Metcalfe’s Law and the Value of the Bitcoin Network

Modelling Bitcoin's price as a function of the network effect, plotting the latest Metcalfe's Law graph and its variants against BTC price.

  • Bitcoin
  • Network Effects
  • Valuation Models
Status
Archival

Published September 2018. Kept online as a record of the analysis at the time, not current market guidance.

Published

17 September 2018

Length

5 pages

Format

Whitepaper · PDF

Status

Archival

The Analysis

Key Findings

Network value scales with the square of its users.

Metcalfe’s Law, proposed by Bob Metcalfe, founder of Ethernet, in the 1980s. The formula: a network’s value is proportional to n(n-1)/2, the number of all possible connections between n users, asymptotically proportional to n squared.

Wallet count stands in for active users.

The report uses the number of wallets generated on the Bitcoin network as its proxy for user count.

Three rival models challenge Metcalfe’s original exponent.

Sarnoff’s model treats value as directly proportional to users, Briscoe’s as n log(n) rather than n squared, and Wheatley’s finds a log-linear exponent of 1.69, not 2.

All four models fit almost identically well.

Pearson correlation against BTC price: Sarnoff 0.918, Metcalfe 0.918, Briscoe 0.914, Wheatley 0.918.

The report’s own conclusion

Bitcoin was overvalued, even after the crash. “It seems the BTC price is still overvalued at the moment, even after the recent crash in January, and might not be sustainable considering the shrinking growth in the active user data.”

How It Was Done

Method & Sources

Modelled using wallet count as a proxy for active Bitcoin users, comparing four network-value formulas (Sarnoff, Metcalfe, Briscoe, Wheatley) against historical BTC price via Pearson correlation. Data drawn from active Bitcoin address growth and weekly wallet creation figures since 2012.
Looking Back
Since Publication

How This Aged

Looking back, the model was right for a while, but incomplete over the longer term. It correctly identified that Bitcoin was overvalued during the 2017-18 bubble, the price subsequently fell dramatically, so the warning about unsustainable growth was justified at the time.

What it missed was that wallet activity alone isn’t enough to value Bitcoin. Bitcoin evolved from primarily a transaction network into a scarce investment asset attracting far larger amounts of capital per participant, including institutional money. The model also underestimated a feedback loop: falling prices themselves reduce wallet activity, which the original approach didn’t account for.

If I were building this today, I’d still use Metcalfe’s Law, but as one component rather than the whole valuation model, combined with capital per participant, institutional adoption, liquidity, supply dynamics, long-term holder behaviour, and macro conditions. Bitcoin today behaves not just like a network, but like a scarce global financial asset too.

The wallet-count proxy specifically hasn’t held up as cleanly either. Custodial exchanges weakened it fundamentally, one exchange wallet can represent thousands or millions of users, while one individual can control many addresses. I’d still use wallet activity as a network-usage signal, but today I’d favour entity-adjusted active users, combined with transaction value, exchange activity, and capital flows, rather than treating wallet count as a direct measure of adoption.
Take It with You

Read The Full Report

5 pages, charts included. No email required.

Scroll to Top