Research · Floor Convergence
Five Ways to Find the Floor
At cycle bottoms, five independent methods agree on where Bitcoin's value floor is. One predicts it. The others discover it after the fact. That convergence is not a coincidence.
The Predictive Floor
The power law floor is the only method here that requires no price history. It is computed from first principles — a power law fit to Bitcoin's adoption curve — and produces a floor price for any future date without consulting what Bitcoin has actually traded.
The formula
where T is days since genesis (2009-01-03). No price inputs. No moving averages. No on-chain lookups.
This is the reference line everything else gets measured against. All four other methods are evaluated by how closely they track this floor at cycle lows.
Five Methods
Each method measures something different. Cost basis. Thermodynamic energy. Market sentiment. Smoothed price history. None of them were designed to track each other. At cycle bottoms, they all arrive at approximately the same number.
Market consensus smoothed over 200 weeks (~3.8 years). The most widely cited "ultimate support" metric in Bitcoin analysis. Has never been closed below on a monthly basis.
Aggregate cost basis of all circulating coins — each coin valued at the price when it last moved on-chain. MVRV < 1 (price below realized price) has historically marked cycle bottoms within weeks. Requires Glassnode integration.
Market Value to Realized Value ratio. When MVRV = 1, market cap equals aggregate cost basis — every holder is at breakeven. Each cycle bottom has touched or briefly breached this level before reversing. Requires Glassnode integration.
All-in cost to produce one Bitcoin: energy, hardware, overhead, depreciation. Miners selling below this level operate at a loss. Sustained bear markets that push price below all-in cost trigger miner capitulation and are historically self-limiting. Requires hashrate + energy data.
Charles Edwards' thermodynamic model: Bitcoin's fair value derived from the energy expended to produce it and the efficiency of mining hardware over time. Not a moving average or cost basis — a physical derivation. Converges with the power law floor at cycle lows.
Convergence Chart
Power law floor and 200-week MA plotted on a log scale from May 2014 to present, with analytical projections through 2028. Remaining three methods will be added once Glassnode integration is complete.
Red vertical line: 2020-11-28 — the structural crossover when the 200 WMA permanently separated from the floor. After this date the 200 WMA is a midpoint indicator, not a floor proxy.
Key moments — floor vs 200 WMA
| Date | BTC Price | Floor | 200 WMA | WMA / Floor | Context |
|---|---|---|---|---|---|
| Aug 2015 | $230 | $252 | $231 | 0.92× | Cycle 2 bottom |
| Dec 2018 | $3,690 | $2,536 | $3,215 | 1.27× | Cycle 3 bottom |
| Mar 2020 | $6,484 | $4,955 | $5,562 | 1.12× | COVID crash |
| Nov 2020 ★ | $18,114 | $6,881 | $7,331 | 1.065× | ATH breakout / crossover |
| Nov 2022 | $17,169 | $16,632 | $24,050 | 1.45× | FTX — floor touch |
| Dec 2025 | $87,085 | $51,999 | $56,271 | 1.08× | Today |
Why They Converge
Each method approaches Bitcoin's value from a different angle.
The 200-week MA smooths price history until short-cycle noise disappears — what remains is the slow-moving adoption trend. The realized price captures the aggregate willingness-to-hold: when price falls below what everyone paid, panic selling has run its course. MVRV = 1 is the same signal in ratio form. Miner production cost establishes the physical lower bound — below it, production stops, supply contracts, price recovers. The energy value model derives fair value from work done, independent of markets entirely.
These are not the same calculation. They are five different measurements of the same underlying phenomenon: Bitcoin's adoption-driven value floor. At cycle bottoms, price reverts to that floor, and every lagging metric catches up simultaneously. The convergence is what you would expect if the floor is real.
The power law floor predicts where that convergence will occur before it happens. The other methods confirm it after.
The Key Distinction
Predictive vs retrospective
Four of the five methods are lagging. The 200-week MA is an average of the past 200 weeks of prices — it cannot tell you where the floor will be in the next bear market without knowing what prices will be between now and then. Realized price and MVRV are on-chain lookbacks. Miner cost reflects current hashrate and energy prices, not future ones.
The power law floor requires none of this. Given today's date, it produces a floor price. Given a future date, it produces a future floor price. No price history. No on-chain state. No energy data. Just T.
That asymmetry is what makes the floor useful as a financial primitive. You can write a loan against it, price an option off it, or anchor a withdrawal rule to it — because you know what it will be when the payment is due.
What's Pending
The 200-week MA convergence is documented above with full empirical data. Three methods await on-chain data integration, one requires energy/hashrate sourcing.
Pending integrations
- Realized Price — full history vs floor, cycle-by-cycle alignment Glassnode
- MVRV ratio — proximity to 1.0 at cycle lows Glassnode
- Miner all-in production cost — bear market floor alignment Hashrate index
- Energy Value model (Edwards) — thermodynamic floor convergence Capriole
When these datasets are integrated, this page will be updated with the full five-method convergence chart, a quantitative convergence table showing proximity of each method to the power law floor at each historical cycle bottom, and regression analysis of the lead/lag relationships.
Methods & Disclosures
Model: Santostasi power law — β = 5.688, log A = −16.493, genesis 2009-01-03. Floor multiplier 0.432 (N(0.432, 0.051), min 0.314) from Monte Carlo analysis across 5,674 daily closes.
Out-of-sample R²: 0.546. Effective sample size: ~24 (block-bootstrap, autocorrelated residuals). Standard errors are wider than OLS suggests.
200 WMA data: 200-week simple moving average of end-of-week closing prices. Empirical series begins May 2014 (first date with 200 weeks of price history). Analytical series computed assuming price = trend.
This page presents research analysis, not financial advice. Historical floor non-breach does not guarantee future non-breach. The power law model may break or decelerate as Bitcoin's adoption curve matures.