6 ms·

Motocoin whitepaper – a cryptocurrency based on proof-of-play in a 2D game [pdf]

27 points by networked 11y ago

jordanwallwork 11y ago
Quite frustrating that this currency seems to rely on a game yet the whitepaper never seems to specify what it is. Since the game is supposedly too difficult to write an AI for I'd be extremely curious to hear more about it. (also unfortunate timing mentioning games where humans are superior like 'go' given the current news!)
- yorwba 11y ago
  > Unlike other crypto-currencies, mining motocoins > is not a passive process. You will have to play > a 2D motorbike simulation game called Motogame, > which is inspired by ElastoMania. http://motocoin-dev.github.io/motocoin-site/#mine http://motocoin-dev.github.io/motocoin-site/#mine
  - zero_iq 11y ago
    Given how well games like Mario have been mastered by machine learning algorithms (to beyond expert human levels), I doubt a motocross-maniacs-style side scroller would present too much of a challenge. It would be interesting to see what, if anything, they've changed in the game to make it harder for a computer to play.
Animats 11y ago
"Proof of grinding"? This is really a CAPCHA scheme; a task that's easy for humans but hard for computers. But computers are too good at such tasks now. The result has been CAPCHA solving services, both automated [1] and manual.[2] Today's rate: $0.75 for 1000 CAPTCHAs. [1] http://www.captchatronix.com/ http://www.captchatronix.com/ [2] https://2captcha.com/ https://2captcha.com/
tinco 11y ago
edit: I'm sorry I'm a bit tired, this post is needlessly critical. The idea of the paper is fun, and can serve as a source of inspiration for good thought, I myself have enjoyed thinking about proof-of-play for a long time and I still hope I'll think of a way to make it work some day. (/edit) This system is in no significant way distinguishable from proof-of-work systems. Unfortunately the only 'difficulty' introduced here is that the computer has to interact with a physics engine. There's a lot of claims made in the paper but no references to any scientific research. Some examples: The article claims playing the game is hard for a computer, but never says how hard precisely, what's the complexity? The article claims games exist that are easier to play for humans than they are for computers such as Go, obviously for Go this has been shown to likely be false the past week, but also in general. Are there really games that are hard for computers but easy for humans? I played with the idea of a game-based cryptocurrency, as the basis for a MTG like economy in a roguelike game, but failed to come up with any mechanism that would be easier for humans to play than computers in a fundamental way, but I didn't get so serious as to look up scientific work, and neither has the author of this paper. The article claims that should a bot arise that can beat this game like humans can, they can simply increase the difficulty. Can they? he didn't calculate what effects the variables of the game have on the complexity. Who says that if the game becomes more difficult for computers, it is still playable (let alone fun) for humans?
- jchrisa 11y ago
  I'm slowly progressing on Document Coin which is predicated on building value from human relationships rather than CPU time. http://www.wired.com/2014/07/document-coin/ http://www.wired.com/2014/07/document-coin/ To do this we track public keys that have held each coin, and machine learn on the blockchain data. Latest update here https://qconlondon.com/presentation/hash-histories-toy-blockchains-great-danger https://qconlondon.com/presentation/hash-histories-toy-block...
  - im3w1l 11y ago
    That's a very interesting concept! My prediction is that it will turn into Celebrity Coin very quickly. Which may or may not be a bad thing. Giving incentives for celebrities to endorse his currency sounds like a powerful idea.
im3w1l 11y ago
> But in many games (for example Go) humans are still superior to computers. Was this submitted ironically?

27 points by networked 11y ago

Quite frustrating that this currency seems to rely on a game yet the whitepaper never seems to specify what it is. Since the game is supposedly too difficult to write an AI for I'd be extremely curious to hear more about it. (also unfortunate timing mentioning games where humans are superior like 'go' given the current news!)

yorwba 11y ago
> Unlike other crypto-currencies, mining motocoins > is not a passive process. You will have to play > a 2D motorbike simulation game called Motogame, > which is inspired by ElastoMania. http://motocoin-dev.github.io/motocoin-site/#mine http://motocoin-dev.github.io/motocoin-site/#mine
- zero_iq 11y ago
  Given how well games like Mario have been mastered by machine learning algorithms (to beyond expert human levels), I doubt a motocross-maniacs-style side scroller would present too much of a challenge. It would be interesting to see what, if anything, they've changed in the game to make it harder for a computer to play.

Animats 11y ago

"Proof of grinding"? This is really a CAPCHA scheme; a task that's easy for humans but hard for computers. But computers are too good at such tasks now. The result has been CAPCHA solving services, both automated [1] and manual.[2] Today's rate: $0.75 for 1000 CAPTCHAs. [1] http://www.captchatronix.com/ http://www.captchatronix.com/ [2] https://2captcha.com/ https://2captcha.com/

tinco 11y ago

edit: I'm sorry I'm a bit tired, this post is needlessly critical. The idea of the paper is fun, and can serve as a source of inspiration for good thought, I myself have enjoyed thinking about proof-of-play for a long time and I still hope I'll think of a way to make it work some day. (/edit) This system is in no significant way distinguishable from proof-of-work systems. Unfortunately the only 'difficulty' introduced here is that the computer has to interact with a physics engine. There's a lot of claims made in the paper but no references to any scientific research. Some examples: The article claims playing the game is hard for a computer, but never says how hard precisely, what's the complexity? The article claims games exist that are easier to play for humans than they are for computers such as Go, obviously for Go this has been shown to likely be false the past week, but also in general. Are there really games that are hard for computers but easy for humans? I played with the idea of a game-based cryptocurrency, as the basis for a MTG like economy in a roguelike game, but failed to come up with any mechanism that would be easier for humans to play than computers in a fundamental way, but I didn't get so serious as to look up scientific work, and neither has the author of this paper. The article claims that should a bot arise that can beat this game like humans can, they can simply increase the difficulty. Can they? he didn't calculate what effects the variables of the game have on the complexity. Who says that if the game becomes more difficult for computers, it is still playable (let alone fun) for humans?

jchrisa 11y ago
I'm slowly progressing on Document Coin which is predicated on building value from human relationships rather than CPU time. http://www.wired.com/2014/07/document-coin/ http://www.wired.com/2014/07/document-coin/ To do this we track public keys that have held each coin, and machine learn on the blockchain data. Latest update here https://qconlondon.com/presentation/hash-histories-toy-blockchains-great-danger https://qconlondon.com/presentation/hash-histories-toy-block...
- im3w1l 11y ago
  That's a very interesting concept! My prediction is that it will turn into Celebrity Coin very quickly. Which may or may not be a bad thing. Giving incentives for celebrities to endorse his currency sounds like a powerful idea.

im3w1l 11y ago

> But in many games (for example Go) humans are still superior to computers. Was this submitted ironically?