Model Fit Score

Model Fit

Each dot is one trading day; colour deepens to BTC orange over time. The orange line is the OLS regression on log-log axes. A clean proxy hugs the line tightly (high R²) with a slope near 1.0.

Rolling Model Fit

Rolling log-log R² over the selected lookback window. Useful for spotting recognition events (R² rising) or regime breaks (R² collapsing). Dashed line = 0.50 dropdown threshold. Dotted line = 0.40 recognition gate.

Leaderboard

Top 25 BTC-holding companies, ranked by lifetime log-log R². Rows below the 0.40 recognition gate are dimmed — these are companies that happen to hold Bitcoin, not Bitcoin plays. Click any row to focus on that company.

Company / Ticker	R²	Slope	Classification	n	BTC stack

What Model Fit Measures

What R² means

A pure Bitcoin treasury play is one where the share price is primarily driven by BTC value per share:

Share Price = mNAV × BTC Value per Share

Model Fit is the R² of that relationship — how tightly the equity tracks the line over its lifetime since BTC adoption. A high R² means the equity is functioning as a clean BTC proxy; the market is pricing it on its Bitcoin stack. A low R² means the share price is still being driven by something else — non-BTC business, regulatory baggage, geographic discount, or no story yet.

What slope means

Slope is the second number worth reading. It's the log-log elasticity — the percentage change in share price per 1% change in BTC value per share. Slope ≈ 1 (0.6–1.8) means a clean proxy with a stable premium. Below 0.6 means dampened — the premium has been compressing as the BTC stack outgrows the equity multiple. Above 1.8 means amplified — the equity moves harder than the underlying NAV, often a small-cap with rising premium or a recognition catalyst. At or below 0 means broken — the equity has moved opposite to BTC NAV per share over the regression window. Slope is a lifetime average; the current mNAV ratio is a separate (instantaneous) snapshot and tells you a different thing.

Classification taxonomy

The leaderboard groups every company into a tier using R² and slope together. Proxy range is slope ∈ [0.6, 1.8]:

Pure BTC Proxy. R² ≥ 0.80 and slope in proxy range. The equity tracks BTC value per share tightly and at proportional elasticity — the market is pricing it on its Bitcoin stack, full stop.
Amplified / Tracks Dampened. R² ≥ 0.80 but slope outside proxy range — amplified above 1.8 (the equity moves harder than NAV), dampened below 0.6 (premium compressing as the stack outgrows the multiple).
Confirmed BTC Proxy. R² ≥ 0.60 and slope in proxy range. Tracks BTC NAV cleanly but with more noise. Most established treasury companies live here.
Confirmed Fit (Off-Slope). R² ≥ 0.60 with slope outside proxy range. The market is pricing BTC in, but at a dampened or amplified elasticity.
Emerging Fit. R² ∈ [0.50, 0.60). The pattern is locking in but isn't a confirmed proxy yet. Watch for graduation.
Emerging, Hidden by Default. R² ∈ [0.40, 0.50). Above the recognition gate but statistically loose; dimmed unless you raise the filter.
Below Recognition Gate. R² < 0.40. Companies that happen to hold Bitcoin, not Bitcoin plays. Dimmed in the leaderboard by default.
Inverted / Broken. Slope ≤ 0. The equity has moved opposite to BTC NAV per share over the regression window — not a proxy at all.

When the score actually matters

Below 0.40 R², ignore. Even if BTC fundamentals look strong, the market hasn't started pricing the company as a BTC play, so capital is better deployed in companies whose prices have already connected with their NAV. Strong fundamentals plus low fit means waiting; strong fundamentals plus rising fit means a candidate.

The four-gate framework

Model Fit is the first of four gates in the broader allocation framework. The others — BTC stack, forecast CAGR, and days to flippening — only earn weighting once Model Fit clears the gate. Full method on mNAV Insights.

What we're watching

How this page connects

Each company's underlying BTC accumulation history is at BTC Stacking Curve — filter by ticker for the Richards/Gompertz fit and the per-week buys.
The cohort accumulation race is at BTC Stacking Race — same companies, ranked by stacking pace.
The Bitcoin price model this work measures against is at BTC Power-Law.
For Strategy specifically (the leaderboard #1), the per-week ATM mechanics live at Strategy ATM.
Click any leaderboard row above to focus this page on that company; click the company ticker to jump to its Stacking Curve drilldown.