485
Coding chess
(startrek.website)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 29 Jan 2024
485 points (98.4% liked)
AnarchyChess
5206 readers
315 users here now
Holy hell
Other chess communities:
!Chess@lemmy.ml
!chessbeginners@sh.itjust.works
founded 1 year ago
MODERATORS
I really hope this is satire. Otherwise I'm scared to ask how long it took.
It's definitely satire. 2 million lines of code is an absurd under-exageration. This post had me looking up the number of possible chess games, because if you coded chess like above you would have to have an if statement for every outcome, and it's 10^120 different possible games.
https://en.wikipedia.org/wiki/Shannon_number
The way I understood it, it's two million lines and nowhere near finished.
Anyway, satire.
That's the number of possible games, the number of possible board states is much lower, 10^40.
Although you're still clearly correct in the end anyways because it's still an absurd number of board states and it's not even formatted to be one state per line.
I read that there are less atoms in the universe than possible chess games, which is quite insane
By an extremely significant margin. Here's another fun one: getting a unique shuffle in a deck of cards is 1/52!. So if you wanted to count all of the different possible arrangements of cards, counting one per second, you can:
Start walking around the equator at a leisurely pace of one step per billion years.
Once you've made it around the earth, remove a single drop of water from the Pacific Ocean and walk around the earth again.
Once the Pacific Ocean is empty, re-fill it and lay a sheet of paper on the ground. Keep stacking a new sheet every time you've re-emptied the ocean drop by drop every time you circle the earth at one step every billion years.
When the stack of paper reaches the sun, you're about a third of the way there!
The way I like to put it is that every single time you randomly shuffle a deck of cards, you are guaranteed to get an order that has never been seen before, by anyone in history. That will be the case for every person who ever shuffles a deck of cards for the rest of time.
To be fair there was that one time a perfect shuffle led to a perfect bridge deal, each player getting a full suit. Sometimes even a fair shuffle goes weird lol.
You only have to code a fraction of those as the computer should take the same move for several of the user inputs.
The post is satire, but I remember being ~8-9 and trying to create a "game" in Microsoft Word with hyperlinks between documents and nothing else. I had hundreds of documents (each representing a game state) before I got tired of that project.
That's pretty clever for that age and shows, that your brain already had the right wiring for (software) engineering. You set out to solve a problem by leveraging the tools you had and bending their functions to your needs. There's a lot of abstract thinking involved.
Today I will remind everyone of DRAGON: A Game About a Dragon
This game was developed by someone who didn't know anything about programming outside of IF statements, integers, and strings. Here is an excerpt of the massively long source code
For a second there I thought the 100% science-based dragons game had been made.
Holy shit... That's nuts. I do see the occasional boolean in there. But it might as well be ifs ints and strings like you said.
Man went full static C programming in C# and coded the whole thing in what looks like one class. Instances?, loops excepting XNA gameloop? What's that? Whos that? Where's that?
Honestly I'm not even mad, that's dedication. Dedication and a refusal to learn more than just the basics. Kinda want to buy the game now just to see how far a fucktonne of if statements can get. By the looks of it, I'd say straight into the dragons maw.
Yandev as well. Guy put the ai for every single character in a single function separated with if/then statements. It was also called every game tick.
My eyes! The ~~global~~ variable list is huge!
Edit: nm, I looked again and they're in a class. Still insane either way.
Literally impossible to code every bosrd state, so forever
The person who downvoted you must not have thought it through.
For comparison to that 10^120^ possible game states, the number of atoms in the universe is only 10^82^ and the amount of time until the "Dark Era" of the universe (after all the stars have died and even all the black holes have decayed via Hawking radiation) is only 10^114^ to 10^117^ seconds. In other words, it really is literally impossible either to build a computer big enough to store all the board states or to write them all down even if you did.
I mean a good program could do a 100 million states/s . So if you could make a computer that lasts until the end of the universe it could go through all of them I guess?
With supercomputers, probably way more, that's just the figure for like a laptop. Plus you could filter out all the states that are symmetric and stuff and lower it by one order of magnitude.
I guess, maybe?
I'm not prepared to speculate on the performance of a computer literally larger than the universe.
that's the most fun type of speculation. No fun speculating over computers that could emulate realistic things like all states of a rubiks cube, cause that's probably already been done.
...hey, wait a second, you pulled a fast one on me with that "100 million states/s" silliness!
I just remembered what the time boundary I was talking about was actually trying to measure: it had nothing to do with the speed of computation; it was the speed of typing in the program code. That's why 1 state per second was a reasonable estimate (if not overly optimistic). If you tried to type in all those
ifs andprints manually, that's what would take you longer than the heat death of the universe.Besides, even executing the program can't do 100 million states per second because it only does one state transition and then waits for user input.
Oh I was assuming you'd write another program to create this program, like the "4 billion if's" blog post if you've read it.
There's a mathematician that figured that there's 10^120 possible chess games, as a lower bound. https://en.m.wikipedia.org/wiki/Shannon_number#:~:text=Shannon%20showed%20a%20calculation%20for,a%20Computer%20for%20Playing%20Chess%22. That's a 1 followed by 120 zeros for just the number of possible games. With this method they'd have to manually go through every move for every one of those games. If we say a game lasts 30 turns on average and they'd take 1 second to code each turn (realistically it'd be longer) it'd take 6.9*10^109 (69 followed by 108 zeros) times as long as the age of the universe.
So it's doable? That's all I needed to hear.
Are you a theoretical computer scientist?
Nice
for just the MINIMUM number of possible games. (lower bound)
I did say lower bound, yes
I responded to the wrong comment. Sorry.
this is definitely satire, otherwise it would take longer than the age of the universe to finish coding it lol
I've seen people try to write programs like this.