From CG community
Jump to navigation Jump to search

Default avatar.png ChampionCoder: One question: In the Mars Lander Episode 2, is the coordinates of the landing ground given in the input?

Unable to find it: which is why I'm asking

FlamingComet_9d13: subscribe Evan lin-- gaming

FlamingComet_9d13: you will have a free Nintendo Switch if you subscribe me on youtube

_O-MEGA: for sure no scam :)) free nintendo lmao

FlamingComet_9d13: quick subscribe me on youtube, I wanna see it, lol

jacek: merry Caturday

VizGhar: How come, there are still more moderators online then common users_

mzbear: that user list isn't reliable

jacek: smits finally winning some games vs remi

eulerscheZahl: VizGhar because you only see 2 types of users in the list: moderators and active chatters while moderators see all the active users

eulerscheZahl: see the scrollbar on the right

eulerscheZahl: external clients also show the full list

jrke: damn smits bot !!!

jacek: oO

MSmits: :)

MSmits: just fixing some bugs

MSmits: hope that will get me > 50% vs remi

MSmits: otherwise picking the right randomseed will

jrke: what you did btw

MSmits: he's very deterministic

MSmits: the right randomseed will be like an opening book in 1 number :P

jrke: i think some minmaxing at few steps before endgame to get good gamestate

MSmits: not exactly, but yes, a few steps before the endgame i pick the lines that give me control over the endgame

MSmits: not by minimax though

jrke: ah ok

MSmits: I use Nim theory

MSmits: as I am sure Remi is, in some way

MSmits: my first 50 plies are still random

jrke: i am writing endgame solver its breaking some where (timeouts) but don't know why

MSmits: which, in combination with the winrate, shows you that noone has any idea how to play the early game

jrke: yes i see

MSmits: hm

MSmits: well timeouts can be caused by anything

MSmits: it should be similar to minimax though, except it's not depth limited

SupSon: strings and coins

MSmits: thats a dual representation of D&B


MSmits: my bot printed as strings and coins

MSmits: board

MSmits: a star is a coin (or a box), lines are strings instead of sides

MSmits: instead of setting lines, you cut strings

MSmits: can be helpful sometimes to gain some insight

MSmits: but otherwise it's just a way to print the board

MSmits: end game solver is fun, but jrke, if you want to win significantly more, endgame solver is not enough. It only helped me a tiny bit compared to other players

MSmits: the nimstring stuff is more important

jrke: so didn't you used your endgame solver in this bot?

MSmits: yeah its in there

MSmits: it's 1500 lines out of 3k total :P

MSmits: the nimstring search is like 500 lines

MSmits: the rest of the bot is other stuff

jrke: i am choosing smallest chain and playing it and reverting last 3 moves and then i am minimaxing them and for its eval i again choose smallest chain and it continues

jrke: is it correct for solving endgame^ and getting maximum values gained by me and opp if no one do mistake

MSmits: there's also loops

MSmits: often it's better to open loops first

MSmits: because to control them, you leave 4 boxes

MSmits: so if the opponent wants to keep control he has to give you 4

jrke: i am minimaxing last 3 reverted moves so it should decide to keep control or not

MSmits: there is no one size fits all heuristic here. It depends on the exact combination of chains and loops

MSmits: ah ok

MSmits: well if you minmax it all, then it will be ok

MSmits: but this way might be extremely expensive

MSmits: for some endgames

MSmits: there could be 10-15 such choices to make

MSmits: you need lots of pruning

MSmits: meaning heuristics to choose what options to even consider

MSmits: also, you say "smallest", but what do you do when several have the same size

MSmits: the way they are connected by joint, might have an impact on how the game continues

MSmits: merging some chains etc.

MSmits: well... good news. my CG benchs says 80% vs remi

MSmits: bad news, i see a horrible bug sometimes and i dont know what it is =/

jrke: congrats 1st then

MSmits: might not get it on first submit, he has been pushed a lot

jrke: try 2-3 times then

MSmits: yeah

MSmits: allright, have to feed my daughter :p

MSmits: brb

jrke: bye

Default avatar.png zhoubou: What kind of thing are you guys doing? 3k lines??

Default avatar.png zhoubou: I'm new to this :P

jrke: some search + some heuristic stuff

Default avatar.png zhoubou: Is it tied to CodinGame or some outside project?

jrke: dots and boxes bot


Default avatar.png zhoubou: Oh, I see. Seems there are beasts awaiting me in the future. Here I am battling 100 lines of crude code.

MSmits: 3k lines is insane, most of my bots are barely over 1k

jrke: my average lines atm are 500 from lat 3 months it became 800

jrke: looks 1 submit is enough MS

MSmits: maybe

jacek: just keep hammering

MSmits: will do :)

MSmits: the games i lose vs remi is when in the 5-10 plies before the endgame i have 0 options to change the outcome of the nim value calculation, meaning he has control

MSmits: often he leaves me an option though

MSmits: he has some probabilistic approach, maybe mcts, to get control

jrke: another 50 pointer

jacek: like another over 50 years old?

jrke: 50 points

MSmits: only in spirit jacek :)

k4ng0u: lose first game to keep some suspens then win the next 8 haha

jrke: new first ranker

MSmits: looks like made it, only 1 lost game lol

jrke: yes

jrke: gtg bye

MSmits: bye

jrke: i will be fixing my bot and i will be there

jacek: time to pm remi

jrke: if i got succeded in my bot

MSmits: lol

MSmits: you mean you want me to pm, or are you?

jacek: yes

MSmits: you have a short line to the great Remi?

MSmits: or maybe he will get an automail if you pm him on the forum

jrke: i have like (my lifespan - 13) years to complete my bot lol

MSmits: you do

MSmits: but as i said

jacek: well there is also his twitter

MSmits: you need to focus on the math here

MSmits: this is one of the few games that is about the math

jacek: and he made d&b for phones

MSmits: wow

MSmits: no wonder he is that good

jacek: and jrke is on the screenshot

MSmits: he is really invested in this game

jacek: "Hmm, just gives me a black screen on my iPhone. And an ad. Disappointing"

jacek: hmm

MSmits: lol

MSmits: maybe a troll

jrke: now we need new league in D&B too much in wood 1 atm

MSmits: did you ever check the csb leagues?

jrke: :smiley:

jacek: make new league and smits bot as boss

jrke: a league of 1 person lol

SupSon: need a batman signal for remi

MSmits: I still have 3 more ideas to improve bot though

MSmits: bit tired of looking at this 3k monster atm, but if he comes back, so will i

jacek: :scream:

jacek: so what now? b? Y? something else?

MSmits: well i was gonna rewrite uttt again until someone pushed me to try D&B

MSmits: forgot who it was, maybe jrke

MSmits: but not really feeling like uttt

MSmits: i really want to do bt, but i am a bit out of ideas

MSmits: i can totally see why darkhorse got stuck on his own game there

jacek: oh im out of top10 in uttt :c

MSmits: he thinks like me

MSmits: maybe bit less obsessed

jacek: i am obsessed with bt

MSmits: yeah, well i would be too if i knew how to play it well

MSmits: it's a nice game

jacek: even before NN, id optimize piece-square tables and other params

MSmits: oh so you're saying i just gave up too soon

jacek: if youre not 1st, its always too soon

MSmits: good point

MSmits: not sure yet though, i put off a long list of work i am supposed to do during this christmas break

MSmits: so maybe i'll just let my two meta mcts work while i do actual work

jacek: yeah, heating the house

MSmits: exactly

SupSon: get bopped jacek

jacek: hm?

MSmits: thats just how he says merry christmas

SupSon: bopped you outta top 10

MSmits: ohhh

MSmits: you bopped re curse above me also

MSmits: i dont mind, he needs the boost of confidence

MSmits: :P

YurkovAS: has public info about progress in top10 uttt? some new heuristics or mcts improvements...

3569066: Oo

MSmits: not much has changed

MSmits: it's just shifts due to people submitting

MSmits: and supson improved his bot i think

YurkovAS: ok, thanks

YurkovAS: +-3 month ago amurushkin find great mcts improvement (+5-15%), he told me by secret. And my bot started win Nagrarok.

MSmits: oh ok, cool

MSmits: yeah there are a lot of little things like that

MSmits: but sometimes they dont combine well

MSmits: so A works, but not if you also do B

MSmits: B works, but not if you also do C, etc.

SJ_01: can someone give a source for learning Trilateration please?

MSmits: no idea what that is

MSmits: bilateral is 2 people talking

MSmits: maybe if we find another person, we can have a trilateral

darkhorse64: I guess it's for shadows of the night 2. You have to explore x direction then y direction. Also look at the graphics, they explain which part of the building you can drop.

MSmits: aha

MSmits: well gonna go to celebrate christmas for most of the day

MSmits: dont steal my new nr 1 while i am out

Default avatar.png ChampionCoder: Is there the landing coordinates of the Mars Lander-Episode 2 given?

eulerscheZahl: you are 2 days late

Default avatar.png ChampionCoder: Not sure if it is

darkhorse64: There is a Connect4 in the WIP, I am writing a bot for it. You can do bitboarding, books, mcts minimax

MSmits: we have 2nd christmas day for some reason

MSmits: i guess thats not international

eulerscheZahl: have have a public holiday today as well

eulerscheZahl: we have*

MSmits: ah ok

Default avatar.png zhoubou: Same

Default avatar.png ChampionCoder: Is there the landing coordinates of the Mars Lander-Episode 2 given?

Default avatar.png ChampionCoder: Not able to find it if it is

MSmits: you need to land on a flat area

MSmits: any

eulerscheZahl: 2 consecutive points with same y

Default avatar.png ChampionCoder: Oh so no coordinates?

MSmits: nope

Default avatar.png ChampionCoder: Thanks will try that

MSmits: allright thats it, ttyl guys

eulerscheZahl: bye

johnytep24: hi

VizGhar: yeey made almost 400k score on 2048. Reaching that 600 commands limit with BFS... Now I need to switch to beam and solve the limit problem :)

Default avatar.png zhoubou: Congrats!

VizGhar: tnx

Default avatar.png zhoubou: I went for the simplest solution just to solve it without going Up all the time. I'll need to solve a lot of easier puzzles to practice first...

VizGhar: yeah... Probably the best way


alenic: Hi guys, in optimization puzzle, is it possible to hard code a series of moves computed offline, without writing any algorithm ? (only when the test coincides with validation) Or is it frobidden ?

eulerscheZahl: for some optim puzzles that hardcoding is pretty common at the top ranks

eulerscheZahl: 2048 is probably the only one where you can't fully do it (as there are so many actions and you have a code size limit on codingame)

Default avatar.png zhoubou: Miles-M I solved 1D spreadsheet using 2 simple recursive functions.

alenic: thank you eulerscheZahl, you hit my problem :), I'm thinking about an hybrid solution where the first moves are hard coded and the rest is a tree search. By the way, nice puzzle ;)

VizGhar: if you have already slved it, ypu can see others solutions

alenic: I focused too much on a compression algorithm than a good tree search :joy:

eulerscheZahl: thanks alenic and also thanks for not stopping after the 100% and completing the puzzle quest

eulerscheZahl: zip your path and encode it (base64 for starters, other encodings with unicode are more efficient)

alenic: I used a bit packing (is more efficient than zlib) with a base85 encoding, I reached a 3.7x compression

VizGhar: Yeah main criticis was on "how easy this 2048 is" :P

eulerscheZahl: and if you do some cheap heuristic for your next action and only compress those moves where your heuristic is different?

alenic: nice idea ! I will think about it...

eulerscheZahl: evil me used 2048 to farm language achievements

alenic: ehehe, I'm only at 11 million score, 62 million is very insane...

VizGhar: :D thanks for recommendation eulerscheZahl finally will reach silver on ocaml :D

eulerscheZahl: A*Craft is also great for language achievements

k4ng0u: on 2048 I feel like it's fine to get a good score with a beam search but to reach the top it's actually pretty hard. I don't know how top players do, but they either need a real good eval function, or game state hash or they use a totally different algorithm

k4ng0u: overall my feeling is that beam search usually provides an better than average solution but rarely the best :p (but maybe I am missing some heuristics or good eval)

wlesavo: k4ng0u all top players beam search with snake eval, maybe apart from dbdr

alenic: maybe top palyer use some neural network trained offline used for example to evaluate a grid ? (like Q network)

VizGhar: I wonder how to evaluate snake properly :) I'm only counting how long is the snake and I'm pretty far

VizGhar: Also I feel like I'm making too much operations just to evaluate the board. Everything else looks fine

k4ng0u: wlesavo what is snake eval?

VizGhar: you are trying to keep tiles in descending order in "snake-like" shape

VizGhar: so better is to have 2048-1024-16-4 in first row then 2048-512-1024-16

wlesavo: k4ng0u basically sum of tiles witth discount in snake order, you can use that as hash also, it is not the best way but works for me

k4ng0u: thx VizGhar and wlesavo for the tip, I will try it when I have the time, it'll certainly better than the raw score eval I currently use

k4ng0u: by the way wlesavo, I see you are participating in raic. how would you grade the competitiveness compared to codingame contests?

wlesavo: k4ng0u top200-300 close to cg silver, but finals aka top60 close to cg legend, although it is quite harder to start, since you have to deal with the api they provide. overall it is not much harder but very much different

wlesavo: first raic for me, but cg prepared me well enough

k4ng0u: I read the rules quickly, how do you handle the max time per game? do you allocate time to first time and then divide the rest by the max nb turns afterwards?

eulerscheZahl: you have 40s for the total game (1k ticks) and max 1s per turn

eulerscheZahl: so there is that "time bank" that we are missing on CG

Astrobytes: very different but great fun

eulerscheZahl: and all submits are public, you can always play vs older versions of opponents. and even test matches are public, which makes hiding your bot less unfair

eulerscheZahl: of course you can still hide. at the cost of not testing vs others in the arena

eulerscheZahl: i would agree that finals are similar to CG legend

eulerscheZahl: the tshirt on the other hand is really easy to get if you have 1 or 2 afternoons to invest

Astrobytes: ^

wlesavo: k4ng0u also this one is more heuristic, so with faster languages you basically dont care much about time. i think some tops do, but you can get quite far even without custom pathfinding

k4ng0u: ok I see, very interesting

eulerscheZahl: but you just missed the time for RAIC

eulerscheZahl: maybe there will be a contest on otherwise you have to wait until next year

eulerscheZahl: is all Russian (rules, website, ..) but you are still allowed to play and win prizes

k4ng0u: yep I know :( I discovered it by chance on the chat the other day. I am just looking at the ranking, and it seems like there are a lot of current/previous codingamers participating to it and not doing too bad!

eulerscheZahl: yeah, the CG raid was efficient this year ;)

jrke: i discovered CG this year and started hackerrank a month ago nothing else

eulerscheZahl: hackerrank is pretty much dead

eulerscheZahl: they had contests with real prizes

eulerscheZahl: but now these very rare contests for limited groups only

eulerscheZahl: like for women in the US

jrke: lol

Default avatar.png zhoubou: What is raic?


Default avatar.png zhoubou: Oh, thanks!

leojean890: RAIC games are complex compared to CG. amurushkin recently told that all previous RAIC contests will be soon released as multiplayer games.

eulerscheZahl: cool, I didn't know that

Astrobytes: that will be awesome

leojean890: he told that on #RAIC2020

eulerscheZahl: too bad that many previous bots are public

Astrobytes: yeah I thought I saw him mention it

leojean890: it will be interesting

Astrobytes: Didn't somebody say something about moving to new servers too?

eulerscheZahl: i read about the current one remaining active

eulerscheZahl: missed the info about previous ones

leojean890: maybe, and would be good news, because there are too few games in my opinion^^

eulerscheZahl: next: RAIC SDK

eulerscheZahl: create whatever game you couldn't do on CG because of server limitations

leojean890: previous contests seem to be very hard.. 3D collisions :)

eulerscheZahl: in 2018 they gave a lot of code at least

leojean890: to make some football with cars :p

eulerscheZahl: i played that

leojean890: are you saying that many codes are public ?

leojean890: oh ok, I think that it was hard, no ?:)

eulerscheZahl: yes, most of top10 shared their bots on github

leojean890: arf ok

leojean890: too bad :s

eulerscheZahl: there's even a repo collecting those links

leojean890: so not only PMs but full code :s

leojean890: (on CG it happens as well..)

leojean890: but not often

eulerscheZahl: even I shared some code

eulerscheZahl: because RAIC is over when it's over

leojean890: ok I see^^

eulerscheZahl: no one could expect a revival

leojean890: because it was not planned for a long time that multis would be released..

Astrobytes: it will be nice to have something to play outside of CG. Not that we are lacking games here :D

eulerscheZahl: RAIC is just different

Astrobytes: It was a refreshing change.

eulerscheZahl: finally no guesswork about missing input

Astrobytes: :D

Astrobytes: The API is a little odd, but you get used to it soon enough

leojean890: different and more complex so different challenges:)

leojean890: the API is not very intuitive, yeah :s

eulerscheZahl: CG is like: there's a unit. I won't tell you the current HP

Astrobytes: hahaha, so true

eulerscheZahl: RAIC is like: and the unit has green pants. not sure if you have to know but i tell you anyways

Astrobytes: Totally on point :D

leojean890: there are too much info I think indeed:p

leojean890: like in CG "BOTG"

Astrobytes: Nah it's not *that* bad :P

eulerscheZahl: BOTG lacked rage info

eulerscheZahl: or however it was called

leojean890: rage ?

Astrobytes: rage was MM

leojean890: ah MM yeah:)

eulerscheZahl: when minions attack your heroes and not your minions

leojean890: I didn't understand the rage concept:p

leojean890: in MM

leojean890: oh ok

leojean890: I did BOTG multi in 2 hours so I don't know very well this topic^^ but it was.. complex!

eulerscheZahl: I meant aggro

eulerscheZahl: not given in input

eulerscheZahl: squid is offline, so i can say it: i hated BOTG :P

Astrobytes: I gave up on botg after all those wood leagues. It certainly aggravated me

Astrobytes: How that was a contest is beyond me in all honety

Astrobytes: *honesty

eulerscheZahl: they even removed 2 more heroes

Astrobytes: Still, each to their own I suppose :shrug:

eulerscheZahl: but it's a lesson learned, there won't be such a contest in the future or took place after

Astrobytes: The concept is nice, if not the game itself.

eulerscheZahl: remove some aspects and it fits better to CG

eulerscheZahl: hero selection, items, potions, gold, minions, tower, aggro, groot, bushes

eulerscheZahl: just too much for CG expectations

eulerscheZahl: and I forgot to mention hero skills

leojean890: I didn't struggle with woods, use one liner I think ? but bronze was harder and didn't make it fast, maybe I will go on one day

Astrobytes: Yeah, it was easy to pass but I was totally bored by the time I got to bronze

eulerscheZahl: maybe i'm a little too negative about the game as it was unfinished when the contest started

eulerscheZahl: so that's my first impression and then it's hard to change my mind

Astrobytes: Maybe someone will revisit the concept some day

struct: Well I dont really have an opinion about it, I havent really played it

eulerscheZahl: hell no

eulerscheZahl: and also no to locam 2.0

Astrobytes: :D


jacek: locam 2.0?

Default avatar.png yazilimogrenyegen:

struct: stop spamming

eulerscheZahl: there is a locam version with 2 lanes

struct: use #clash channel

Default avatar.png yazilimogrenyegen: sorry

Default avatar.png yazilimogrenyegen: it is my failiure

struct: np

Astrobytes: the link is on a Cats website I think

eulerscheZahl: and on discord

eulerscheZahl: oh, the channel got deleted

leojean890: I don't want to do LOCAM multi:)


jacek: oO

Astrobytes: thibaud is streaming?

eulerscheZahl: indeed

jacek: eeyup

eulerscheZahl: "Watch me fail at easy puzzles"

Astrobytes: lol

eulerscheZahl: and he has those annoying follower animations

Astrobytes: what annoying follower animations?

eulerscheZahl: animation shown on the whole screen

eulerscheZahl: with the name of the follower

Astrobytes: Oh. II haven't followed and didn't see

eulerscheZahl: someone else just followed when i opened the stream

Astrobytes: we should really be in there trolling him :P

eulerscheZahl: i think he did ?disableChat

eulerscheZahl: but twitch chat might work

Astrobytes: yeah I meant the twitch chat

eulerscheZahl: spoiler: it was the gardener

Westicles: Ooh, his mom just barged in with the vacuum cleaner and they are yelling at each other

Default avatar.png aboudehhub: seks

jacek: oO

jacek: AutomatonNN do you seks

AutomatonNN: eulerscheZahl is there a way to see a game really nice to see the code and come back after a contest

eulerscheZahl: false alert, no one pinged me

jacek: AutomatonNN is that so

AutomatonNN: and I don't know what that is

Default avatar.png zhoubou: There is the answer

Default avatar.png JavaSuck: yo does some1 got covid right now ?

if__name__.__fox__: I bet many do

Default avatar.png Toby75: Hi, has someone time and effort to test something together with me? I wrote a code, that I tried out anywhere and it should work, but the test cases still show errors...

Default avatar.png Toby75: It's the Temperatures challenge written in JS

jacek: test cases or validators?\

Default avatar.png lolo460: me confuse

Default avatar.png Toby75: sorry Validators

Default avatar.png Toby75: [SPOILER ] paste this into the Temperatures example, and view the validators. They say that I must print the positive one, if there is a positive and a negative number equally to zero. Now test it with node or on a web page with a positive and negative number, and you will see, that only positive numbers are getting printed.

Default avatar.png Toby75: So what the Validators say makes no sense

Default avatar.png Toby75: And before I did it another way, there the test cases showed me 100%, but not the validators

BrunoFelthes: man, you misunderstood the problem

Default avatar.png Toby75: what is the problem?

BrunoFelthes: only it the negative and the positive are equal, you must get the positive

BrunoFelthes: *if

Default avatar.png Toby75: That's what I did

Default avatar.png Toby75: When the array is [10, -3, 7, 3] I would need to print 3

Default avatar.png Toby75: That's what my code does

jacek: what if its [10, -3, 7, -3]

jacek: will it print -3?

BrunoFelthes: you even dont do any comparation without the Math.abs..

Default avatar.png Toby75: Then it prints out -3

Default avatar.png Toby75: yes

jacek: hmm

Default avatar.png Toby75: I would not have called this out if I was not sure that there is a big problem

Default avatar.png Toby75: I'm quite sure that my code should work, sadly I can't see which test cases have failed

Default avatar.png Toby75:

jacek: it failed 4th test case

Default avatar.png Toby75: i know

Default avatar.png Toby75: but why??

Default avatar.png Toby75: I tried out the array [10, 5, -5, 7, 11] AND [10, -5, 7, 5, 11] in node, and it always printed 5

Default avatar.png Toby75: so it also does not matter which one is first, the negative one or the positive one

jacek: [42 5 12 21 -5 24] it fails, dunno why yet

Default avatar.png Toby75: I just checked that one in node

Default avatar.png Toby75: exactly that one

Default avatar.png Toby75: it prints 5

Default avatar.png Toby75: not -5

Default avatar.png Toby75: that's why I am so confused

Default avatar.png Toby75: just lmk if you find out what happened lmao

jacek: mhm

Default avatar.png Toby75: got it

Default avatar.png Toby75: was really an issue with codingame

Default avatar.png Toby75: their node version does not support array.includes

Default avatar.png Toby75: but it did not show me an error

jacek: huh

Default avatar.png Toby75: the weird thing it, that every other node runner supported array.includes

Default avatar.png Toby75: I tried it on several websites and on my own node on my pc

Default avatar.png Toby75: that one used to work, instead of the .includes

Default avatar.png zhoubou: My browser supports `includes`, but it still returns -5

Default avatar.png Toby75: That is so weird

Default avatar.png Toby75: all pages like showed 5 not -5

Default avatar.png zhoubou: Ohh I see the error

Default avatar.png zhoubou: Inputs is an array of strings

Default avatar.png zhoubou: Not ints

Default avatar.png Toby75: bruh

Default avatar.png zhoubou: Of course it doesn't find -5 in [..., '-5', ...]

Default avatar.png Toby75: thought that would not make a difference

Default avatar.png Toby75: +"" would have fixed all

Default avatar.png zhoubou: :)

Default avatar.png Toby75: So just newer versions of node dont't care about types in the .includes

Default avatar.png Toby75: that's why I could not find that out on any other pages

Default avatar.png Toby75: yeah, nvm at least I know that now :joy:

jacek: this one would work


Default avatar.png zhoubou: It doesn't work in

Default avatar.png Toby75: it does for me lol

Default avatar.png Toby75:

Default avatar.png zhoubou: It doesn't. You have the wrong inputs

Default avatar.png zhoubou: Inputs is an array of strings in the puzzle

Default avatar.png Toby75: oh now I get what you mean

Default avatar.png Toby75: that's probably the problem when coding without type definitions

Default avatar.png zhoubou: I'm proud of my 108 characters Python solution, and I'm not even a golfer :P

Default avatar.png Toby75: xD

Default avatar.png Toby75: python is just on another level

Default avatar.png zhoubou: Ikr

jacek: mhm

Default avatar.png zhoubou: Gtg. Bye

elePHPant: I am hosting a meetup in 1.5 horus for devs to share their code with other devs Code Clash - meetup!

jacek: oO

VizGhar: 4.3M points in 2048... 58M more and I'll be first :D

Uljahn: guess 20M is doable in python

VizGhar: I'm trying C++ :) came with pretty fast eval function, but it's doing weird stuff after reaching 8192 tile :|

darkhorse64: Signed/unsigned or overflow in eval

darkhorse64: ?

VizGhar: most probably...

VizGhar: need to clear that spahgetti

mzbear: i claimed ~100k score with a hardcoded DLDLDLDR-loop without even bothering to inspect the board state at any point ;D

mzbear: couldn't quite decide on internal board representation format so i never really got started with it

mzbear: the dreaded analysis paralysis ... sometimes it would be better to just start writing code instead of thinking too much

VizGhar: 64 bits is nearly enough

Westicles: Fun problem. Only 14 of the cases can get the 131k tile (the rng needs to spit a 2 at the right time or it is impossible)

Westicles: or maybe a 4, I forget which

cazka: why is String.replaceAll() not a function in this IDE for JavaScript?

struct: use replace

struct: with global flag

struct: replace(/a/g, 'b');

cazka: thanks

ZzzTRAPzzZ: Tychkorg and Matheiu OP

Default avatar.png NguyenDinhDuong: :stuck_out_tongue_closed_eyes: