Game Theory in Blockchain?

The blockchain consensus seems to be very related to the Game Theory, when each participant tries to achieve their "favorites", where defining "what favorites means" is not very easy, since a participant may ex. intends to gain more coins, while another one tries to destroy the system and decrease its reputation as a "competitor crypto-currency" etc.

However, in the Game theory Wikipedia page, there is a list of "Game types" such as :

Cooperative / Non-cooperative,
Symmetric / Asymmetric,
Zero-sum / Non-zero-sum,
Simultaneous / Sequential,
Perfect information and imperfect information,
Combinatorial games,
Infinitely long games,
Discrete and continuous games,
Differential games,
Evolutionary game theory, Stochastic outcomes, Metagames, Pooling games, Mean field game theory ...

Regarding to above explanations, what types of these Game Theory are more involved in the blockchain consensus?

Questioner

Posted 2018-10-20T13:52:15.727

Reputation: 906

Answers

I believe your question is flawed. Consensus isn't about game theory, it's an agreement. While game theory is analysis of strategies dealing with competition. Yes game theory has been applied outside of gaming. It's been used in war, business, even biology. Also it's not far from decision theory. While each individual in a consensus may have opinions facts to validate their positions they're still being judged by people. Game theory attempts to reduce decision making to mathematics, but ultimately we're talking about people here not calculators. Sometimes your best answer is just to eat the grapes & not pick side A or B. Meaning sometimes during a competition there is a winner. In consensus it's not always about a winner as much as it's about reaching an effective agreement which may be a 3rd option not yet revealed. Ultimately people can be irrational & do unexpected things.

OrcaMan42

Posted 2018-10-20T13:52:15.727

Reputation: 31

Thank you, but I think in ex. PoW when miners have to choose a strategy by which either gain more rewards (coins) or maybe destroying that system as a competitor cryptocurency, so as a result, their strategies affect other miners decision, maybe a good example would be Selfish mining or Block-withholding in Bitcoin : (https://arxiv.org/pdf/1311.0243.pdf%7C) Do you think the selected strategies by miners about working on which branch of blockchain is not a game theory? when their decisions affect other miners strategy. Thanks

– Questioner – 2018-11-19T12:30:47.363

Again, you're not thinking clearly. Miners offer a service. Their job is to "mine". That said the more blockchains & off shoots of blockchains there are for them to mine, the more money is in it for them. The "large miners" mine all the coins they can. So it's not one coin vs another as you're thinking. There's no bidding war that says I can mine this coin & you can't. It's simply a matter of what coin is most productive for me as a miner to mine. It's possible that if one coin becomes too expensive to mine a miner could switch to mining another coin & that happens all the time. In addition some are "under contract" to mine a specific amount. Some will just mine empty blocks or blocks that need confirmations. Again this all depends on various factors that go into mining. What they mine is usually up to them. No one says they have to be under contract. It would make no sense for a miner to destroy something that they could potentially go back to & harvest more money at a later time should situations change for one reason or another. Their decision is either to be in the game or not. If they're a miner their decision is made. If they change their minds ok fine, but 1 miner can't destroy something without have a 51% majority. Which is why its important for everyone to mine. So no one person / company can do that.

OrcaMan42

Posted 2018-10-20T13:52:15.727

Reputation: 31

Thank you, but when I say "destroying crypto-currency", I mean "decreasing value of the system (such as Bitcoin) and its related crypto- currency by attacking the system successfully." -> – Questioner – 2018-11-19T18:07:53.930

-> I can propose you to read this paper : (https://arxiv.org/pdf/1311.0243.pdf%7C) and this one : (https://arxiv.org/pdf/1411.7099.pdf) where you can see that gaining more rewards (more bitcoins) is not always the target of a miner if they can achieve more rewards by block-withholding instead of broadcasting new block immediately (i.e. a selfish miner) or even by losing usual rewards, they may decide to attack a system successfully to devalue target cryptocurrency. ->

– Questioner – 2018-11-19T18:08:48.103

-> So, we can see that there are several strategies with different probability of success depending on the miner's power and many other parameters. In such a situation, miner is playing a game in which they need to choose the best strategy to achieve their goal that for each miner may be different. – Questioner – 2018-11-19T18:08:59.187

These papers are from 2014. Policies have changed since then. I suggest you read the NY agreement which happened because of these issues this paper put forth & read the following news articles that came out of the bitcoin community directly since then. Because of disagreements in policy & procedure bitcoin has been forked so many times it might as well be a roasted hog with an apple in its mouth. We have bitcoin core, bitcoin cash, bitcoin plus & "Bitcoin". Yes what you suggested did happen prior to the NY agreement. https://en.bitcoin.it/wiki/New_York_Agreement Which brought forth: https://medium.com/@DCGco/bitcoin-scaling-agreement-at-consensus-2017-133521fe9a77 & https://github.com/btc1/specifications/blob/master/PCS/PCS-2017-0002-segwit2x.md

OrcaMan42

Posted 2018-10-20T13:52:15.727

Reputation: 31

I am not talking about only this particular attack, but it was only an example to show that consensus can be a game theory. Thanks – Questioner – 2018-11-19T22:47:29.977

consensus | kənˈsensəs | noun [usually in singular] a general agreement: [as modifier] : a consensus view | a consensus of opinion among judges. – OrcaMan42 – 2018-11-20T00:27:32.120

game theory | ˈɡeɪm ˌθɪəri | (also games theory) noun the branch of mathematics concerned with the analysis of strategies for dealing with competitive situations where the outcome of a participant's choice of action depends critically on the actions of other participants. Game theory has been applied to contexts in war, business, and biology. Compare with decision theory. – OrcaMan42 – 2018-11-20T00:27:45.087

decision theory | dəˈsɪʒən | noun the mathematical study of strategies for optimal decision-making between options involving different risks or expectations of gain or loss depending on the outcome. Compare with game theory. – OrcaMan42 – 2018-11-20T00:28:00.107

Where are the differences? NOT the similarities, but the differences. Game theory is math, consensus is an agreement. You can use one with the other, they go well together, like a coffee & a bagel. But if I see you putting cream cheese in your coffee & creamer on your bagel I'll shoot you myself for being an alien. The two aren't interchangeable. – OrcaMan42 – 2018-11-20T00:33:35.333

I do not say that game theory = consensus ! , sure they are two different concepts obviously. What I asked is : what type of "game theory" may be used in "consensus" ? meaning that e.g. a miner can use game theory strategies in a consensus mechanism such as PoW to enhance probability of their winning (For example). – Questioner – 2018-11-20T09:44:54.793

-> Meaning that investigating strategies of the miners in consensus process from point of view of game theory since in a consensus process, miners (or in general validators) need to decide about their strategies (and maybe decision theory) – Questioner – 2018-11-20T09:51:45.723