Blockchain Mining Games

Slides:



Advertisements
Similar presentations
Introduction to Game Theory
Advertisements

CPS Bayesian games and their use in auctions Vincent Conitzer
Price Of Anarchy: Routing
This Segment: Computational game theory Lecture 1: Game representations, solution concepts and complexity Tuomas Sandholm Computer Science Department Carnegie.
MIT and James Orlin © Game Theory 2-person 0-sum (or constant sum) game theory 2-person game theory (e.g., prisoner’s dilemma)
Study Group Randomized Algorithms 21 st June 03. Topics Covered Game Tree Evaluation –its expected run time is better than the worst- case complexity.
Online Ramsey Games in Random Graphs Reto Spöhel Joint work with Martin Marciniszyn and Angelika Steger.
An Introduction to Game Theory Part II: Mixed and Correlated Strategies Bernhard Nebel.
Computational Game Theory
CPSC 668Set 10: Consensus with Byzantine Failures1 CPSC 668 Distributed Algorithms and Systems Fall 2006 Prof. Jennifer Welch.
Selfish Caching in Distributed Systems: A Game-Theoretic Analysis By Byung-Gon Chun et al. UC Berkeley PODC’04.
1 search CS 331/531 Dr M M Awais A* Examples:. 2 search CS 331/531 Dr M M Awais 8-Puzzle f(N) = g(N) + h(N)
Query Incentive Networks Jon Kleinberg and Prabhakar Raghavan - Presented by: Nishith Pathak.
Near-Optimal Network Design with Selfish Agents By Elliot Anshelevich, Anirban Dasgupta, Eva Tardos, Tom Wexler STOC’03 Presented by Mustafa Suleyman CIFTCI.
On Bounded Rationality and Computational Complexity Christos Papadimitriou and Mihallis Yannakakis.
DANSS Colloquium By Prof. Danny Dolev Presented by Rica Gonen
Network Formation Games. Netwok Formation Games NFGs model distinct ways in which selfish agents might create and evaluate networks We’ll see two models:
Network Formation Games. Netwok Formation Games NFGs model distinct ways in which selfish agents might create and evaluate networks We’ll see two models:
Chapter 9 Games with Imperfect Information Bayesian Games.
Auction Seminar Optimal Mechanism Presentation by: Alon Resler Supervised by: Amos Fiat.
Standard and Extended Form Games A Lesson in Multiagent System Based on Jose Vidal’s book Fundamentals of Multiagent Systems Henry Hexmoor, SIUC.
Games with Imperfect Information Bayesian Games. Complete versus Incomplete Information So far we have assumed that players hold the correct belief about.
On a Network Creation Game PoA Seminar Presenting: Oren Gilon Based on an article by Fabrikant et al 1.
Beyond selfish routing: Network Games. Network Games NGs model the various ways in which selfish agents strategically interact in using a network They.
Beyond selfish routing: Network Games. Network Games NGs model the various ways in which selfish users (i.e., players) strategically interact in using.
1. 2 You should know by now… u The security level of a strategy for a player is the minimum payoff regardless of what strategy his opponent uses. u A.
Game tree search Chapter 6 (6.1 to 6.3 and 6.6) cover games. 6.6 covers state of the art game players in particular. 6.5 covers games that involve uncertainty.
Vasilis Syrgkanis Cornell University
5.1.Static Games of Incomplete Information
Subverting Bitcoin David Evans and Samee Zahur. Mining Why do we need miners?
Network Formation Games. NFGs model distinct ways in which selfish agents might create and evaluate networks We’ll see two models: Global Connection Game.
2.5 The Fundamental Theorem of Game Theory For any 2-person zero-sum game there exists a pair (x*,y*) in S  T such that min {x*V. j : j=1,...,n} =
Network Formation Games. NFGs model distinct ways in which selfish agents might create and evaluate networks We’ll see two models: Global Connection Game.
CSCE 668 DISTRIBUTED ALGORITHMS AND SYSTEMS
Satisfaction Games in Graphical Multi-resource Allocation
Contents of the Talk Preliminary Materials Motivation and Contribution
On a Network Creation Game
Lecture 13.
The Duality Theorem Primal P: Maximize
Tools for Decision Analysis: Analysis of Risky Decisions
Great Theoretical Ideas in Computer Science
Minimum Spanning Tree 8/7/2018 4:26 AM
Introduction to Game Theory
Extensive-form games and how to solve them
Introduction to Operations Research
Artificial Intelligence Problem solving by searching CSC 361
Game Theory Developed to explain the optimal strategy in two-person interactions. Initially, von Neumann and Morganstern Zero-sum games John Nash Nonzero-sum.
Asymptotic Notations Algorithms Lecture 9.
The Importance of Communities for Learning to Influence
The Subset Sum Game Revisited
Game Theory in Wireless and Communication Networks: Theory, Models, and Applications Lecture 2 Bayesian Games Zhu Han, Dusit Niyato, Walid Saad, Tamer.
Multiagent Systems Game Theory © Manfred Huber 2018.
Alternating tree Automata and Parity games
Learning 6.2 Game Theory.
Kevin Mason Michael Suggs
Introduction Wireless Ad-Hoc Network
NIM - a two person game n objects are in one pile
Network Formation Games
Vincent Conitzer CPS 173 Mechanism design Vincent Conitzer
Pruned Search Strategies
Games with Imperfect Information Bayesian Games
CSCE 668 DISTRIBUTED ALGORITHMS AND SYSTEMS
Example: multi-party coin toss
Clustering.
Network Formation Games
Switching Lemmas and Proof Complexity
Class 11 Max Flows Obtain a network, and use the same network to illustrate the shortest path problem for communication networks, the max flow.
Majority is not Enough: Bitcoin Mining is Vulnerable
CPS Bayesian games and their use in auctions
Presentation transcript:

Blockchain Mining Games Kiayias, Aggelos, et al. "Blockchain mining games." Proceedings of the 2016 ACM Conference on Economics and Computation. ACM, 2016. Presented by Zhenpeng Shi 7/19/2019

Introduction Selfish Mining Miners withhold releasing mined blocks in order to get larger revenue than their fair share. Simulation results shows that to avoid selfish mining, at least 67% miners must be “honest” miners. Can we prove it theoretically? How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you. Fig. 1. Mining power α above which selfish mining trumps honest mining, function of the propagation factor γ.[1] [1] Eyal, Ittay, and Emin Gün Sirer. "Majority is not enough: Bitcoin mining is vulnerable." Communications of the ACM61.7 (2018): 95-102.

Introduction Blockchain mining games Game theory: the study of mathematical models of strategic interaction between rational decision-makers. Nash equilibrium: a state where no player has anything to gain by changing only its own strategy. What we want to see: at Nash equilibrium, every miner gets its fair share of reward, which is proportional to its relative computational power. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Introduction Blockchain mining games Each miner is a player in the game, its object is to maximize reward. Miners are selfish, which yields a non-cooperative game. Miners’ strategies are: (1) which block to mine (what to compute next) (2) when to release mined blocks (when to release the results of computation) How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Model Two complete-information game models (1) immediate-release game Every miner releases immediately the mined blocks Can show that the threshold is 0.361≤ ℎ 0 ≤0.455 How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Model Two complete-information game models (2) strategic-release game Miners can withhold releasing blocks. To ensure complete information, add the restriction: miners must immediately announce a successful mining of a block, but not the block it self. Can show the threshold is bounded by ℎ 0 ≥0.308 How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Model Parameters 𝑛 – the number of miners (players) 𝑝=( 𝑝 1 ,…, 𝑝 𝑛 ) – the probability that miners succeed in solving the crypto- puzzle, which reflects their relative computational power; sum up to 1: 𝑖=1 𝑛 𝑝 𝑖 =1 𝑑 – the depth of the game, the payment for mining a new block is paid after 𝑑 new blocks attached to it 𝑟 ∗ – reward for mining a block; assume 𝑟 ∗ is constant and normalized to 1 How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Model Definitions Public state: a rooted tree. Every node is labeled by one of the players. The nodes represent mined blocks and the label indicates the player who mined the block. Every level of the tree has at most one node labeled 𝑖. Private state: similar to the public state except it may contain more nodes called private nodes and labeled by 𝑖. The public tree is a subtree of the private tree and has the same root. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Model Definitions Strategy: A pure strategy of player 𝑖 consists of two parts: Mining: to select a node of the current public state to mine Releasing: a (perhaps empty) private part of the player’s state which is added to the public state Frontier strategy: to release any mined block immediately and select to mine one of the deepest nodes. In a fair blockchain mining game, frontier strategy is expected to be played by all players. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Model Definitions Payments: For some nodes of the tree, the miners who discovered them will get a fixed payment (normalized to 1). The payments comply with the following rules: (1) the nodes that receive payment must form a path from the root, which implies that at every level of the tree exactly one node receives payment. (2) among the nodes of a single level that satisfy the above path restriction, the first one which succeeds in having a descendant 𝑑 generations later receives payment. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Assume that all miners who follow the Frontier strategy act as a single miner. 2-player game: Miner 1 is the miner whose optimal strategy we wish to determine and has relative computational power 𝑝; Miner 2 is assumed to follow the Frontier strategy and have collective relative computational power 1−𝑝. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Prune away abandoned branches. The state will be a long path followed by two branches, one for each miner of lengths 𝑎 and 𝑏. Denote the state by (𝑎,𝑏), notice that 0≤𝑎≤𝑏+1. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Mining states M: both miner keep mining their own branch. Capitulation states C: Miner 1 gives up its branch and continue from a block of the other branch. Winning states W: Miner 2 gives up its branch and continue from the newest block of other branch. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you. M-orange, C-green, W-red

The Immediate-Release Game Let 𝑔 𝑘 𝑎,𝑏 denote the expected gain of Miner 1 when the frontier advances by 𝑘 new levels starting from an initial tree at state (𝑎,𝑏) Constant 𝑔 ∗ is the expected gain per level in the long run. Define: Potential function 𝜑(𝑎,𝑏) denotes the advantage of Miner 1 for currently being in state (𝑎,𝑏). How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Consider three cases: (1) 𝑎,𝑏 ∈𝑀. Both miners continue mining, with Prob. 𝑝, the new state is (𝑎+ 1,𝑏), with Prob. 1−𝑝, the new state is (𝑎,𝑏+1) (2) 𝑎,𝑏 ∈𝐶. Miner 1 abandons its branch and the new state is (0,𝑠), where 0≤𝑠<𝑏 (3) 𝑎,𝑏 ∈𝑊. Miner 2 abandons its branch and the new state is (0,0), Miner 1 gains 𝑎 By definition, 𝑔 0 𝑎,𝑏 =0. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Get a similar recurrence for 𝜑 Fix 𝜑 0,0 =0. Notice that potential function 𝜑(𝑎,𝑏) is always non-negative. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Frontier is a Nash equilibrium if, having fixed the strategy of all miners except 𝑖 to Frontier, the best response of a miner 𝑖 has expected gain per level equal to 𝑝 𝑖 (the probability that miner 𝑖 succeeds in mining a new block, or its relative computational power), i.e., 𝑔 ∗ = 𝑝 𝑖 . Theorem: In the immediate-release model, Frontier is a Nash equilibrium when every miner 𝑖 has relative computational power 𝑝 𝑖 ≤0.361. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Important lemmas in proof: (1) Miner 1’s winning probability starting at state (𝑎,𝑏) is no more than 𝑝 1−𝑝 1+b−a . (2) If Miner 1 capitulates from state (0,1), no other state except (0,0) will ever be reached. (3) If 𝑝<0.361, then state 0,2 is not a mining state. (4) For 𝑝<(3− 5 )/2, if (0,2) is not a mining state, then (0,1) is not a mining state. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Immediate-Release Game Upper bound of 𝑝 Theorem: When Miner 2 plays Frontier, the best response strategy for Miner 1 is not Frontier when 𝑝≥0.455. Idea of the proof: Fix the depth 𝑑 (e.g. let 𝑑=3), then construct 𝜑(𝑎,𝑏) so that it’s a valid potential function of the game, from which we get a relationship between 𝑔 ∗ and 𝑝. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Strategic-Release Game Similar to the immediate-release game. 𝑎 𝑟 ≤𝑎 denotes the number of released blocks on Miner 1’s branch. All blocks on Miner 2’s branch are always released. It can be 𝑎>𝑏+1. Assume that (1) if 𝑎≤𝑏, then 𝑎 𝑟 =𝑎; (2) if 𝑎>𝑏, then 𝑎 𝑟 =𝑏, that is, 𝑎 𝑟 =min⁡{𝑎,𝑏} How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Strategic-Release Game Define the expected gain of Miner 1 as Consider the following two cases: (1) 𝑎≤𝑏. Miner 1 can choose to capitulate or to mine, which is the same as in the immediate-release game. (2) 𝑎>𝑏. Miner 1 has another option to release an additional block and lead the game to state (𝑎−𝑏−1,0), in which case Miner 2 will capitulate. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

The Strategic-Release Game How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you. Theorem: In the strategic-release model, Frontier is a Nash equilibrium when every miner 𝑖 has relative computational power 𝑝𝑖≤ 0.308

The Strategic-Release Game Idea of the proof: For states 𝑏+1,𝑏 , Miner 1 probably will release all blocks. For states (𝑎,𝑏) with 𝑎>𝑏+1, Miner 1 is safe not to release blocks until Miner 2 catches up within distance one. Define a potential function 𝜑 (𝑎,𝑏) which tends to overestimate the optimal potential. Prove that when 𝑔 ∗ =𝑝 and 𝑝≤0.308, 𝜑 (𝑎,𝑏) satisfies the recurrence of 𝜑 (𝑎,𝑏) The optimal gain per step cannot exceed 𝑝, that is How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Conclusion Provide a game-theoretic model to analyze bitcoin miners’ behavior. Identify the thresholds of the relative computational power below which Frontier is a Nash equilibrium. For the immediate-release game, the threshold is 0.361≤ ℎ 0 ≤0.455 For the strategic-release game, the threshold is bounded by ℎ 0 ≥0.308 Future work: Tighten the thresholds. Consider the incomplete-information game. How presentation will benefit audience: Adult learners are more interested in a subject if they know how or why it is important to them. Presenter’s level of expertise in the subject: Briefly state your credentials in this area, or explain why participants should listen to you.

Thank You!