Thread by @ParkerMerritt

𝗙𝗢𝗟𝗟𝗢𝗪𝗜𝗡𝗚 𝗙𝗟𝗢𝗪𝗦 𝗩: 𝗣𝗢𝗢𝗟 𝗖𝗥𝗢𝗦𝗦-𝗣𝗢𝗟𝗟𝗜𝗡𝗔𝗧𝗜𝗢𝗡
In today's @coinmetrics State of the Network, we examine on-chain flows b/t mining pools and their counterparties, uncovering hidden links to help assess the true degree of mining decentralization 🔍

Any proper mining pool analysis starts with the "0-hop" and "1-hop" heuristic.

0-hop addresses (typically pools) are the direct recipients of new BTC issuance, while 1-hop addresses (typically miners) are any address linked to pools by at least 1 outflow.

To link 0-hops to pools, we make use of the "coinbase signature." When pools mine blocks, they insert arbitrary data into the $BTC issuance tx. By standard, most include a claim of ownership.

@coinmetrics exposes this via ATLAS's extra_data field, easily mapping blocks to pools.

Looking at the flow of issuance & transaction fees to mining pools, there's already troubling signs of centralization.

In 2023, the top 2 pools (Foundry & AntPool) earned the majority of mining rewards— around 53% between the two.

Generally, links b/t 0-hop addresses & pools are mutually exclusive, each controlling an isolated set.

However, some pools exhibit overlap. Examining the 7 0-hops associated w/ Binance Pool & BTC​.com, 2 addresses received payouts from both pools.

The overlap is brief, with 22 BTC​.com blocks paying out to Binance Pool addresses in 2023.

Still, any shared flows lend credence to the theory that Binance Pool is white-label BTC​.com, complicating efforts to directly measure pool centralization.

After taking custody in a 0-hop address, pools distribute rewards to their constituent miners.

Still, we can't assume all 1-hop addresses = miners.

Some miners withdraw earnings directly to an exchange, providing optionality to liquidate BTC & cover operating costs.

Binance, Coinbase, & Kraken are top benefactors, receiving 21k BTC. U.S. pools & exchanges have clear alignment— Foundry made up 89% of pool flows to Coinbase, Kraken, & Gemini.

However, Seychelles-based MEXC initially appears to be shockingly dominant in pool-to-exchange flows.

Other sources tie the MEXC addresses to Cobo​.com, a custodian led by F2Pool founder Discus Fish. We’ll call this group the “MEXC/Cobo cluster.”

Since Jan. 2023, this cluster received 63.6k BTC via pool payouts. AntPool contributed 45.7k BTC—nearly 47% of their total inflow.

Many pool-to-exchange flows contain inputs from multiple pools.

Based on the “common input heuristic," this tx structure typically hints a set of addresses shares a common owner.

At the very least, it suggests a high degree of coordination b/t pools moving funds to exchanges.

Since 2023, there were 1,200+ pool-to-exchange flows w/ shared inputs. Contributors include AntPool, Binance Pool, F2Pool, Luxor, BTC​.com, Braiins, & Poolin.

Most also include inputs from MEXC/Cobo, indicating the entities play a hands-on role in consolidating outflows.

Looking at the remaining 1-hop addresses, there’s a puzzling amount of transfers from multiple pools to a singular address, 3BH…WGb.

Last year, analysts at @TheMinerMag_ spotted this suspicious consolidation of funds, pinning the blame on AntPool.
theminermag.com/news/2023-12-2…

@TheMinerMag_ On top of +36.3k BTC from AntPool, 3BH…WGb received contributions from 7 of the other top mining pools.

In aggregate, 57.4k BTC flowed from pools to this address, making up around 18% of the 1-hop flow total (excluding exchanges) since Jan. 2023.

3BH…WGb doesn’t appear to share inputs with other pools/exchanges, making its ownership difficult to pinpoint via heuristical clustering.

Nonetheless, the consolidation of funds into this address implies the existence of an entity backstopping multiple pool operations.
atlas.coinmetrics.io/address-detail…

@TheMinerMag_ Many pools offer Full Pay Per Share (FPPS), sending miners consistent compensation regardless of whether the pool mines blocks.

While convenient for miners, FPPS introduces liquidity risk, as short-term variance in “pool luck” can throw flows off balance.
minebest.com/blog/pps-vs-fp…

Poolin's collapse in 2022 was largely attributed to a FPPS imbalance, with the illiquid pool ultimately suspending withdrawals and triggering an 80% exodus in pool hashrate.
coinmetrics.substack.com/p/state-of-the…

Still, pools have had little choice but to embrace FPPS. In 2022, $RIOT broke off from Braiins, a pool offering Pay Per Last N Shares (PPLNS). After a period of bad luck, on-chain flows show Riot switched to Foundry, the top FPPS pool. A year later, Braiins switched to FPPS.
coinmetrics.substack.com/p/state-of-the…

For small pools, absorbing long bouts of bad luck requires a larger, liquid partner. 3BH…WGb appears to be this partner.

3BH…WGb also receives many flows from the MEXC/Cobo cluster. The common thread b/t these entities is Loop, an asset transfer network created by Cobo.

Given FPPS challenges, pools likely leverage Loop for payout mgmt. This reduces liquidity risk, but adds counterparty risk, inserting another middleman into mining.

Miners can switch pools, but contributors oversee ~50% of hashrate, leaving few choices for fully opting out.

Pool centralization is a big concern in the Bitcoin community, with Foundry/AntPool dominance elevating censorship & network disruption risk.

Though the remaining hashrate is well-distributed, links b/t pools are a concern, signaling a hidden consolidation of power in mining.

I hope you found this thread insightful, but check out the full @coinmetrics report for a more detailed breakdown of our latest mining flows analysis!

And be sure to subscribe— lots of great mining pieces in the pipeline heading into the halving ⛏️
coinmetrics.substack.com/p/state-of-the…