Six models · one reference
Bitcoin Valuation Models Compared: S2F, Rainbow, MVRV & More
By Sahil Kaila · Alphabit founder and risk-model builder. About the author
A Bitcoin chart can describe scarcity, holder cost basis or price momentum. Those are different questions. Here is what six popular models can tell you, where they have fallen short, and how to read them together.
Reviewed October 2, 2026 · Definitions and historical examples, without live price targets
The comparison at a glance
These are six widely used approaches, rather than an exhaustive list. “Useful for” describes an interpretation, not a tested trading advantage.
On a small screen, scroll the table sideways. Full explanations follow below.
| Model / category | What goes in | Useful for | Miss or limitation |
|---|---|---|---|
| Stock-to-FlowSupply model | Existing supply ÷ annual new issuance, fitted to market value | Makes the declining issuance schedule easy to compare across halvings. | Supply scarcity alone does not specify demand or the timing of price changes. |
| Rainbow ChartPrice-history fit | A fitted growth curve with colored price bands | A simple visual overview of price relative to a long-term fitted trend. | The old bands broke in 2022; revised versions change the reference curve. |
| Mayer MultiplePrice extension | BTC price ÷ 200-day simple moving average | Shows how stretched price is against its recent trend. | A low multiple can persist during a falling market; a threshold is not a price floor. |
| MVRV / MVRV Z-ScoreOn-chain valuation | Market cap versus realized cap; Z-Score adds normalization | Adds a network cost-basis proxy to price analysis. | Coins last moving on-chain does not always mean they were bought; extremes do not time reversals. |
| Logarithmic regressionFamily of price fits | A specified relationship between price and elapsed time | Compresses large price changes and shows deviations from a chosen growth path. | Results depend on the equation, start date and fit window. Extrapolation is an assumption. |
| Pi-Cycle TopMomentum trigger | 111-day SMA crossing above 2 × 350-day SMA | Flags unusually rapid price extension with a reproducible rule. | Triggered near April 2021’s local high, then missed the higher November peak. |
1. Stock-to-Flow: a supply hypothesis
PlanB’s March 2019 model fitted Bitcoin’s market value to its stock-to-flow ratio. The original article forecast roughly $55,000 after the May 2020 halving. Keep that original model separate from later S2F variants and other forecasts attributed to its creator. Read the original model.
What it leaves out: the regression has no explicit demand variable. A scheduled reduction in issuance cannot explain every change in liquidity, adoption or selling pressure. This is a structural limitation, rather than a claim that every forecast from every S2F version failed. Evaluating a price target requires the specific formula, forecast date and horizon.
2. Rainbow Chart: an accessible picture of a fitted trend
Rainbow bands make a long price history readable, but their color labels are not measured probabilities. BlockchainCenter documents the old chart breaking below its bands in 2022, a V2 curve published on November 21, 2022, and a newer dynamic regression. Its creator describes the chart as a playful historical view, not a price predictor. Read the chart’s history and current method.
What went wrong: a band that appeared to contain earlier prices did not keep containing future prices. A revised curve can help describe new data, but judging old decisions against today’s revised bands would introduce hindsight. Save the version that existed when the signal was observed.
3. Mayer Multiple: distance from the recent trend
Divide price by its 200-day simple moving average. A value of 1 means price equals that average; 2 means price is twice it. This is an interpretable measure of extension, not an estimate of intrinsic value. See the chart publisher’s definition.
Where interpretation fails: below-average price does not imply the decline is over. The average itself can keep falling. A fixed threshold selected from earlier cycles may behave differently later, so “cheap relative to trend” needs a separate position-sizing and drawdown rule. Read the Mayer explainer.
4. MVRV: market value versus an on-chain cost-basis proxy
MVRV divides market capitalization by realized capitalization. Realized cap values coins at the price when they last moved on-chain. It therefore adds information beyond a price moving average, although transfers between a person’s own wallets are not necessarily purchases. Glassnode explores both extreme levels and time spent above or below them. Read Glassnode’s MVRV analysis.
MVRV Z-Score is a different calculation: market cap minus realized cap, divided by the standard deviation of market cap in the conventional formulation. Check the provider’s history window and any rolling variant before comparing thresholds. See Glassnode’s Z-Score definition.
Where interpretation fails: an extreme reading describes current unrealized profit or loss; it does not supply a reversal date. Treating realized price as a guaranteed support level, or a Z-Score as a literal probability of a crash, goes beyond what the formula measures. Read the MVRV Z-Score explainer.
5. Logarithmic regression: check the actual equation
This label covers multiple approaches. A logarithmic price axis changes the display; a regression specifies a fitted relationship. A power law in time and a logarithmic growth equation are not automatically the same model. BlockchainCenter’s own history shows how different growth curves can sit underneath a similar-looking chart.
Where interpretation fails: a good fit to the same observations used to choose the equation does not establish forecast accuracy. Start date, excluded observations, band widths and refitting can all change the apparent signal. Future bands assume the chosen relationship continues. Explore Alphabit’s regression chart and check its stated method before interpreting the bands.
6. Pi-Cycle Top: a specific crossover, with a small sample
Philip Swift created Pi-Cycle Top in April 2019. It triggers when the 111-day simple moving average crosses above twice the 350-day average. Earlier historical matches therefore predate publication. Read the creator’s formula and publication date.
What it missed: Alphabit’s daily-close reconstruction records a signal in April 2021 but none for the higher November 2021 peak. That makes April a useful local-top signal, not a complete call of that cycle’s highest price. Our checked daily series starts in September 2014; we do not independently verify the widely cited 2013 matches. See the reconstruction and its coverage limits.
A silent crossover does not establish that risk is low. Counting only successful triggers while ignoring tops with no trigger gives an incomplete assessment.
How to judge any Bitcoin model fairly
- Define the claim. Is it describing valuation, calling a peak, or forecasting a price by a date? Test that specific claim.
- Freeze the rules. Keep the original formula, thresholds, fit window and data version. Distinguish publication from earlier historical fits.
- Count misses as well as hits. Include false alerts, peaks with no alert, waiting time and drawdowns after a supposed cheap signal.
- Use later data. A test on observations unavailable when the rules were chosen is more informative than a fit to the development sample.
- Check overlapping inputs. Mayer and Pi-Cycle share price history. Rainbow and regression can share a growth fit. Agreement is not automatically independent confirmation.
Comparing trading strategies also requires explicit contribution schedules, fees and rules. A valuation chart by itself does not establish portfolio performance.
Frequently asked questions
- Which Bitcoin valuation model is best?
- There is no established universal winner. MVRV describes on-chain cost basis, Mayer describes price extension, Pi-Cycle is a crossover trigger, and Rainbow or regression charts describe fitted trends. A fair ranking needs a defined goal and a test on data that was unavailable when the rules were chosen.
- Are MVRV and MVRV Z-Score the same?
- No. MVRV is market capitalization divided by realized capitalization. The conventional MVRV Z-Score normalizes the difference between those capitalizations using the standard deviation of market capitalization. Their scales and thresholds cannot be used interchangeably.
- Can the Bitcoin Rainbow Chart predict a future price?
- A projected band is an extrapolation of a fitted curve. BlockchainCenter describes its Rainbow Chart as a way to view historical price movements and says it cannot predict Bitcoin’s price. Its curve has also changed between versions.
- Do several agreeing indicators make a signal more reliable?
- They can add context, but agreement alone is not proof. Mayer, Pi-Cycle and price regression all use Bitcoin price, while Rainbow and regression may share a fitted growth curve. Overlapping inputs can repeat the same information.
Educational reference. Model limitations above include our interpretation of the formulas; documented historical misses are identified separately. See the disclosure.