Mining pools are key players in the bitcoin ecosystem. They accumulate the bulk of the network’s computing power and directly affect both security and performance. Despite many studies on the architecture of the bitcoin network, the behavior of the mining pools themselves is still not well understood. Questions like:
- Why do pools mine empty blocks?
- How do they choose transactions to include in a block?
- What affects their revenue?
- How do their actions affect ordinary users (e.g., in terms of confirmation rates and fees)?
In this paper, the authors conducted a systematic study of pooling behavior based on a detailed analysis of:
- over 156,000 blocks (including ~257 million transactions) between February 2016 and January 2019;
- over 120 million unconfirmed transactions collected in real-time from March 2018 to January 2019.
Key findings of the study:
Centralization of power
Despite the decentralized nature of the network, a small number of mining pools consistently control the bulk of the hashrate. This raises concerns about the sustainability and independence of the network.
The prisoner’s dilemma between pools
Pools are in a state of constant competition for hashrate share. This is similar to the prisoner’s dilemma: even if all pools would benefit from a slowdown in capacity expansion, each of them individually continues to invest in growth in order not to lose in the short term. As a result, computing resources grow while each participant’s profitability decreases.
The Malthus Trap
The authors point out that there is a limit after which block rewards (in bitcoins) become insufficient to pay off the exponential growth of hardware and electricity. This is referred to as the Malthus trap of mining.
Price and commissions are weakly responsive to chalving
Analysis has shown that neither the BTC market price nor the average transaction fee shows an immediate or pronounced response to halving events (decreasing the reward per block).
Empty blocks are mined faster
An interesting feature: the intervals between empty blocks are on average significantly shorter than those between full blocks. This may be due to speed optimization, when pools deliberately publish blocks without transactions for the sake of time gain.
Feerate is the main criterion for transaction selection
Pools select transactions per block based on the rate of fee per byte (feerate), prioritizing the more profitable ones. This confirms an aggressive strategy to optimize revenue through transaction selection.
Why it matters?
This study provides a clear picture of the real-world practices of mining pools and allows us to:
- better understand the economic incentives in the network;
- identify the risks of centralization;
- assess Bitcoin’s resilience to load and attacks;
- generate ideas for improving the protocol or creating alternative reward distribution models.
