmits: actually uttt is way more interesting to me than othello

MSmits: i never went very far analyzing othello theoretically

1400179: Sacrilegeous

MSmits: I have written so many coordinate systems and such for uttt, to better hash stuff

MSmits: i should say board representations

MSmits: I even have 1 that fits in 127 bit

MSmits: yavalath is interesting to me in the same way. Oware, partially, mostly the endgame stuff was fun to come up with

MSmits: I'm still running a book generator for oware btw... just so you know. I won't be turning it off for this game. Oware and yavalath are the only games i am booking currently

1400179: I'll stay clear from those

MSmits: sure :)

1400179: Not sure what I'm doing next actually

MSmits: after checkers youmean?

1400179: Probably still got some fun with checkers endgame

1400179: Yeah

1400179: If I do chess I may vanish for good

MSmits: if you dont care about playercount, a dependable leaderboard etc and you only want an interesting game, i have some options

jacek: :notebook:

1400179: I only care about having good varied opponents

Westicles: samegame

1400179: If I do chess I'll have thousands of them

MSmits: hmm i was gonna say twixt, but not enough varied opponents currently

1400179: And some I won't ever beat

MSmits: onitama is a great game,

1400179: I'm not sure I enjoy the move mechanic though

1400179: Looked nice at first then I realized it might just become tactics tactics

MSmits: whats that

jacek: like chess

Westicles: Another indian deleted...

MSmits: you just said tactics twice

1400179: Should I have said it thrice?

1400179: Like othello is 99% tactics and 1% eval

1400179: imo

jacek: huh

MSmits: ok i thought you were saying something profound by saying tactics twice, but you just wanted to say, mostly about tactics

Westicles: https://www.codingame.com/contribute/view/604966d6b8bb9b27ed1e2668aec89656d894

1400179: Me saying something profound :huh:

MSmits: not sure what the problem is with mostly about tactics

1400179: Just becomes a search race

1400179: Like uttt, whoever searches more

1400179: Though I may have broken it, maybe

MSmits: not sure

1400179: ?

MSmits: with uttt, you're the strongest bot currently. But every time a stronger bot came around, we thought that was it

1400179: I know for a fact that I search way less than the others around

MSmits: if i submitted my current bot a year ago it would seem unbeatable

MSmits: oh

1400179: Won't give the exact number but still

MSmits: sure

MSmits: if thats what you meant, then you're completely right

MSmits: but I am of the opinion uttt isnt all search either

MSmits: I prune a lot of moves

MSmits: and use quite a few heuristics

1400179: Then your opinion is going with my result

kovi: for the record: i failed to reach legend...without heuristic

1400179: I was convinced you couldn't beat a bot that searches at least twice as much because there isn't enough to read

kovi: even with uttt heuristic comes faster than speed

1400179: But I was wrong

MSmits: oh, yeah i figured that out a while ago losing to bots with only 60k rollouts

MSmits: then they shared their heuristics

MSmits: karlis o also shared a bit a while ago

1400179: Don't stick to the 2x too much, could be 4x or 10x

MSmits: still have to try this

1400179: I'm sure you couldn't beat a bot with 10x search

1400179: Even with those heuristics

MSmits: yeah i dont think so either

Default avatar.png Angecide: is it true karliso has 400k rollout round 2?

MSmits: but the window of performance is only about a factor of 2 wide at the top

MSmits: between the slowest and fastest

MSmits: Angecide no

Default avatar.png Angecide: oh ok

MSmits: he said so at one time

1400179: Yeah but that's not my point

MSmits: but he was wrong, miscalculated

MSmits: it's about 220 i think

struct: he incremented counter twice

struct: Same mistake I did on STC

Default avatar.png Angecide: 220 still immensely impressive

MSmits: you still have to know how he counts this

1400179: ^

MSmits: if you expand a node and do a rollout from each child, you get much more

1400179: As soon as you introduce heuristics those numbers are incomparable

MSmits: for example, i simply assume everyone uses teccles

MSmits: you can get far deeper into the search

MSmits: this affects simcount, though i am not 100% sure how

MSmits: it's a safe assumption, because even if they dont use teccles, it's probably going to be a worse move

kovi: but teccles ifself is just make narrower path isnt it?

MSmits: yeah, that means you expand to 1 child instead of 9

MSmits: i have some more heuristics like that

1400179: Alright broken record incoming

1400179: Number of rollouts is a joke because it doesn't take into account game length

MSmits: yes, games end faster if you rollout smart

kovi: true. if you manage to win early...

MSmits: also if you rollout smart, the calculations are slower

MSmits: so its hard to gauge the result properly

Default avatar.png Angecide: ye I think jacek does like 15k rollout round 2, and he is 14th or something

MSmits: for people just getting started with uttt and not worried about heuristics and pruning and all that, it's a fine metric

jacek: but i may count it differently than others

MSmits: those bots are all the same basically

MSmits: jacek, yeah maybe you count expansions and then simulate multiple times

MSmits: in 1 expansion

jacek: i count only when i do rollout

Default avatar.png owent3877: good mornin

Default avatar.png owent3877: im back

jacek: ohai

MSmits: rollout is also a vague term, it's used for the entire process from selection to end of simulation and also used for a single random rollout after expansion

Default avatar.png owent3877: yup

MSmits: but since you can do multiple random rollouts from 1 expansion its hard to say what you mean

jacek: i mean as playout, simulation

Default avatar.png owent3877: hi

MSmits: ah ok

Default avatar.png Angecide: maybe depth is a better metric?

MSmits: not with mcts

MSmits: it's going to be all over the place

jacek: mcts is asymmetric

MSmits: that works well with minimax

Default avatar.png Angecide: ye I see

1400179: cough ab

Default avatar.png owent3877: what are you talking about?

MSmits: i should say ab?

jacek: ultimate tic tac toe

Default avatar.png owent3877: wow

kovi: hmm, teccles is not the strongest heuristics...

MSmits: what is stronger then?

jacek: i use teccles only for some openings as 1p

1400179: ab makes it incomparable again

Default avatar.png owent3877: im working MIME Type

MSmits: ah ok

kovi: i have just added another variation on avoid 9b

kovi: 9board

MSmits: you mean to avoid giving away a free move?

kovi: yes

MSmits: yeah thats fairly important and it could be stronger than teccles, but it's not as reliable

MSmits: you need to tweak it a lot

1400179: If ab prunes a ton of nodes because my eval is crap, doesn't matter if I reach depth 10 in 1 ms

Default avatar.png owent3877: have you tried code.org?

1400179: That depth 6 in 100ms might be better

MSmits: ahh ok

MSmits: does it actually prune more with a bad eval?

jacek: with eval = 0 you prune very much

MSmits: i would expect the opposite

1400179: Wouldn't work in my example would it

1400179: Not the point

MSmits: no i got it, eval matters

MSmits: also when you introduce move ordering and history heuristic and all that

1400179: Just specific cases of ab

1400179: As soon as ab is there

1400179: Forget metrics

MSmits: hmm you can use the metric of counting the number of applied moves

MSmits: it's not perfect, but it is something

1400179: Small improvement

1400179: The metric that matters is wdl

kovi: wdl?

1400179: win/draw/loss

MSmits: it matters for the leaderboard sure. But I find it very hard to judge things by winrate these days, when a single move made can change it so much

Default avatar.png owent3877: i do that a lot

1400179: You write a bot to play a game

1400179: At the end of the day that's what you optimize for

MSmits: yea

Default avatar.png owent3877: what are you working on?

Default avatar.png owent3877: in doing Coders Strike Back


Default avatar.png owent3877: hello?


Default avatar.png idknow: doing clash of code

jacek: need some help with CSB?

Default avatar.png Angecide: I heard it is all about smitsimax

kovi: not for very top

eulerscheZahl: hi kovi are you playing topcoder?

Wontonimo: hey reCurse, thanks for the link to the MCTS review. it was a good read

kovi: hello euler! no, i have lost motivation there

jacek: the survey?

1400179: You're welcome

eulerscheZahl: TC in general or just that game?

DomiKo: which TC game?

eulerscheZahl: https://imgur.com/a/myiwKh2

MSmits: reCurse if i did it correctly: 3 pieces state count: 199520

eulerscheZahl: https://www.topcoder.com/challenges/b846362a-218e-4a3d-b86c-3d38f472e199?tab=details

MSmits: too many it seems

1400179: How many bits per state

MSmits: I'm afraid you need a byte

1400179: Hmm

MSmits: because it's not WLD

1400179: Might fit still

MSmits: with a byte you could go up to 255 turns left till win

MSmits: total state count: 6408836

MSmits: 4 pieces

MSmits: so you see the size of the jump

MSmits: between dbs

1400179: Yeah 4 pieces would need some shortcut

1400179: Ah but you're doing pieces

1400179: What about kings

MSmits: basically what you would need then is all 2 piece + part of 3 piece

MSmits: at least thats the order in which you generate

1400179: Ohh

1400179: Why didn't I think of this before

1400179: You don't even need the whole db

MSmits: well the earlier dbs are negligible in size

1400179: Even 4 pieces could fit

MSmits: but yeah within size 3, you can do a part

MSmits: how so

1400179: You only need anchors roughly spaced by the depth of your search

MSmits: hmm

MSmits: but most states could still happen

MSmits: so you would just randomly select a bunch of them?

1400179: Not randomly

1400179: Semi-randomly

1400179: Pick a random state, exclude everything reachable in 5 moves, rinse and repeat

MSmits: seems difficult to do that properly

1400179: I don't think so :)

MSmits: but states are reachable by multiple paths

1400179: You just need *a* path to win

1400179: And you don't need the path to lose

MSmits: that works for opening books, not sure if it works that well for endgame books when there are so many states to start from

1400179: It's a tradeoff

1400179: But a good one I think

MSmits: worth a try

MSmits: oh btw i was wrong, you dont have 255 turns left to win, you need to include the loss as well

MSmits: so +127 or -127 in an sbyte

MSmits: +100 would mean 100 turns till you win

1400179: Like I said, I don't see the point to including the loss

MSmits: -100 would mean 100 turns till you lose

MSmits: well i dont mean that

MSmits: i mean you look up from the perpective of the person who's turn it is

MSmits: unless you would only look it up on 1 of the 2 players

1400179: Yeah, and with my hypothesis you don't need the perspective of the loser

MSmits: right cuz thats not one of the anchors

MSmits: you'd have to check what the longest path to win is for your selected state. Probably you need 8 bit, maybe 7. 6 is too little i think

1400179: If anchors work there's not much to worry about number of bits per state

MSmits: allright

MSmits: so basically you're using this to avoid repetition, but you dont mind that it doesnt always cut short your search immediately

MSmits: sometimes requiring 3-4 more plies

MSmits: because your anchor isnt at that particular depth, but a little deeper

1400179: Yeah, you still have search to leverage

1400179: If you know you're in an endgame state, that's a lot

MSmits: well you have to think about the fact that the book isnt just used from an endgame state

1400179: Sure

MSmits: it's also used from a pre-endgame state, from which the search enters endgames

1400179: I'm assuming training will take care of most of what matters there

MSmits: makes sense

MSmits: i still think selecting those anchors is not gonna be easy. I already think generating the endgamebook itself is pretty messy. Now you also need to be smart abou tit

1400179: I'm not worried

MSmits: well let me know if you succeed. I dont need the details, but it would be cool to know if something like that works in general

1400179: Yeah if I get there will do

MSmits: kk

1400179: I'll start by plugging that 7 piece db in my training

MSmits: be careful about the fact that all those DB on that site are WLD

MSmits: DTM is something you have to generate yourself, but you can use their code

1400179: So how do they resolve endgame with WLD

Default avatar.png owent3877: cool

MSmits: you mean, without infinite repetition. I think there exist some heuristics as well. But i am not well versed in checkers theory

1400179: There has to be otherwise you can't prove it

MSmits: I know how to mate a king with a rook and king in chess... but doing stuff like that in checkers is different

MSmits: reCurse yes, but thats retrograde analysis, its very expensive, not done in a live game

1400179: No I know

1400179: But for an engine to use it

1400179: It can't just say "oh btw it's a proven win"

MSmits: thats right, so i think it uses WLD to cut short the search when possible and then when it actually has to play out the win, it uses heuristics to end the game

MSmits: probably taking a lot more time than the DTM result

1400179: Hmm

MSmits: chinook does not play perfect DTM

MSmits: it will take more time to win than necessary

1400179: Oh so simplification maybe

1400179: For checkers especially since it's the win condition

MSmits: the thing is, as long as you use a heuristic that will speed up the end of the game, you can always still lookup the WLD value to make sure you dont accidentally lose, so whatever you do it's safe

1400179: Yeah makes sense

MSmits: but on CG it's different, because of the 300 turn thing

struct: this happens with syzygy chess table bases too, a forced mate is a forced mate

MSmits: WLD db can make you lose

1400179: Never gave tablebases too much thought

struct: doesnt matter if its 2 or 3 moves

1400179: Yeah but I'm gambling on moves left not to matter

MSmits: i should say wld can make you draw sry

1400179: Or at least not as much as having a bigger tb

MSmits: it doesnt matter for small db

MSmits: the larger the db, the longer the longest possible game

1400179: ofc

MSmits: and I think you're right that whats currently on the leaderboard is going to be wrecked either way by a NN with an endgame book

1400179: Well with my flawed training and improvements I get rank 3

1400179: Top 2 will require more work, but already need to restart training with all the fixes

MSmits: my bot is an early playout termination bot that basically just counts kings and men with a param

MSmits: dont compare with that :P

MSmits: a few guys above me, top 3 is good i think

jacek: oh im above you

MSmits: doesnt surprise me

MSmits: I did actually spend a lot of time on checkers, but mostly to test a minimax version vs a ept version

1400179: ept?

jacek: i just half-assed NN and it works

MSmits: it's basically the way you do mcts with a NN

jacek: early playout termination

1400179: Ah

MSmits: but with a typical eval

MSmits: its what i use in most games now

MSmits: i guess if i ever do a nn, i can just plug that in :P

jacek: i could train it more but its bugged and im too dumb to copy paste workable bitboard implementation

jacek: every github project has gazillion abstracts

jacek: why they dont make hackatonish code

Cybersick: Why codingame server is always empty

MSmits: i'm not sure if i have a great sim either. It's a bit hacky with a bunch of pdep pext and all that

Cybersick: And there are too few people

jacek: clash?

Cybersick: Yes

Cybersick: Hello ?

jacek: maybe its not this time of day

jacek: if there are only few people, there will be bots in their place

MSmits: reCurse, i think it's about 15k with just kings in 3 piece db

struct: 15k states?

MSmits: yeah

1400179: Woo

MSmits: there's a lot of ways to have 2 kings and 1 "man"

jacek: at this point if one side doesnt have king its loses, doesnt he?

1400179: So 4k is definitely in reach

MSmits: or 1 king and 2 men

jacek: thats so gay

struct: What is the longest state out of the 3 kings one MSmits?

struct: ah wait

struct: Forget it

MSmits: i didnt finish my retrograde calculator, tis is just counting states

MSmits: i needed to do this first, to decide how far to go with it

Default avatar.png Angecide: on the topic of intrinsics, I have found great use for pdep to find n'th significant bit, but why would anyone use pext? It just extracts the mask and shifts it all to the right, what would be the usecase for this?

MSmits: project is on hold anyways

MSmits: Angecide lookups

Default avatar.png Angecide: hmm

MSmits: you take a pattern from the board

elePHPant: Hosting a meetup Monday in which we are working on the genome sequencing problem. https://www.codingame.com/ide/puzzle/genome-sequencing

Default avatar.png Angecide: oh I think I see

MSmits: say 6 bits, spread out

MSmits: compress them to the first 6 bits

MSmits: then have a small lookupindex

MSmits: very useful in yavalath

MSmits: where win patterns are spread out over 61 bit

jacek: and probably in bishop moves

MSmits: yeah

MSmits: jacek uses it for ntuples i assume

eulerscheZahl: whoo!

MSmits: whoo to you as well eulerscheZahl

jacek: nah

eulerscheZahl: thanks oreshnik

MSmits: why not?

jacek: when i use rows, i just shift and mask

jacek: for squares, i also use some hacks

MSmits: oh right, i guess the patterns are simple in othello

jacek: my pc doesnt have pext/pdep

Westicles: Euler approved just in time for the weekend

MSmits: ohh which one

MSmits: he had multiple

Westicles: tryangle catch

MSmits: ahh the pretty one

MSmits: I love it, but it seems hard

eulerscheZahl: so blockout isn't pretty?

MSmits: i didnt have a good look at it :P

MSmits: i actually studied tryangle a bit

MSmits: seems doable till you get the option of making new paths

MSmits: thats what makes it seem hard, but definitely worth trying out

elePHPant: Anyone interested in joining me on monday to work on the code sequencing problem?

elePHPant: Meetup via zoom

eulerscheZahl: wanted to keep you busy for at least 10 days ;)

MSmits: yeah that would definitely have worked

kovi: yay, i try to catch toad

eulerscheZahl: and kovi is beating me already :(

elePHPant: We are going to watch brief videos about DNA and then jump into the problem

MSmits: ah well

elePHPant: everyone will have the opportunity to share screen and talk about their code

MSmits: biohacking?

eulerscheZahl: monday... we have jobs

elePHPant: Monday evening

jacek: hopefully not php jobs :v

elePHPant: 6pm eastern

MSmits: we already have a virus out there, dont make more please

elePHPant: lol

elePHPant: and lol

kovi: so many games for wood4?

eulerscheZahl: i'm not even allowed to install zoom on the company computer, IT isn't happy about their data privacy :D

eulerscheZahl: maybe because everyone submitting at once

eulerscheZahl: so you get your games + other games

elePHPant: love it when devs demote PHP, it means more work for me

kovi: oh, true

MSmits: I hope they didn't go by rumor on that one, didnt they fix their issues? Zoom is definitely the nicest program

eulerscheZahl: and community contributions are weird in general, not a classic wood

eulerscheZahl: i never tried zoom

MSmits: wish i had zoom for teaching

eulerscheZahl: we have MS Teams

MSmits: we use google meet, but no breakout rooms yet :(

elePHPant: teaching - thats part of what we do in the meetup

MSmits: i use teams as well

MSmits: for colleagues

MSmits: for some reason my school uses both microsoft and google

kovi: oh, boss is also on submit

kovi: never seen that

eulerscheZahl: yeah, bosses get more matches then regular players

elePHPant: just in case someone is interested ... https://www.meetup.com/social-coding-virtual

eulerscheZahl: wow, that's midnight in my timezone

elePHPant: Ouch

elePHPant: where is everyone from?

eulerscheZahl: smits can tell you how realistic it is that i will join :D

eulerscheZahl: central Europe

eulerscheZahl: hover over our avatars to see the location

elePHPant: Oh ok, I feel like the "odd PHP guy from the United States"

elePHPant: LOL

jacek: oO

elePHPant: oO

jacek: do you feel irrelephant

eulerscheZahl: odd PHP guy from US is taken already

eulerscheZahl: that's PatrickMcGuinessII

Westicles: you beat me to it

Westicles: unless he can top the sex swing...

struct: Euler did you manage to find a fix for the display bug on blockout?

eulerscheZahl: no :(

JLukeSkywalker: pinball

eulerscheZahl: my fix is not looking too closely

JLukeSkywalker: we need a pinball game

struct: From what I understood I think the invisible pieces

struct: also can have an impact on hit

struct: But I dont know anything about 3d rendering

struct: So I could be wrong

struct: on it*

eulerscheZahl: hm, i could try to play with it: if invisible move faaaaar off

struct: yeah

eulerscheZahl: but not today anymore

struct: Doesnt need to be very far

struct: Just not on the board

kovi: no promotion till boss settles i guess

kovi: westicles at negative score :o

Westicles: Did I win?

kovi: none

eulerscheZahl: promotion popup is working \o/

eulerscheZahl: you can't test that in the demo environment

kovi: hmm im not that good with attack

eulerscheZahl: i outcommented mine

kovi: me2

kovi: not sure if it is better or worse

MSmits: that's weird, tryangle catch opened on php

MSmits: i guess i have no choice now

MSmits: time to learn php

eulerscheZahl: did you test CotR in PHP?

kovi: wood3 boss beating me

MSmits: eulerscheZahl do you mean cgfunge, thats the one where you recommended php

eulerscheZahl: i got a perfect 100% vs boss

MSmits: never did it in cotr

eulerscheZahl: i mean code of the rings

eulerscheZahl: as you did that recently

MSmits: ohh now i get you

eulerscheZahl: IDE saves your last language

MSmits: you mean it stayed on php

kovi: yeah, i see my old weakness

MSmits: it's C#

MSmits: and after that i did GoD, also C#

MSmits: so i dont understand

MSmits: maybe Patrick hacked me

kovi: wow, top20 uttt....

MSmits: do you search kovi? Or some heuristic?

MSmits: wait you're working on two multis at the same time now?

Westicles: When he says CG hacked his GPU, are we just politely quiet or is there something to it?

Default avatar.png Angecide: insane climb on uttt kovi

kovi: nah, i tested catchangle earlier

MSmits: i dont know, doesnt webgl use your gpu, doesnt mean it's hacking does it?

eulerscheZahl: well done kovi UTTT was my nightmare on legend leagues

kovi: uttt lucky run maybe

kovi: yeah, i failed with pure vanilla mcts

eulerscheZahl: there isn't that much luck

eulerscheZahl: some randomness at the top

eulerscheZahl: but top20 is top20 and not top50

MSmits: luck happens when bots are similar in strength 10-30 might be similar

MSmits: but 50 is definitely worse

kovi: well...today i improved from gold to top20

MSmits: gj

jacek: oh my

kovi: yeah i think top20-40 is the proper range

MSmits: if you test vs my bot you might get the occasional win, but try vs karli so and re curse, they're almost unbeatable as p1

kovi: draw is ok

MSmits: you'll lose most as p2 as well

kovi: true

MSmits: the game is lopsided enough that you'll win some when you're p1

MSmits: even with a weaker bot

kovi: i think that is th reason im so high :)

MSmits: hard to say really. Best way to test is to just place 10 opponents spread out over the top 30 of the leaderboard. Then do 1000 CG bench games

eulerscheZahl: struct i still get the missing surfaces when I move invisible blocks out of range

dbf: eulerscheZahl, congratulations with approved game

eulerscheZahl: thanks :)

kovi: i need to look into mcts solver

MSmits: input is huge

Default avatar.png Angecide: kovi u got 23th with vanilla mcts?

MSmits: kovi, small winrate bonus, but ++ coolness factor when you output solved games 15 plies before

Default avatar.png Angecide: ah I guess u mean u don't do anything special with terminal nodes

kovi: i failed legend with vanilla. added heuristics

kovi: ok, some perf improvements also came late today

kovi: yes terminals

MSmits: so much input eulerscheZahl, or maybe it's because i switch to C# and it's so verbose

kovi: my code is still messy and im sure those can be backpropagated better

MSmits: ye i am used to seeing that cin << << << << crap

eulerscheZahl: you mean for Tryangle catch?

MSmits: yes

Default avatar.png Angecide: my code is only mcts solver, whenever I add heuristic, somehow it just makes my bot worse

Default avatar.png Angecide: I need to learn the proper way of involving heuristics in my bot

eulerscheZahl: you can ignore parts of it

MSmits: i would not go C++ with this game

MSmits: so I have to build a graph right

MSmits: been a while

MSmits: well unless you count mcts

dbf: Angecide what language do you use for uttt?

Default avatar.png Angecide: c++

Default avatar.png Angecide: currently 31 legend

kovi: afk a bit

struct: euler, what do you mean missing surfaces?

eulerscheZahl: the sides of the blocks

eulerscheZahl: you can just see through

eulerscheZahl: i already closed it again, going to bed now

struct: ok, gn

MSmits: gn euler

jacek: mcts solver gave me like 55-60% winrate in uttt

struct: lose rate %?

jacek: hm?

MSmits: you mean compared to a bot without it

struct: loserate as p1

MSmits: i am sure he did a p1/p2 alternated test

MSmits: vs his own bot

jacek: eeyup

MSmits: without mcts solver

jacek: locally with thousands of games

jacek: it was long before i put many heuristics and improvements i have now

Default avatar.png Angecide: ah, I just saw it on cgstats, it says 56% win, which seems to match ur number as well with mcts solver

MSmits: not sure how well it generalizes to other bots with other circumstances, but 100% sure it helps

Default avatar.png Angecide: in my case

jacek: from my experience it never hurts

MSmits: i mean the solver only really starts to use resources when it solves

MSmits: and when it solves it helps because you lose those nodes to explore

jacek: in othello it may not help but doesnt hurt either, and during self-games tests its faster as 10 plies are solved so you can test more games

MSmits: in othello it helps a *ton*, my bot solves at ply 43 or so

jacek: your bot is fast, avx guy :c

MSmits: true

struct: o.o

struct: 43?

MSmits: I am happy it still did ok without book, was afraid it would drop below 10 like it did when i did my messed up yavalath test

MSmits: still not sure what went wrong there, i interleaved the book into the search too much

MSmits: jacek when you use a book, do you only use it at the top of the tree and just forego search or do you use i on expansion, as part of the search?

MSmits: i do that last one in yavalath and its very messy. Not in other games though

struct: We need more board games

MSmits: lol not really

struct: From all I ported I only like Y and Amazons

MSmits: I mean i love them, but

MSmits: there's a lot now :)

MSmits: hmm

struct: I will only stop when there are more multies than clashes

MSmits: lol

jacek: MSmits i only use book in Y and just directly the move

jacek: not for search

MSmits: ah ok

struct: We need knightthrough

MSmits: dont know it

struct: like breakthrough

struct: but with knights

MSmits: did you just make that up?

struct: no

struct: its real

MSmits: mmh I think you made it up

jacek: https://www.researchgate.net/publication/316147806_Generalized_Rapid_Action_Value_Estimation

MSmits: ow

MSmits: he didnt make it up

MSmits: I could solve Rookthrough

struct: is that one real?

MSmits: No I made it up

MSmits: you see how that works?

MSmits: :P

MSmits: rookthrough would be so boring, you start and poof you're on the other side

Astrobytes: Aren't there variations of amazons using knights and other pieces?

Astrobytes: Sure I saw that somewhere

struct: no idea

jacek: crazyhouse breakthrough!

Astrobytes: breakthrough960... :P

MSmits: yeah!

MSmits: 960

MSmits: make a new contrib for this struct

struct: Ill just replace Yinsh with it

struct: So I dont have to go through approval

jacek: not much love for yinsh eh

Astrobytes: Yet.

kovi: how can you shuffle bt?

MSmits: Keep yinsh, its a fine game like paper soccer and twixt, it just didnt get popular

Astrobytes: kovi: that's the joke ;)

struct: astro I added random positions, pieces can only be next to each other if they are in a different 4x4 square

struct: otherwise not

Astrobytes: Oh that sounds good

MSmits: ah

MSmits: so 1 piece of both sides in each 4x4?

struct: yes

Astrobytes: Yep

MSmits: not bad, needs some playtesting though

MSmits: but tric trac is good for that

MSmits: he fixed onitama too

Astrobytes: It might just create chamber mayhem very quickly and be uninteresting or it might be interesting

struct: yeah, Its all up for testing

trictrac: struct tomorrow I will make a bot for amazons

kovi: gotta go now

struct: thanks I should also do the fireworks one tomorrow

Astrobytes: And hopefully I will finish my fireworks bot trictrac

struct: This weekend I will have it ready

kovi: try angle has 4 leagues and only 5 players

struct: :(

Astrobytes: So many games

MSmits: ye

trictrac: thanks and perhaps you can approve :slight_smile:

struct: othello got 78 players since puzzle of the week

jacek: and how many in wood1

struct: 0

Astrobytes: lol

jacek: buy buy quest map

jacek: meh

struct: its free points

struct: 78 players = 780 points for rnak 1

JohnnyLuke: made a clash about cheating on an exam, and accidentally show that you can get a better score than the person you cheat off of .....

JohnnyLuke: :grimacing:

MSmits: yeah, only cheat on the questions you dont know the answers to

Wontonimo: that is the principle behind "boosting"

Wontonimo: several weak classifiers can make a strong classifier

dbf: nice title of the preview page for a game: https://clip2net.com/clip/m572140/e1c5a-clip-50kb.jpg?nocache=1

MSmits: just a fyi for any of my students reading this chatlog after the fact. I was speaking hypothetically :P

JohnnyLuke: lol

struct: lol dbf

MSmits: lol practice disappointment

Astrobytes: Hahaha dbf and MSmits

MSmits: i bet he did not realize this when he added the tags :P

dbf: Can I be 2% if I practice 5%? :thinking:

1400179: lmao

Astrobytes: I think he did MSmits in all honesty

MSmits: maybe

MSmits: funnier if he didnt

1400179: Where is that preview from anyway

Astrobytes: He will be happy tomorrow when he sees it I'm sure.

struct: email?

struct: Im not sure

dbf: reCurse it is a link preview in telegram messenger

struct: ah

Astrobytes: Hahaha excellent

MSmits: telegram? Isnt that where all the rightwing extremists went when they were kicked off twitter?

1400179: No

Astrobytes: huh?

MSmits: mmh must be confused with something else

MSmits: they went from parler to something else

dbf: MSmits, yes extremists and some CG players

Astrobytes: Telegram is just a messaging platform, Russian-based, fairly secure (or was)

1400179: TTT still too low to my liking :/

struct: ttt?

1400179: Time to Trump

MSmits: https://www.nbcnews.com/tech/tech-news/telegram-recent-haven-far-right-purges-extremist-content-n1254215

1400179: Uh? Must have missed that development.

1400179: Then again I'm trying to detox now

MSmits: it was one of the last things i saw before i completely ghosted the US

Astrobytes: Yeah I didn't see that

MSmits: havent watched the US news for a few days now

MSmits: feels great

Astrobytes: lol, we're all detoxing now eh

Default avatar.png zhoubou: What is this news you talk about?

Astrobytes: It's almost a relief, yet also worrying

MSmits: better you dont know

JohnnyLuke: i wish i could not watch US news, unfortunately stuck here

JohnnyLuke: Y'all lucky

MSmits: i'm sure lots of crap is still happening Astrobytes, but it's at a different level for at least a while I think

Astrobytes: You are not your government JohnnyLuke

Astrobytes: MSmits: yeah I'm not gonna talk about it anyway, peace and quiet for now :)

MSmits: JohnnyLuke, you actually only have to watch enough to know whats happening and make informed decisions. You don't have to follow it 24/7

MSmits: i was watching way too much of it

MSmits: Astrobytes agreed :)

Default avatar.png zhoubou: I wouldn't have a clue if a war broke out in my neighborhood

Default avatar.png zhoubou: That's how much I read

MSmits: you'd think it was fireworks?

Default avatar.png zhoubou: Yeah, probably

Astrobytes: Another quake!

MSmits: not unreasonable btw, in my country

Default avatar.png zhoubou: Exactly Astrobytes

Default avatar.png zhoubou: And a 1000 more!

Default avatar.png zhoubou: Jeesh

Default avatar.png zhoubou: I don't care if there were 27k quakes below 2.0 or something

MSmits: are you old enough to actually have been in a war zhoubou?

MSmits: I mean it's not that long ago is it?

Default avatar.png zhoubou: No MSmits

Default avatar.png zhoubou: Fortunately no

MSmits: good i suppose

Astrobytes: That was a horrible time.

MSmits: definitely, some scars in my country from that war

MSmits: even

MSmits: some of our military had to stand by while genocide happend =/

Astrobytes: I won't talk about it. Too much for CG chat in all honesty.

MSmits: ye

MSmits: sorry

Astrobytes: No, no worries.

Astrobytes: Anyway, zhoubou, what you working on now?

MSmits: trying to get a tryangle catch bot going

MSmits: ye zhoubou, sup?

Astrobytes: Too many inputs to do quickly no?

Default avatar.png zhoubou: Astrobytes MSmits Well, studying mostly

struct: just use the starter kits provided by cg

struct: :zipper_mouth:

Astrobytes: Aha, almost forgot about the studying there. When's the exam again?

Default avatar.png zhoubou: Other than that, going through CS career path on Codecademy, and helping someone out on puzzles

MSmits: ohh

Default avatar.png zhoubou: 9.2.

MSmits: you have a pro account?

Astrobytes: ah yeah

MSmits: for codecademt?

Default avatar.png zhoubou: Yeah

MSmits: cool

Astrobytes: struct: lol, not my style

Default avatar.png zhoubou: I could spare some money

MSmits: I have 2 students doing their html stuff for free

Default avatar.png zhoubou: It's not worth it if you're not me imo :P

Default avatar.png TheFlashSaxophone_e85b: are you guys bots?

Default avatar.png zhoubou: But it's worth it for me

Astrobytes: Yes.

MSmits: just me TheFlashSaxophone_e85b

MSmits: i wouldnt call Astrobytes a bot, he's just a handful of heuristics

Default avatar.png zhoubou: TheFlashSaxophone_e85b Of course not. Beep boop.

struct: 2210437 users

Astrobytes: True AI MSmits

MSmits: lol

trictrac: kovi : almost 0 win against your bot !!

Default avatar.png zhoubou: AI is always just a bunch of if statements, amirite?

VladimirAngelov: no

Astrobytes: The One True AI is yes

Astrobytes: (it's a fairly old CG meme)

Default avatar.png **zhoubou googles. Fails.

MSmits: from some guy whining about people using a search right?

MSmits: on the forum

MSmits: instead of just doing if statements

Astrobytes: Yeah, can't remember the contest.

Astrobytes: Also you can't mention True AI without Prune Arrays and cegimax

MSmits: of course

MSmits: it's very relaxing to work on a bot in C#

MSmits: no worries about catastrophic bugs

Astrobytes: ...

MSmits: or performance

MSmits: I was think maybe getting a bit creative with Linq and stuff

MSmits: because i have the functional programming class soon

Astrobytes: Yeah, the linq stuff is handy, very handy

MSmits: i need to get better with it

dbf: MSmits do you teach FP?

MSmits: no i am both student and teacher

MSmits: for this i am student

MSmits: hopefully done by summer

MSmits: then i will be qualified as high school CS teacher

Astrobytes: 'done' you say... :)

MSmits: finished i mean :P

Default avatar.png zhoubou: Does CG have self-assesment?

MSmits: mmh depends on your definition

Default avatar.png zhoubou: Like if I want to check my knowledge.

MSmits: not like on codecademy

Astrobytes: Not by any really measurable metric no

Default avatar.png zhoubou: Ah ok

MSmits: but puzzles have tags

MSmits: so you can practice in a more focused way

Astrobytes: unless you keep track of what you're learning on the way - even then, hmm

Default avatar.png zhoubou: Oh I meant the jobs part of CG

Astrobytes: No I don't think it does

MSmits: oh yeah companies can arrange assessments here

Default avatar.png zhoubou: But I can't arrange one for myself?

MSmits: for job interviews

MSmits: hmm dont think so

Astrobytes: Yeah, interview assessments but not personal

Default avatar.png zhoubou: Hmm ok

MSmits: want us to assess you?

Astrobytes: Not that I'm aware of but perhaps message Thibaud

Default avatar.png zhoubou: MSmits I would fail currentl :D

Default avatar.png zhoubou: currently*

MSmits: maybe

Astrobytes: Failing is a large part of succeeding.

Default avatar.png zhoubou: Good point

Default avatar.png zhoubou: How would you assess me anyways?

MSmits: tbh, that sounds like a garbled yoda quote :P

MSmits: doesnt make it any less true though

Astrobytes: lol

Default avatar.png zhoubou: A large part of succeeding, failing is.

MSmits: no idea zhoubou

Default avatar.png zhoubou: There, fixed it for yah

Astrobytes: :)

Astrobytes: brb

MSmits: well there you go, you pass

Default avatar.png zhoubou: :D

MSmits: Yoda certificate of excellence

Default avatar.png zhoubou: Can you provide a pdf?

MSmits: sec

MSmits: even better

MSmits: https://imgflip.com/i/4vw654

Default avatar.png zhoubou: :tada:

Default avatar.png zhoubou: I'll print it and put it on my wall. Thanks!

MSmits: np :)

TBali: cool :-)

Default avatar.png highsgod: Ayee

AntiSquid: https://imgflip.com/i/4vwbks

TBali: so all my bots are a big success after all?

MSmits: yoda is always right

TBali: "A succeeding failing is large"

MSmits: "failing large, is succeeding"

MSmits: thats the one you need

TBali: So Coelho is a 900y old yedi master

Default avatar.png zhoubou: Large is succeeding, failing

Astrobytes: Succeeding, a large failing is

Default avatar.png TheFlashSaxophone_e85b: test

TBali: permutate()

Astrobytes: Failed

Default avatar.png TheFlashSaxophone_e85b: sun

MSmits: input is now read Astrobytes

MSmits: finally finished debugging the input

Default avatar.png TheFlashSaxophone_e85b: ai

Astrobytes: lol

Default avatar.png TheFlashSaxophone_e85b: bot

MSmits: i forgot that C# doesnt actually stopping you from making the mistakes, it just tells you so you can fix it right away

MSmits: lots and lots of fixing

TBali: next multi - yoda contest - the worse your bot is, the higher you are on the leaderboard

TBali: but crashing or timeout does not count

Astrobytes: I think it's kinda language independent MSmits

Astrobytes: :D

MSmits: maybe :)

MSmits: so next up, roaming around the map randomly

TBali: no, php does not tell you

Astrobytes: That's what I mean TBali, the errors are yours to make, and yours alone.

TBali: it is like corporate bullshit - outside everything is shiny, internaly it rots

MSmits: php rots?

TBali: MSmits, I see you are physics teacher. What age group do you teach?

MSmits: it was in my IDE when i opened it either

MSmits: earlier

MSmits: TBali 16-18

MSmits: I also teach a bit of computer science, same ages

TBali: My son is 17.5, wants to be physics researcher

MSmits: a few hrs a week only

Astrobytes: "rots" - reminded me of my favourite Swedish phrase: min pumpa ruttnar

MSmits: thats cool TBali

TBali: focuses on it quite a lot, now

TBali: had his first paper accepted

MSmits: either that is your only Swedish phrase Astrobytes, or i just discovered a new side of you

Default avatar.png zhoubou: Astrobytes I love Swedish. I'm actually slowly learning it through Duolingo

MSmits: TBali wait what?

Astrobytes: How did it go TBali? I remember you were looking for lightweight GUI ideas for him

MSmits: a paper at that age?

MSmits: I only ever wrote 1 paper

Astrobytes: MSmits: I have a good friend in Sweden :)

MSmits: ah I see

TBali: yes, quite proud now

MSmits: i can see that

Astrobytes: But I learned that in a restaurant from a waitress.

TBali: https://iopscience.iop.org/article/10.1088/1361-6404/abd989/meta

Astrobytes: zhoubou: Nice! It's an odd but beautiful language, I should put time in to learning it properly some day

Default avatar.png zhoubou: Yeah, same :)

MSmits: ah TBali, so it's an extension of his childhood basically. First you bounce balls, then you write papers about bouncing balls :)

AntiSquid: initial public offering science ?

TBali: :)

Astrobytes: eh? IOP, institute of physics

TBali: One of the peer reviewer commented: nice, but it requires only high scool physics

AntiSquid: ya i can't read properly anymore

TBali: Well he is high school. But actualy there was more than high school maths in it

AntiSquid: i actually thought i see IPO

Astrobytes: Screen break for squid

MSmits: TBali if they say "it only requires high school physics" then thats poor criticism. If you can solve a problem noone else has, with simpler tools, so much the better

Astrobytes: Agreed. It's not exactly a trivial problem.

TBali: he studies in a spec math class , best in country so could use some analysis

MSmits: I used high school physics to prove there's a black hole in the center of our galaxy, with my students

Astrobytes: "chaoticity" though - that a real word?

TBali: now it is accepted, after some changes needed to submitted

MSmits: i mean it wasnt original but we used the original data

Astrobytes: Nicely done MSmits!

Default avatar.png zhoubou: That's cool

MSmits: funny thing is, 1 week after they handed in the project, the scientists that used the data got the nobel prize

MSmits: complete coincidence

Astrobytes: Nice coincidence though eh

TBali: Astrobytes, yes this was the same paper the simulation in Java/processing was needed

MSmits: ye

MSmits: so... my point is, nothing wrong with high school physics :)

Astrobytes: Thought so TBali, well done to him (and you for helping him out where needed)

Astrobytes: MSmits: it's the foundation

MSmits: exc

MSmits: exactly

Astrobytes: Doing biology, chem and physics at the same time for me in highschool was confusing but rewarding when they converged at points

MSmits: I just received an e-mail from a student just like your son TBali. He was the 11th best student in my country last year, to take the exams

AntiSquid: is that not normal Astrobytes ?

MSmits: did similar stuff

Astrobytes: That's awesome.

AntiSquid: like 2 hours or so of each of those ?

Astrobytes: No I took them as main subjects

AntiSquid: you had choice eh

Astrobytes: Not much. Had to fight for it, because I wanted to take art and music too

Astrobytes: Was talking about that yesterday

MSmits: oh it's normal here, to do all those together

AntiSquid: our entire classes we arguing with the school director about having spanish instead of french and the guy shut us down saying no ...

AntiSquid: class *

AntiSquid: were *

AntiSquid: sleepy T_T

MSmits: sometimes it's either/or with bio and phys, but many do both. Chem is mandatory for them

Astrobytes: Yes for the first 2 years of highschool you do everything, next two you choose, then you choose again for the next two

MSmits: here its choose after 3, so first 3 no choice

MSmits: that is, at the highest level

AntiSquid: smits do you have Bacalaureat like in france in your country? there's no such thing in UK afaik

MSmits: otherwise it's after 2

Astrobytes: Most of the time teachers advise students pretty well

MSmits: whats bacalaureat

AntiSquid: https://en.wikipedia.org/wiki/Baccalaur%C3%A9at

Astrobytes: The baccalaureate is only French (or maybe Canada? And maybe another few euro countries?)

AntiSquid: post-highschool exam

TBali: Abitur

TBali: in Germany

AntiSquid: bunch of EU countries have it

TBali: we have the same concept

jacek: matura?

MSmits: do you mean a full exam at the end of high school? or after?

Astrobytes: We had (in Scotland) 6th year studies, which was like university foundation level

Astrobytes: but that was a while ago

TBali: Now in Hungary it also acts as the Uni entry exam for state unis

Default avatar.png zhoubou: jacek Same as us

MSmits: we do have a full exam at the end AntiSquid, that encompasses about 70% of the entire curriculum. Is that what you mean?

MSmits: not all subjects have this. Computer science doesnt

MSmits: physics does

AntiSquid: "now" TBali ? wasn't it always like that ?

MSmits: physics exam takes 3 hours

MSmits: it's pretty nerve wracking for some students

TBali: there is even an international version https://en.wikipedia.org/wiki/International_Baccalaureate

MSmits: the grade is 50% of the total grade on your diploma

MSmits: the other 50% is all the stuff you did before the exam

TBali: AnitSquid - no, in my time I had separate Baccalaurate exam, and Uni entry exam 3 weeks later

MSmits: it's important to have such a national exam, to prevent schools from messing with the bar students have to pass. they can make it too easy otherwise

AntiSquid: cool, maybe i can go to Uni and do gender studies now TBali

TBali: My son did the USA Sat Physics Subject test just to see how it goes, but in the end he did not apply for USA univ

TBali: https://collegereadiness.collegeboard.org/sat-subject-tests/subjects/science/physics

MSmits: i dont understand why gender studies is a thing. What do you learn there?

MSmits: dont know if they have that here

AntiSquid: clearly you're intolerant?

MSmits: lol

MSmits: please enlighten me :)

TBali: I think it is overblown by politics, both sides

MSmits: i dont even know what it teaches

Astrobytes: As usual, real issues become polarising and divisive because of hype.

MSmits: hm ok, but what is the actual issue?

AntiSquid: gender studies battle inequality !

MSmits: != ?

AntiSquid: generate new genders, release genders on the gender stock market and pump up their values

MSmits: mmh i doubt thats what i tis

MSmits: that sounds more like genetic experimentation

MSmits: so is this part of a politics curriculum then?

AntiSquid: geez there are scholarly articles now trying to promote it lol sorry my sarcasm can't beat that

MSmits: i gave up and googled it

Astrobytes: There is real inequality, real hatred, real issues

AntiSquid: you probably learn how to be a smooth political talker with that degree, it develops proper bullshiting skills

Astrobytes: Squid subscribes to the media-led description of it

MSmits: there always is

MSmits: AntiSquid that's a common critisicm of social sciences in general. I don't think it should be limited to gender studies.. speaking as someone who just read about it on wiki for the first time

MSmits: dont have a very high opinion of social sciences myself, as a physicist

MSmits: sure it has value, but it also seems to be messy and often wrong

AntiSquid: social sciences is a great ideal

AntiSquid: but that's not what they're pushing

Astrobytes: Nor I as a biologist, but they exist and they exist to question

MSmits: who is they

AntiSquid: the ones teaching the curriculum for the gender studies degree i mean

Astrobytes: What about the gender studies degree do you disagree with?

MSmits: from what i read on the wiki, which again, is all i know about gender studies atm, it seems to use a lot of vaguely defined words

MSmits: that alone is problematic

PatrickMcGinnisII: I can't get an additional degree because i initially took physics w/ calculus and the new degree requires physics for engineers ... wtf is the difference?

MSmits: no idea

MSmits: but physics is a large field

Astrobytes: Society in general is problematic, why wouldn't studying it be problematic?

MSmits: for example, when people talk about "feminism" they have wildly different ideas about what it even means

AntiSquid: there being 50 genders or more part and many other mostly theatrical studies about social aspects of our world Astrobytes, that's what i disagree with

Astrobytes: Antisquid: that's not gender studies

MSmits: AntiSquid gender is a word that has apparently evolved to have a different meaning to some people. Noone would deny humans have 2 genders biologically

MSmits: the problem is that the definitions seem unregulated. That's probably what you have a problem with also

Astrobytes: It's more about the relation of genders to each other and the reasons for inequality: very broad subject. Same as studying race relations

MSmits: inequality of opportunity or inequality of outcome ?

MSmits: it's a hugely different issue

MSmits: people just call it inequality, but thats bad

Scarfield: FemstroBytes

Astrobytes: The 50 gender thing: yes there are a lot of (imo) total idiots around touting ridiculous stuff but that's the only part picked up by the media.

AntiSquid: i made a joke about the woke aspect of it, which is very real, didn't meant to go with chat completely offtopic to discuss it

Astrobytes: Both MSmits. It's a subject worthy of study is all I'm saying.

MSmits: woke is another term that is too vaguely defined and is mostly used to trigger/shame/punish/encourage people without being too specific about what you're saying

AntiSquid: even if you take just the male vs female part of gender studies you'll see it's partly pure rubbish

AntiSquid: or maybe mostly

Astrobytes: AntiSquid you read way too much alt media :D

MSmits: Astrobytes i agree it is

AntiSquid: that's an assumption

Astrobytes: Scarfield: LOL where have you been

MSmits: words matter though, they leak into society and take on a life of their own, that's the problem

MSmits: like the gender thing

AntiSquid: judgement is based on what i heard directly from the horses mouth

Astrobytes: The whole identify as a stonekin, 350 gender thing is just a meme

Astrobytes: now

Scarfield: just didnt feel like coding lately, but not the first break i have had :)

Astrobytes: That doesn't mean there aren't people out there pushing that though

MSmits: AntiSquid you do have to be careful about rightwing media. They sometimes have a point, but more often than not they will make false connections

Astrobytes: But it's not in mainstream gender studies that is for sure.

AntiSquid: the memes are there to point out the flaws, but discussing toilet signs is part of gender studies

MSmits: like when it's about gay rights, they start comparing to bestiality or pedophilia and such

AntiSquid: for example

Astrobytes: Scarfield: all good man, I know how it goes

Astrobytes: And: BREAKFIELD

TBali: Astrobytes, yo you are a biologist?

Scarfield: I honestly dont understand why its even being argued to be "genders"

Astrobytes: Yes TBali, well, by qualification at least.

Scarfield: yay, pun name, i have missed that xD

AntiSquid: MSmits you're also making assumptions

TBali: What area are you working in?

MSmits: rightwing media is obsessed with who should go to which toilet :P

Astrobytes: Pharma regulatory TBali

MSmits: arent we all AntiSquid

struct: Current media is trash

Astrobytes: MSmits: that's a real issue, for all parties concerned

struct: I cant even watch the news on tv anymore

TBali: Coding for work, or just for fun?

struct: its unbereable

Scarfield: i will make up a new gender, that by definition can only belong to one person, so private toilet

MSmits: it's an issue because rightwingers make it one is what i meant

MSmits: it shouldnt be

MSmits: who cares about that, just let people go where they want

MSmits: nothing ever happens

Astrobytes: TBali: I do a little for work (scripts to help me out etc) but the rest for fun yes. I did have 1 dev job back in 2000-2002;

AntiSquid: no MSmits, was specifically talking about shit you hear from gender studies crowd directly

MSmits: ah ok, i dont listen to the gender studies crowd, so i dunno. Maybe i twould be interesting

TBali: I am wondering, how much % of CG community is pr developer?

Astrobytes: MSmits: Sure, I mean that in some places people get a bit off with the wrong genders being in their toilets - potential for abuse.

MSmits: yes, they seem to think there is so much potential, yet you almost never hear about someone who pretended to be a different gender to spy on the toilets... that just doenst happen

Astrobytes: In Sweden, the toilets are mixed.

MSmits: its a right win fantasy

MSmits: +g

AntiSquid: you might have seen it in your work place but didn't pay attention to it or didn't realize the flaws

Scarfield: but sweden also doesnt use words for him/her anymore

Astrobytes: No, there are cases of abuse. But they are a lot rarer than those news outlets would have you believe.

MSmits: and those rare cases. Do you think a law would have stopped them

Astrobytes: Sweden's been cool about the gender thing for ... as long as I've ever visited there at least :D

TBali: him/her word I don't even know how am I suppose to use it 'properly' in English. 'They'. and then isse the verb in singular 3rd?

Astrobytes: Hell no MSmits. But it's about the illusion of safety.

AntiSquid: there's a push of 50% stats instead of skill for example

AntiSquid: that's not wrong?

Scarfield: they have gone a bot overboard imo, there is a difference between the genders. Trying to push that there isnt as some politically correct agenda just seems weird. But they arent hurting anyone, but us danes can joke about these weird swedes :p

Astrobytes: Scarfield: perhaps officially? But they most certainly do personally!

AntiSquid: also all the BAME politics is also part of it

AntiSquid: you heard about BAME in your work place right ?

TBali: Scarfield, you weork at Evil Corp, so that does not count :-)

TBali: Mr. Robot - what a great show, really liked it

Astrobytes: Yeah, but the fact it's official Scarfield, it makes a lot of people feel really comfortable. In everyday life/speech they use whatever but they are respectful

Scarfield: wait, they only added a gender less word.. lol i thought they "banned" the gendered words xD

MSmits: I dislike policing speech

Astrobytes: Scarfield: They also hide from their neighbours if they are about to both leave their apartments at the same time :D

Default avatar.png zhoubou: Hmm I might like this Sweden :thinking:

Astrobytes: MSmits: they don't police the speech

MSmits: then it's fine

MSmits: there was a big problem in canada when they policed speech about gender

AntiSquid: who doesn't police speech?

Astrobytes: Sweden is geek heaven btw

Scarfield: yea, i actually thought that was what they did, at least i looked it up

AntiSquid: you have to be very careful what you say these days :/

Astrobytes: Context squid, context. I was referring to the Swedish.

MSmits: AntiSquid depends on what you do mostly

MSmits: if you could get fired for what you say

Scarfield: also i think i saw some statistic that the swedish had the highest percentage of beautiful people to population

MSmits: sometimes it's good to punish people for what they say though... it depends. Cancel culture is real though

AntiSquid: diversity, inclusion ... doublespeak is all over sweden Astrobytes

struct: (╯°□°)╯︵ ┻━┻

Astrobytes: Now I KNOW you're reading those particular news outlets

Default avatar.png Anonymous12321: hi

Default avatar.png Anonymous12321: im

Default avatar.png Anonymous12321: learning python

AntiSquid: ya sorry struct, was just intending to do a quick joke, not start a debate

Default avatar.png Anonymous12321: is class of code for adults?

MSmits: what's wrong with diversity and inclusion?

Default avatar.png zhoubou: Lmao AntiSquid

struct: Wasnt related to discussion astro

struct: anti*

AntiSquid: https://www.tandfonline.com/doi/full/10.1080/10439463.2019.1611818 Astrobytes

AntiSquid: look at the god damn vocabulary

AntiSquid: you know what, i actually done a gender studies course

struct: That is longer than the mcts paper

Astrobytes: What about the vocabulary?

Scarfield: and doesnt teach MCTS,, pff

Astrobytes: *SWEMCTS

Scarfield: xD

Astrobytes: Ah welcome back EcksDeeField

Scarfield: :innocent:

Astrobytes: What's your next multi then Scarfield-a-tron

TBali: Hmm, I see TryAngle Catch is accepted

Astrobytes: Yes it certainly is.

TBali: could have been a great community contest

Scarfield: hmm not sure, failed hard at the last contest. Maybe ill finish that, but will probably look at something else first

AntiSquid: listed 2 words frequently used in their papers / studies / showcases you say i took it from right wing media ... Astrobytes

Astrobytes: You may now practice AI, Disappointment and 5% TBali!

TBali: might need to try this. But I have 24 untouched community multis on todo list

TBali: It is hard to motivate for community multis - only the best are there, so leaderboard is filled with good bots only

Default avatar.png Anonymous12321: im need help

Default avatar.png Anonymous12321: how find prime

Astrobytes: AntiSquid: didn't say right wing

Default avatar.png Anonymous12321: oof

struct: :tada:

struct: one person completed potw

Astrobytes: Oh nice

AntiSquid: do i need to quote chat history now :/

TBali: what is potw

struct: puzzle of the week: othello

TBali: I mean which is it now

Scarfield: prisoner of the wild

TBali: I still have my golfed submission there :-)

TBali: 86/156

Scarfield: xD

AntiSquid: you know what i am referring to and i am sure you are capable to see the issues yourself, but like to either play devils advocate or w/e

TBali: not bad from 126 bytes

Scarfield: "leaderboard would be modified to TrueSkill point PER source code chars."

Scarfield: sad that this hasnt been changed yet :p

TBali: C++ bots would fail miserably

TBali: Ruby would rule the world

Scarfield: hmm Pengwins or pokerchip race will probably be my next multi btw furryBytes

TBali: actually I am planning to try out MCTS is a fast language. (I coded mcts only once, for uttt and in php). Shall I try C++ or Rust? which might pose less difficulties for a beginner?

Default avatar.png duckymirror: I didn't manage to get a Rust version working without unsafe code...

TBali: I struggled with Rust vene for short solo puzzles

Default avatar.png duckymirror: Or rather, without unsafe code and little overhead

struct: Might be easier to find stuff on c++ than on rust, but dont quote me on that

TBali: I am a bit afraid of C++ memory management

AntiSquid: in what sense

struct: on cg just allocate and dont touch it

TBali: Making coder mistake, hard to debug?

TBali: Being inexpereinced with pointers

Scarfield: pre allocate a nodepool, assign nodes as the algo goes on

AntiSquid: pointer headaches are fun once you get used to it

TBali: Okay, but if my code is too fast than I get out of memory

TBali: so need to reuse

AntiSquid: then ascend to pointers referencing pointers of pointers for more fun

Default avatar.png duckymirror: My Rust MCTS doesn't run out of memory

struct: You either have a bug or dont use all the memory

Scarfield: you dont have to use pointers, have an int index for the firstChildNode and an int for the number of children for example

TBali: And you keep all created nodes intact?

struct: yes

AntiSquid: i love pointers, i am holding onto them ... i let go of recursion, that's already a big emotional loss

TBali: In php I usually turn off garbage collector. When I run 2048 and Samegame offline (for more time) I had out of memory...

AntiSquid: besides a friend of mine with 20+ years of C++ experience convinced me to use pointers

THEASIAN: hello

Scarfield: ooh, forgot about the 2048 multi.. i should make my own to do list

TBali: optim it is

Scarfield: yea

TBali: but offline result is so long that cannot fit in the source 100k hardcoded :-)

TBali: But with some encoding it still boosted my points

TBali: 2 bits per move, than base64

Scarfield: that is still something i havent touched yet, local search -> encoding, could be fun to do

TBali: (zlib is left out from CG's php...)

TBali: I played it jhust when php v8 with the new JIT apeeared - used for some benchmarking

TBali: And actually JIT makes a difference for CG-like workloads (not for web apps)

TBali: 2-3x speedup

TBali: but still slow

TBali: too bad CG is using 2 major release older version...

struct: how much code would break if they updated version?

TBali: not much imho

TBali: unless you use really not according to best practice

TBali: the type system became stronger but it is fully opt-on

TBali: opt-in

Scarfield: optim, heh heh ugh

TBali: not sure about C++ It is much better established, yet Rust seems shiny with the promise of memory safety

TBali: If I learn how not to fight borrow checker

Default avatar.png JBM: high-level C++ is memory-safe enough

Default avatar.png duckymirror: You get used to the borrow checker.

TBali: Funny thing . just made a C++ skill assessment on Pluralsight - it is completely flawed. It says "Expert". Flattering but far from truth

TBali: "Great start! You did better than 87% of your peers." Maybe lot of people do this as beginner. But that does not make me an expert

struct: expert = ready for easy puzzles on cg

TBali: Yeah I am for that. I know for, if, ...:-)

AntiSquid: are there other online assessments like that?

TBali: don't know. Pluralsight is subscription based

TBali: It was 20-30 quiz questions

TBali: Maybe I was lucky, with some educated guesses

TBali: lots of lambdas, objects, pointers

AntiSquid: does the assessment help ?

Scarfield: Which is the correct term: 1. If loop 2. For loop

TBali: But did it immediately after watching 5h of C++ tutorial

TBali: that helped

Default avatar.png JBM: Scarfield: yes

Scarfield: xD

TBali: loophole

Default avatar.png selmaDa: :neutral_face: how?

Astrobytes: gn all!

AntiSquid: bye bye

Scarfield: :wave"

Scarfield: lol

Default avatar.png JBM: gn8

struct: gn

Scarfield: gn

TBali: :wave:

AntiSquid: is there a quick contribution tutorial for lazy folks btw?

Default avatar.png JBM: sure

Default avatar.png JBM: "don't"

Default avatar.png JBM: contributions aren't for the lazy

Default avatar.png selmaDa: :joy:

TBali: or the faint hearted

TBali: :)

AntiSquid: i will keep it under WIP lol

struct: just the sdk tech io thing I guess

Default avatar.png JBM: yeah WIP scrubs it all

struct: https://www.codingame.com/playgrounds/408/tech-io-documentation/create-a-playground

AntiSquid: promise it will be less sloppy than your TVC

Default avatar.png selmaDa: can i ask you guys?

Default avatar.png JBM: like my tvc is sloppy

Default avatar.png JBM: it's got regression tests

Scarfield: less sloppy..? how could it be anything but less sloppy, TcV is perfection

Default avatar.png JBM: a proper MVC separation of concerns

TBali: is it approved now?

Default avatar.png JBM: not yet

Default avatar.png selmaDa: tcv better

Scarfield: god i cant type today

AntiSquid: i can throw something like that together, but no worries, want to do somewhat decent animations, got some funny idea

TBali: I cannot solve it. I am rooting for the troll. My soulmate. Cannot think of beating him

Default avatar.png JBM: in-jokes and easter eggs

Default avatar.png selmaDa: ok :slight_smile:

Default avatar.png JBM: a proper learning curve

Default avatar.png JBM: per-league upgrades and oodies

Default avatar.png JBM: *goodies

AntiSquid: learning curve? let me check me rank there again

Default avatar.png JBM: debug mode

Default avatar.png JBM: <blink>

AntiSquid: is it down ?

Default avatar.png selmaDa: can i ask for something ?

Default avatar.png JBM: i re-submitted it a few days ago

Default avatar.png JBM: it's up all right

struct: ask selmaDa

Default avatar.png JBM: or ask us

Default avatar.png selmaDa: can we make money from codinGame?

Default avatar.png JBM: but don't ask to ask

AntiSquid: oh i dropped a bit

Default avatar.png JBM: sure

Default avatar.png selmaDa: how?

AntiSquid: maybe the the learning curve theory is real for tvc

Default avatar.png JBM: what's your ranking AntiSquid?

Default avatar.png JBM: selmaDa: you tell us

Default avatar.png selmaDa: 3:joy:

AntiSquid: 43, was 11 without trying hard

Default avatar.png selmaDa: my level 3

struct: Go back to 2016 win the accountant contest

struct: profit

Default avatar.png selmaDa: 2016!

TBali: solved on D... lol

Default avatar.png JBM: :you_rock:

Default avatar.png selmaDa: me ? :upside_down:

Default avatar.png JBM: no, not yet for you

Default avatar.png selmaDa: ok :joy:

Scarfield: :you_paper:

struct: Only way to make money here is to win a contest I guess

struct: and even then the prizes might not be money

Default avatar.png JBM: i've seen people ready ot pay forsolutions in a specific language here already

Default avatar.png JBM: wouldn't match my rate, though

TBali: really? omg

AntiSquid: D next contest, no matter what, call me out if i dodge again, was lazy last time worried i need to google too much

TBali: And how you rate D?

TBali: 1 to 10?

Default avatar.png selmaDa: if is not money ? what's the prizes

Default avatar.png JBM: depends on the language

AntiSquid: idk, didn't try using pointers yet TBali

AntiSquid: it's alright i don't see why people hate it

AntiSquid: better than java for sure

Default avatar.png selmaDa: @JBM

Scarfield: Hoodies, posters, gift cards and such is what i remember

TBali: on a very novice level D is just C, no?

Default avatar.png JBM: i'm lost wrt D

struct: https://www.codingame.com/contests/finished

struct: selmaDa check there

Default avatar.png JBM: it used to be my "better C++" hope

Default avatar.png JBM: but apparently now it's GC and all the shit

struct: Unleash the geek had 3k amazon gift card as first prize

Default avatar.png JBM: a good thing for me rust came up

TBali: converted 30 puzzles, but did not hve too much impact on my

Default avatar.png selmaDa: yes perfect

struct: But that was a sponsored contest

struct: Nowadays im not sure what the prizes are

AntiSquid: because it's not outlandish syntax-wise

struct: beside a tshirt for top 20

Default avatar.png JBM: more like top3

AntiSquid: canvas to 3 ?

Scarfield: you could sell the tshirt for profit

AntiSquid: and shirts top 20

Default avatar.png selmaDa: does there are more in 2021?

TBali: everlasting glory for top10

Scarfield: when it eventually arrives

struct: "May 6" selmaDa

struct: But it may change

AntiSquid: ML contests give good prizes and some of them don't have strong competition

Default avatar.png selmaDa: why ?

struct: I dont know

struct: Sometimes it gets delayed

Default avatar.png selmaDa: ok

Default avatar.png selmaDa: What did you gain from your presence here?

struct: knowledge

Scarfield: enemies and bald spots

Default avatar.png selmaDa: :thumbsup_tone2::joy:

TBali: fun, trolls

TBali: castels

Default avatar.png selmaDa: me nothing

AntiSquid: that's a controversial topic selmaDa

Default avatar.png selmaDa: i khnow

TBali: gender studies :-)

Scarfield: xD

AntiSquid: i have a certificate for it

TBali: That's a strange book: http://leanpub.com/thinking-functionally-in-php

Scarfield: but seriously learned a lot, and had fun doing so

Default avatar.png selmaDa: heyy!

TBali: Not in the spirit of "right tool for the rtight job", but still interesting

Default avatar.png JBM: dat book titel

TBali: I bought it

AntiSquid: selmaDa what matters is what you want to gain, focus on that

Default avatar.png JBM: so how's your train of thought now

TBali: read it 70% already

Default avatar.png JBM: if you could think functionally in php, you wouldn't code in php to begin with

Default avatar.png JBM: change my mind

TBali: Well, it was interesting

Default avatar.png JBM: yeah

TBali: My mindset is not functional.

Default avatar.png selmaDa: i want to learn programming and earn money !

Default avatar.png JBM: it's made you like FP

Default avatar.png JBM: and want to stop PHP

Default avatar.png JBM: goals achieved

AntiSquid: he could have read a book about stock exchange in that time and now be coding a bot for trading, right JBM ?

Default avatar.png JBM: could've

TBali: I solved a few puzzles in Haskell, it was really nice

TBali: found quite hard to change mindset

TBali: and my codes are most likely ugly

PatrickMcGinnisII: besides the speed, what is wrong with php?

AntiSquid: $$$$

Default avatar.png JBM: i don't have all night

TBali: Read part if this, http://learnyouahaskell.com/starting-out

AntiSquid: we want bitcoin symbol

Default avatar.png selmaDa: :person_frowning_tone2:

TBali: Even saw 4h of haskell tutorial on Pluralsight

TBali: But still hard to think of a problem in FP way

PatrickMcGinnisII: i make $ with php

Default avatar.png selmaDa: @JBM are you freelancer

TBali: Okay, than I a worse - I do php as hobby...

AntiSquid: not sure anymore i finished the book learnyouahaskell, probably skimmed through last few chapters

Default avatar.png JBM: selmaDa: what's your budget?

TBali: I did not finish

TBali: just the basic chapters

Default avatar.png selmaDa: 0$:slight_smile:

TBali: what is that in €?

AntiSquid: hired, what are your skills selmaDa ?

Default avatar.png selmaDa: programming and desing

Default avatar.png JBM: who doesn't like to dese in the weekends

AntiSquid: dresing ?

Default avatar.png selmaDa: yes

TBali: dressing? what skill is that?

Default avatar.png selmaDa: desing

Default avatar.png JBM: something inaccessible to us programmers

Default avatar.png selmaDa: not dresing

AntiSquid: he's probably playing visual novels where he has to dress up the grill

AntiSquid: or she ?

TBali: they

TBali: :)

AntiSquid: ow

Default avatar.png JBM: pump us the jam; dress up the grill

PatrickMcGinnisII: i use php to easily make custom PDFs on the fly from session & cookie data and email it. What would it take for me to do that easily in another lang?

**AntiSquid takes cover

PatrickMcGinnisII: sql as well

TBali: Sure, I do like PHP, while seeing its shortcomings

Default avatar.png JBM: “easily”

Default avatar.png JBM: “on the fly”

Default avatar.png selmaDa: :repeat_one: can we build app for free

AntiSquid: really depends what you're used to imo PatrickMcGinnisII

Default avatar.png selmaDa: but online for lasy :joy:

Default avatar.png JBM: oh, and “custom”

AntiSquid: you can do that in python @_@

TBali: Well it was not planned properly, but eveolved naturally..

TBali: But at least moving in the right direction in past 5 releases

PatrickMcGinnisII: well i don't use .asp ... i kinda hate .asp - everything i made with it ran slow too

PatrickMcGinnisII: the php is easier

Default avatar.png selmaDa: good night :wave_tone1:

AntiSquid: tell us your gender first

PatrickMcGinnisII: i guess i should be an azure addict now?

Default avatar.png selmaDa: girl

TBali: In CG, there is no real difference to more "modern" langauges, except speed

AntiSquid: ok good night

TBali: all constructs there, almost

Default avatar.png selmaDa: you too

Default avatar.png selmaDa: see you tomorow

TBali: bye selmaDa

AntiSquid: get and avatar btw, don't be like that high level lurker

PatrickMcGinnisII: rofl, a girl shows up and yall lose your ... nevermind

therealbeef: allegedly

PatrickMcGinnisII: I just wrote 100 lines for euler's TAC game and it still won't do anything real

TBali: C++ seems to have more "quirks" now, than php

TBali: was it 100 loc docblock?

PatrickMcGinnisII: well, i have class now. :P

TBali: good start :-)

PatrickMcGinnisII: tbali , you try chess with php? i can't get a minimax to run past depth 2

TBali: In all the multis, I make nice classes, etc. Very readable. Than I realize the sim speed will suck

PatrickMcGinnisII: i dunno if using alpha beta is even worth it

AntiSquid: feel free to identify as a girl to fill up the void or if you think it gets you more attention lol

PatrickMcGinnisII: there r some real cool women on here, they just don't show up much

TBali: did not try chess yet

AntiSquid: doubt CG has "girl" roleplayers, there are no ingame items to trade

TBali: and after I realize speed sucks, I go for heuristics -

TBali: Don't know female % of CG community. my prejudice is it is fairly low

PatrickMcGinnisII: i think turning off the gc asnd even error reporting help alot

PatrickMcGinnisII: i started dropping data into static classes and it does run a little faster, dunno why

TBali: gc - I do that. Error_reporting? Ebven if there are no errors?

TBali: some checking gets turned off by that?

PatrickMcGinnisII: i think there is some try {} catch that gets bypassed

TBali: I like to use type hinting - even if it actually slows down beacuse of extra runtime checks

TBali: OK, I will try out that one next time.

PatrickMcGinnisII: oh i use it for validating code... but turn off for submit

TBali: I also usually use declare(strict_types=1); which might have perf penalty


TBali: but we do not do this for speed, don't we :-)

TBali: how can you turn of type hinting ?

TBali: you either declare it or not, no?

TBali: (I mean function arguments, return types)

TBali: from 7.4 class methods could have been also type hinted

TBali: I mean properties

PatrickMcGinnisII: chess can be done without classes, but i did it with classes

TBali: I muss struct

TBali: miss

PatrickMcGinnisII: the board can be modified with a history stack, which i do...but the tree is too big at depth 3

TBali: classes help so much to organize code, & data. But really slows down things

PatrickMcGinnisII: struct, lol still used in C

PatrickMcGinnisII: well using intval() versus using (int) , i dunno if there is a speed thing to it

TBali: do you use bitboards for chess?

PatrickMcGinnisII: no i didn't, just a linear array of 64 values

TBali: I found replacing bitwise operators with arithmetihic usually helped

TBali: In UTTT

TBali: I mean vice versa

TBali: it is too late...

TBali: actually a simple vector type would help for php speed. Array is jjust too verstaile

PatrickMcGinnisII: thing is with these games you have 3 states for each square or cell...enemy piece, no piece or your piece ... So there's so many compares (ifs) that anything in a loop gets slowed way down

TBali: using sinternally a sorted hashmap with zvals as keys and as values is just so overkill if I need just a fixed list of int

PatrickMcGinnisII: I try to use bitboards with xor operations, but its like re-inventing the wheel

TBali: Might be, chess is complex.

TBali: For Tic Tac Toe, I found good bitboard ideas on the net

PatrickMcGinnisII: I want to use splfixedarrays, but the functions are limited

TBali: implemented it

TBali: And I read that splfixedarray is actually not that great in speed

TBali: (too old)

PatrickMcGinnisII: priority queues are impossible speedwise on CG

TBali: I used for some puzzles.

TBali: A* requires it

PatrickMcGinnisII: well u figure out any more speed stuff, plz let me know

PatrickMcGinnisII: gonna go eat some buffalo wing cauliflower

PatrickMcGinnisII: glhf

TBali: :)

AntiSquid: that's a new one to me ... next up pork wing cauliflower

Default avatar.png shuo502: h

Default avatar.png Anonymous12321: h

MadKnight: h

Recal: i

Default avatar.png Birdington: hi

Default avatar.png Joselillo: Hi to everyone

Default avatar.png Thirumalai15: hello

Default avatar.png Thirumalai15: any php devs

ErrorQwerty: ew

ErrorQwerty: e

Default avatar.png shuo502: hello any py devs

Kepax: sexy

Aayush.Curious: Hey guys; anyone here interested to be in a hackfest?

Aayush.Curious: It's in Frb.

Aayush.Curious: Feb*

Aayush.Curious: And..I want a really good coder for my team..