Introduction to Decred
Decred is a cryptocurrency designed for P2P payments in a similar vein as Bitcoin, albeit with a focus on governance through a hybrid PoW/PoS consensus architecture. The network was launched in February 2016 by the developers of btcsuite, a Bitcoin implementation written in Go. While not a fork of Bitcoin, Decred utilizes several of its design elements and explicitly aims to improve upon Bitcoin’s perceived shortcomings. Decred emphasizes governance input for all stakeholders through an on-chain voting and public proposal system; it was the first blockchain to implement an on-chain, token holder-approved hard fork. Initial project development was funded through a developer pre-mine of 4% of the total DCR supply; there was no VC financing or token sale of DCR. An additional 4% was airdropped to the community to spur adoption. Block rewards are split 60/30/10 between PoW miners, PoS stakers, and a development fund controlled by community vote.
Decred’s founders’ outlook on Bitcoin informed the project’s development. Decred’s original developers’ work on btcsuite led them to believe Bitcoin had several flaws, including:
- Ineffective and inefficient governance
- Lack of funding for protocol development
- Outsized influence of Proof of Work miners on development decisions
Jake Yocom-Piatt, who would later lead Decred project development through Company 0, discusses these issues at length in a late 2015 blog post. Indeed, the Bitcoin community has debated each of these areas considerably, with many viewing these factors as significant roadblocks to Bitcoin’s progress. Bitcoin has no formal governance structure, and decisions to alter the protocol are made entirely off-chain, typically by insiders/early adopters and heads of large mining operations. Changes to the Bitcoin protocol must be approved by consensus of the Bitcoin Core developers and adopted by miners, and there is no direct way for Bitcoin users or associated groups to vote on protocol changes. In particular, the Bitcoin Improvement Proposal (BIP) system provides a collaborative repository for proposing protocol upgrades, though no formal system for implementation.
One clear example of Bitcoin’s governance difficulties is the multi-year scaling debate saga, with various stakeholders advocating a variety of solutions to network transaction congestion on the Bitcoin network, including increasing the base block size, SegWit, and layer 2 solutions such as the Lightning Network. Potential solutions to the scaling issue have been negotiated off-chain by prominent stakeholders, leading to off-chain coordination mechanisms such as the New York Agreement. Indeed, these protocol upgrades must ultimately be adopted by miners, who are increasingly influenced by a limited number of ASIC manufacturers, such as Bitmain. However, this type of process is often complex and without enforcement measures, which in this case led to a break from the NYA and the eventual BTC/BCH fork in August 2017. Though many groups of financial stakeholders in the Bitcoin network exist, the core developers and large mining entities have a perhaps outsized influence on governance.
Further, the lack of clear development funding methods in Bitcoin is often seen as problematic. The core network software exists as open source code on Github, but it is difficult for developers to directly monetize their contributions to the codebase. Funding for Bitcoin Core developers was entirely donation driven until 2014. Instead, the ecosystem effectively supports development through external businesses such as Blockstream and non-profits such as the MIT Media Lab’s Digital Currency Initiative, both of which employ Bitcoin Core developers directly. Particularly in the case of Blockstream, a for-profit entity, this could create real or perceived conflicts of interest for developers. While developers may support themselves through early investment in BTC or ancillary work, the Bitcoin protocol itself does little to incentivize development work.
Decred’s design and architecture intends to address these issues, building a cryptocurrency with governance rights for an inherently broad array of stakeholders. The origin of the project can be traced back to a 2013 Bitcointalk thread by anonymous user ‘tacotime’, and the subsequent whitepaper titled ‘Memcoin2 (MC2): A Hybrid Proof of Work, Proof of Stake Crypto-currency’. The initial design of Decred was also inspired by the Proof of Activity whitepaper co-authored by Litecoin founder Charlie Lee. These development efforts merged in early 2014 with the open-source software engineering firm Company Zero and CEO Jake Yocom-Piatt, and the network launched in late 2015. Later, Decred released a constitution outlining the guiding principles and philosophy of the network.
Decred utilizes neither solely the ‘1 CPU = 1 vote’ of Bitcoin/PoW nor the ‘1 token = 1 vote’ of a pure PoS protocol, opting instead for a hybrid approach. Transactions on the Decred network are validated through a hybrid PoW and PoS system. At a high level, PoW miners generate blocks that a randomly selected set of PoS validators must validate before the block is appended to the main chain. This consensus architecture is part of Decred’s overall goal of giving various stakeholders input in managing the network. PoS validators act as an explicit check on PoW miners; the former can reject invalid blocks or those that use software versions not preferred by token holders. The block rewards are split 60/30/10 between miners, stakers, and the development pool. The 60% PoW miner reward is reduced proportionally if the block is not approved by all five randomly chosen PoS validators. Participation in the PoS process is voluntary through a ticket system, discussed in the following section. As of mid-2018, the annual PoS reward is approximately 15%.
Decred’s ticket system is closely related to the PoS component of the its consensus process. DCR holders who wish to participate in the PoS validation and governance process must stake DCR to receive ‘tickets’, which exist as a non-tradable cryptoasset on the Decred network. Five ticket holders are randomly selected by a Poisson distribution to validate each block that are created by PoW miners.
Effectively, only a subset of existing DCR is used in the governance process; governance influenced is measured not by DCR balance, but by ticket balance. By requiring a lock up period for the DCR to obtain tickets, Decred hopes that only users invested in the long-term growth of the network will be involved in the consensus process. Short-term speculators and day traders of DCR will not be able to participate in consensus or governance without making their holdings illiquid. Indeed, this design aims to solve a general issue in staking systems: the amount of voting weight to give exchanges or other large holders. Here, a centralized exchange that holds a large amount of DCR could only leverage that to influence voting/governance if it was willing to lock up their DCR for several weeks. For an exchange that allows users to immediately withdraw deposited DCR, this is unworkable. Further, the ability for large token holders to influence the network by acquiring many tickets is limited, since the ticket price dynamically adjusts according to demand in previous periods. A significant DCR holder would face a rapidly increasing marginal cost to acquire a large supply of tickets over a short time frame. Further, allocated tickets must ‘mature’ for 256 blocks before they are eligible for staking rights, which reduces the potential for a rapid attack on the network. Generally, Decred aims to minimize the influence of large PoW/PoS pools, exchanges, and individuals with large holdings in the governance and block validation processes.
The price of these tickets is dynamic according to market demand, and the DCR used to claim them is refunded once the ticket expires. There is a target ticket pool size of 40960 tickets available on the Decred network, designed such that a ticket will be chosen for block validation duties within 28 days on average. Tickets expire if not selected for block validation after 40960 blocks (about four months); any one ticket has a 99.5% probability of being chosen before it expires. Five tickets are randomly selected for an approval vote on each Decred block. At least three of the five selected tickets must approve the block generated by the PoW miners; blocks failing to gain three votes are orphaned. Once selected, ticket holders who vote on block validation split the 30% portion of the total block reward. Tickets can be delegated to staking pools to provide protection against missed votes; if a wallet with a ticket is not online when its ticket is randomly selected, they lose the staking rewards and governance input opportunity.
Decred uses a stake difficulty algorithm to determine a ticket price that maintains a constant ticket supply. The purpose of this is to maintain PoS subsidy returns over time and expected ticket holder influence over the several week timespan that tickets are typically held. Users purchase tickets by submitting the determined DCR price to the network. Up to 20 tickets are available for purchase in each block; in the event that demand exceeds this supply, allocation is determined by transaction fee. Tickets are eligible to be selected for block validation and voting 20 hours after they are purchased. Short term fluctuations in ticket price are limited by a stake difficulty algorithm upgrade implemented in mid-2017.
One concern in the Decred community is that the rising ticket price (about 100 DCR, as of mid-2018) excludes small holders from participating in governance and block validation. To remedy this, Decred will implement a ticket splitter, which effectively allows users to pool funds to purchase tickets. The voting decision of the split ticket is weighted by stake. The splitting process’ implementation reduces the variability in DCR’s locked duration; as a consequence of the change, users can split their claim over many partial tickets and receive unlocked DCR gradually as those tickets are randomly selected, instead of locking 100 DCR for up to 4 months.
The chart below illustrates the trends in staking and the ticket system in Decred since network inception, and compares the financial returns to staking with two partial substitutes. The percentage of the Decred supply staked as part of the ticket system has gradually increased from 25% in mid 2016 to nearly 50% in mid 2018, as shown in red. Given that these staked DCR are locked for an average of 28 days until the ticket is selected, this signals increased confidence in the security of the Decred network and utility of DCR as a store of value. In contrast, the net monthly returns in USD to staking has gradually decreased from over 5% at network launch to under 2% in mid 2018. These returns can be compared to baseline cryptoasset staking alternatives such as Dash, which currently has a monthly return of ~0.6%. However, the comparison with Dash is imperfect, since the staking requirement there is much higher (1000 DASH, ~$300k), there is no illiquidity through token lock-up, and the Dash network is significantly larger by market cap and transaction volume. Further, returns to DCR staking can be compared with baseline risk-free alternatives in the traditional financial markets, such as 1 month US T-bills currently hovering around 2% returns. Of course, staking DCR is not truly risk-free; allocated tickets cannot be redeemed at will to respond to market movements or code bugs. As such, the liquidity profile is somewhat similar to one month T-bills, and distinct from Dash. The data below is sourced from Dash and Decred network statistic explorers, and the US Treasury Department.
Individuals are incentivized to claim tickets for two primary reasons: to accrue a portion of the block reward allocated to the PoS component, and to participate in network governance through on-chain voting rights. Given this dual function, the trends illustrated above are notable, as they display increasing staking participation in tandem with falling block reward returns. There are multiple explanations for these trends. From a purely financial view, where users decide to stake based solely on estimated returns and risk/reward calculations, this is perhaps unsurprising: as trust in the network grows, more users are willing to bid for a finite number of tickets and the market return falls accordingly through this competitive process. The trend may also reflect user’s valuation of DCR’s extra-financial worth: DCR users are willing to stake for a lower effective return because the ticket system also grants them direct input in the Decred governance process. In that case, the value of holding a ticket and participating in staking is more than just the financial return of the block reward; direct governance rights are arguably a desirable feature in cryptoassets. While the chart above and trends specific to DCR certainly do not prove this relationship, as governance decisions in Decred have been uncontroversial thus far, it may be indicative of an emerging trend in valuing governance that will require further observation and consideration when examining and valuing other cryptoassets.
Politeia and Network Proposals
In addition to validating blocks of transactions, ticket holders can vote on-chain for proposed network changes. These changes are first considered through informal community discussions, then published as Decred Change Proposals (DCPs). All DCPs include clear documentation and motivation for the change, as well as code for a working and tested implementation. The DCP system is similar to Bitcoin’s BIP system, except that approved DCPs are automatically adopted by the network at the end of the voting period.
Decred’s voting process for DCPs follows a two-phase schedule to coordinate new code implementation and voting procedures. The Decred documentation includes a detailed guide to this process. First, the network validators must meet the upgrade threshold by utilizing the updated codebase version that uses the code proposed in the hard fork. Once 95% of the 1000 most recent blocks and 75% of the votes cast within the previous 2016 blocks have upgraded to the latest version, a voting period of the next 2016 blocks begins. In this voting period, Decred ticket holders who are randomly selected to validate a block also vote or abstain from voting on the proposal. For a vote to resolve, at least 10% of votes need be non-abstaining and at least 75% of those votes need be affirmative or negative. If these conditions are not met, there is a revoting period over the next 2016 blocks by a new set of chosen ticket holders. Since node operators have already updated with the additions, the protocol change directly causes the hard fork immediately following the voting period. In addition to the stake difficulty algorithm hard fork, the Decred network voting has also signaled support for developers to begin work on Lightning Network integration and to implement new opcodes for the Lightning Network. Given that code for proposed network changes must be developed prior to voting, it is perhaps unsurprising that the voting results for the three implemented hard forks suggested the changes were uncontentious; less than 2% of participating voters rejected the proposal in each case.
To manage information about governance proposals, Decred will utilize an in-development proposal system called Politeia. Politeia stores a timestamped record of data related to Decred’s governance off-chain, in order to reduce on-chain costs. Described as ‘git plus timestamping’ for governance proposals, it will also feature a user interface for voting on development fund allocations. Users can submit projects to the Decred community for funding from the 10% development fund block subsidy. This is similar to the Dash treasury system. The development fund will function as a smart contract-controlled DAO with on-chain voting for allocations. Prior to this network upgrade, the development funds accrue to the fund wallet.
Company 0 and CEO Jake Yocom-Piatt oversaw initial project development, with the network launching in February 2016. Early developers split a premine of 8% of the total DCR supply (1.68 million DCR) as compensation for their work, and individuals contributing to technological advancement and with demonstrated interest in Decred received a curated airdrop. The developers were compensated for their initial work at a rate of $0.49 per DCR.
The target Decred block time is five minutes, and each block has a maximum size of 393kb. Through mid 2018, network congestion has not been an issue in Decred as it has in Bitcoin. The supply schedule of Decred is similar to that of Bitcoin; the long term supply is capped at 21 million DCR, with a decreasing block reward schedule that divides rewards between PoW miners, PoS stakers, and the development fund. The last block reward will be created in 2039; afterward, the network will be secured by transaction fees alone. Blocks that receive fewer than the maximum of 5 PoS votes receive a proportionally slashed block reward; as such, a portion of the 21 million maximum DCR supply is effectively removed from the supply and not allocated to any user. The total block rewards for Decred decrease by 1% every 21 days; this is in contrast to Bitcoin, where block rewards decrease by half every four years. This is intended to minimize sudden changes to mining profitability and ensure a more stable ecosystem. Decred transactions use the UTXO model similar to Bitcoin, as opposed to the account model of networks such as Ethereum. Decred is developing integration with the Lightning Network, and was involved in the first atomic swap with Litecoin in September 2017.
The PoW component of Decred uses the Blake-256 hashing algorithm. In early 2018, ASIC mining units for Blake-256 hashing algorithms were released, quickly increasing the total hashing power of the Decred network by a factor of 20. These ASICs can also be used in Sia. Similar to Bitcoin, hashing power in Decred is relatively concentrated to a small number of mining pools. Due to Decred’s hybrid consensus mechanism, where PoS stakers can reject fraudulent blocks, the introduction of ASICs and mining centralization are arguably not as crucial issues as in Bitcoin or other pure PoW networks. However, the development of ASICs will likely contribute to greater security for the network and generate miner interest in DCR.