Your Premier Source To Capitalize On The Bitcoin Revolution! 


Join Our Home Based Business Opportunity Today.



COINOMETRICS BRIEFING #1 - THE 50% CLUB

Posted by Unknown in on -


Authors
Jonathan Levin –  jonathan.levin@coinometrics.com
Andrew Geyl – ooc@coinometrics.com


Executive outline

---The 50% issue could be a revenant issue within the Bitcoin network because of the underlying economic incentives facing individual miners


---Miners with solely slightly but five hundredth will accomplish terribly similar varieties of attack as a five hundredth pool


---The random nature of the block determination method implies that a pool might simply have a hashrate on top of five hundredth for a brief amount even supposing longer horizon estimates counsel that they're smaller


Introduction
The existence of one entity dominant five hundredth of the mining capability on the Bitcoin network may be a failure of Bitcoin, not as a result of the entity might act maliciously with no recourse or it's a negative impact on the value however as a result of Bitcoin will now not be referred to as a suburbanised system
1
. the aim of Bitcoin as a suburbanised system was to distribute trust between many various entities that might be unable to interact against the interests of the users. The existence of a mining pool that's even near five hundredth breaks this and therefore the foundations of Bitcoin. This making known may be a study of the past performance of the Bitcoin network once one entity approaches the five hundred mark and explains however, at Coinometrics, we tend to still monitor this example terribly fastidiously. the five hundred club has been threatening to accumulate a lot of members for reasons that require to be resolved by the community. The return of the problem is caused by the present economic incentives that face individual miners. Individual miners square measure absolve to choose a mining pool of their selection and also the price of switch between them is negligible. but there exist 2 incentives to hitch the most important pool.
1.

The reason for mining pool’s existence is reduction in variance of returns. Mining returns square measure random and so the temporal order of the reward for solo mining is extraordinarily unsure. Miners need to extract the utmost price from their current hashing power at the present problem as every rise in Bitcoin problem reduces the effectiveness of their hardware. the very best variance reduction is found within the largest pools. 2.

For a bigger pool, miners pay less time on the average performing on redundant mining issues. once a jack solves a block they straightaway begin performing on ensuing block whereas all different miners got to wait to verify it for themselves. Larger pools solve a lot of blocks making atiny low advantage. data takes time to travel across the Bitcoin network and as a result some miners pay longer than others performing on previous blocks which will ne'er create it on to the blockchain. This will increase the expected come per hash tired a bigger mining pool. whereas these economic incentives square measure terribly troublesome to avoid we are able to produce AN incentive compatible

1

Malicious activity includes reversing users transactions, ignoring other miner’s blocks, blocking certain transactions and selfish mining. See here for a very good summary (Hacking Distributed article).



The 50% Club

system by making it prohibitively expensive to run pools over certain sizes. This is currently being addressed in many proposals to change the proof of work.


Bitcoin relies on probabilistic outcomes. the problem of making new blocks relies on finding the answer to a scientific discipline puzzle. the safety of the network so depends on the result of random method and isn't settled. A mineworker with forty eighth will reverse ten confirmations with associate eighty fifth success rate.
Even at the chosen four-hundredth voluntary cap, the mining pool may overturn six confirmations with five hundredth likelihood. Recognising that this is often associate unregulated market and also the miners can decide the ultimate outcome, the community should vocalise what levels of market concentration area unit acceptable given the probabilistic nature of security. area unit we tend to comfy with one entity having the ability to overturn six confirmations with five hundredth success rate?


Estimation
Estimates of the network hashrate and market share of the mining pools area unit virtually by definition inaccurate. The hashrate of pools that don't unleash public figures should be inferred from the amount of blocks resolved by the mining pool over a protracted amount.

 For the mining market to serve its purpose in delivering the service of security and spread to the Bitcoin network through intense competition, pools ought to be inspired to publish their hashrate in public. At Coinometrics we tend to use a mixture of various observation periods associated pool reportable knowledge to come back up with an correct live of the network hashrate and also the market concentration of Bitcoin mining.


Analysis
There has ne'er been a pool or identifiable  solo mineworker that has resolved over five hundredth of the network blocks for any week (Monday to Sunday, UTC). within the past there has been many instances of a pool determination over 1/2 the blocks in a very day. but Ghash.io has ne'er resolved over 1/2 the blocks in a very single day. within the past measures were taken to confirm that one mining pool didn't dominate the market. archangel Marsee of BTC gild enforced changes to their protocol that caused the sharp decline of their network hashrate in could 2013 (See Figure 1). Voluntary acts by pools will so simply scale back the danger of a pool connection the five hundred club. Figure one shows the 3 pools that have come back nighest to passing the five hundred mark. the road graph shows the weekly smoothened estimate of the amount of blocks resolved and also the individual points show the observations. There area unit individual points within the proportion of daily blocks resolved round the five hundredth mark as shown within the figure. The pools is shown to simply surpass twenty fifth and thirty third of the network for important periods of your time, which can additionally threaten network security and competition.


 The Bitcoin network doesn't have an interior construct of your time. The timestamps place into blocks area unit generally wide inaccurate because the criterion for a legitimate timestamp is fairly loose. Bitcoin’s construct of your time is outlined by the discharge of latest blocks on the network. we are able to so modification our unit of interest from time to a rolling window of the amount of blocks. this permits United States of America to check periods that don't align with universal time


Source: Rosenfeld (2012) Analysis of hashrate-based double-spending https://bitcoil.co.il/Doublespend.pdf


Using a shorter amount yields additional innacurate measures requiring larger confidence intervals. there's a trade-off as employing a seven day average introduces a big lag to this estimate of the hashrate.



See here for a proof of however massive concentrations of hashing power will introduce incentives to refuse the Bitcoin protocol. Monopoly power additionally could raise the long-term equilibrium dealings fees.
Bitcoin relies on probabilistic outcomes. the problem of making new blocks relies on finding the answer to a scientific discipline puzzle. the safety of the network so depends on the result of random method and isn't settled. A mineworker with forty eighth will reverse ten confirmations with associate eighty fifth success rate.


 Even at the chosen four-hundredth voluntary cap, the mining pool may overturn six confirmations with five hundredth likelihood. Recognising that this is often associate unregulated market and also the miners can decide the ultimate outcome, the community should vocalise what levels of market concentration area unit acceptable given the probabilistic nature of security. area unit we tend to comfy with one entity having the ability to overturn six confirmations with five hundredth success rate?


Estimation
Estimates of the network hashrate and market share of the mining pools area unit virtually by definition inaccurate. The hashrate of pools that don't unleash public figures should be inferred from the amount of blocks resolved by the mining pool over a protracted amount.


 For the mining market to serve its purpose in delivering the service of security and spread to the Bitcoin network through intense competition, pools ought to be inspired to publish their hashrate in public. At Coinometrics we tend to use a mixture of various observation periods associated pool reportable knowledge to come back up with an correct live of the network hashrate and also the market concentration of Bitcoin mining.

The 50% Club
 days and weeks. We use a 1000 block rolling total (approximately 6 to 7 days) and a 150 block rolling total (approximately one day). There were no results close to 50% of the network for the 1000 block rolling total, but many for the 150 block rolling total.


Figure 1. Historical percentage of network blocks solved by the biggest block solvers





Source: Coinometrics, Organ of Corti


The results can clearly be seen in Figure 2. There are far more instances of pools exceeding 50% of the blocks solved in the 150 block window than the daily window applied initially. We find this method to be more accurate than relying on the UTC timestamps of the blocks. On one particular day Ghash.io solved 31 blocks according to the UTC timestamps but their 150 block median percentage of the network was above 50%. DeepBit also had many more exceedences using the rolling window method





Evaluation and concluding remarks

Even when the estimate of the intensity function (the smooth lines in Figure 2) is only a approaching 45% of the network, the stochastic nature of the block solving process means that a block solver could easily have a hashrate above 50% for a short while. This needs to be considered when deciding the tolerance level of market concentration. The three members of the 50% club have in the past had a sufficient proportion of the network to successfully mine selfishly, or perform double spends with a good probability of success for significant periods of time. As more complex and valuable services are built on the blockchain calculating the risk of these events occurring is still in its infancy. Action must be taken, BTC Guild, has been the only member of the 50% club to implement a protocol change that successfully lowered their share of the network. We applaud the recent actions of Ghash.io to remedy the situation but stronger actions must be taken while consensus builds around the different technical solutions. At Coinometrics we are committed to publishing accurate measurements and analysis on the structure of the mining industry.








F T G+




Leave a Reply