Is bitcoin’s mining cost a price floor?
Not as a trading signal. From January 2015 to August 2026, rules that held bitcoin when its price was cheap relative to the estimated cost of mining reached a Sharpe ratio of 0.76 at best, below a rule that knows only the number of days since the last halving (1.07). Each halving doubles the cost overnight, so “cheap versus cost” mostly means “shortly after a halving”.
The claim
Hayes models bitcoin’s marginal cost of production from electricity prices, mining-hardware efficiency and network difficulty, and argues that it anchors the price. In the 2019 back-test, regressions and vector autoregressions show the cost explaining bitcoin’s price, the 2017 bubble resolving by converging back to the model, and prices that “tend to this bound and not collapse to zero”. Mining cost is now widely quoted as bitcoin’s floor.
What we tested
Daily bitcoin closes against a public estimate of the network’s electricity cost per coin, January 2015 to August 2026, a window with three halvings. We ranked the price-to-cost ratio against its own trailing history — one year, two years, and within each halving epoch — and tested 24 rules that hold bitcoin when the ratio is cheap or expensive, with costs. The control was a “calendar twin”: the same ranking built from nothing but the days since the last halving, with no price or energy data.
Result
The best cost-based rule had a Sharpe ratio of 0.76, buy-and-hold over the same window 0.70, and the calendar twin 1.07. No rule stands out once the 24 tries are accounted for. The reason is mechanical: at every halving the reward per block halves, the cost per coin doubles, and the ratio drops by half — 1.14 to 0.55 in July 2016, 1.04 to 0.45 in May 2020, 0.95 to 0.50 in April 2024. Post-halving years have historically been strong, so the “value” signal is the halving calendar in disguise, with three events behind it.
Nor is it a floor. At the November 2022 low bitcoin traded at 0.64 times its estimated cost. The estimate also trails the price by about two weeks, because it depends on which machines are profitable to run, and that depends on the price.
Holding rules, January 2015 – August 2026
| Rule | Sharpe ratio |
|---|---|
| Best of 24 price-to-cost rules | 0.76 |
| Buy and hold | 0.70 |
| Calendar twin: days since halving only | 1.07 |
| Price-to-cost ratio at the Dec 2018 low | 0.84 (bottom 10% of range) |
| Price-to-cost ratio at the Nov 2022 low | 0.64 (bottom 1% of range) |
What it means
Production cost describes miner economics well; a ratio near or below 1 says miners are squeezed. It does not time the market: the price has traded far below the estimate, the estimate moves with the price, and the timing value it seems to have comes from the halving calendar.
What would change the verdict
Three halvings are very little to learn from. The next, expected around 2028, adds a fifth epoch, and a cost series built from hardware and power data that do not depend on the bitcoin price would test the floor idea more cleanly.
This note re-tests a claim made by the cited authors on our own crypto market data; the summary of their method is ours. Tested · published · last reviewed . Research, not investment advice.