Ghost Alpha

← Learn

Mid-range valuation, coins leaving exchanges

Mid-range valuation, coins leaving exchanges

What this means

MVRV sits in the middle stretch of its own history - neither near holders’ aggregate cost basis nor stretched well above it - while BTC is moving off exchanges on net. This has been the single most common state in this classifier’s backfilled history, and its longest confirmed periods run well past a month.

The v1 rule

Valuation reads “mid-range” when MVRV’s percentile against its own full history is above the 25th percentile and at or below the 75th percentile, and MVRV itself is at or above 1.0. Flow direction reads “accumulating” when the net exchange flow (30-day) and exchange reserve change (90-day) signals combine, under v1’s two-vote system, to a net negative reading: more BTC leaving exchanges than arriving. This is a pinned v1 classifier rule, not a permanent definition - a future version could set these thresholds differently.

What would change the state

Valuation would need MVRV’s percentile to fall to the 25th percentile or below (or MVRV itself under 1.0) to move to below-basis, or to rise above the 75th percentile to move to stretched. Separately, the flow and reserve signals would need to flip to a net positive reading - more BTC arriving at exchanges than leaving - for the direction label to flip to distributing. The two dimensions move independently; either can change while the other holds.

Entered 26 times in the comparison window.

Closed periods measured (N)
25
Mean duration
26.7 days
Median duration
14.0 days
Shortest / longest
3 / 86 days

Comparisons use the institutional-era market from August 2020 onward; a metric's own percentile-in-history uses full available history.

Of 25 prior periods in this state, what followed
HorizonMedianQ1 / Q3Worst case
30d+5.5%-1.7% / +11.0%-13.6%
90d+2.3%-14.7% / +35.7%-31.6%
180d-3.0%-30.5% / +38.6%-47.8%
Every prior period (25)

Comparisons use the institutional-era market from August 2020 onward; a metric's own percentile-in-history uses full available history.

Conditioned on: core state (valuation, flows, leverage, halving position); sentiment and short-history modifiers excluded.

Description of the past, not a forecast.

Period30d90d180d
2026-08-19 – present+16.9%n/a - not yet resolvedn/a - not yet resolved
2026-04-07 – 2026-06-02+11.0%-11.1%n/a - not yet resolved
2026-03-11 – 2026-03-26+3.8%-12.3%+12.5%
2025-12-04 – 2026-01-31-1.7%-21.2%-27.8%
2025-11-18 – 2025-11-21-8.0%-25.9%-16.5%
2025-09-03 – 2025-11-14+9.5%-18.1%-38.3%
2025-08-24 – 2025-08-28-1.3%-25.3%-40.1%
2025-08-15 – 2025-08-18-1.6%-14.7%-42.9%
2025-06-12 – 2025-07-09+10.7%+7.5%-12.4%
2025-06-02 – 2025-06-09+2.9%+2.3%-14.2%
2025-03-27 – 2025-05-08+8.6%+23.0%+28.5%
2025-02-05 – 2025-03-06-10.4%+0.1%+19.3%
2024-10-25 – 2024-11-07+47.9%+57.2%+41.4%
2024-09-11 – 2024-09-25+8.8%+68.7%+37.6%
2024-04-15 – 2024-04-20+4.6%-3.8%-0.3%
2024-02-21 – 2024-02-26+22.7%+35.7%+14.8%
2023-11-28 – 2023-12-07+12.8%+44.2%+80.9%
2023-10-18 – 2023-10-23+29.1%+52.4%+123.8%
2023-08-14 – 2023-10-02-10.8%+26.0%+62.5%
2023-06-10 – 2023-07-22+17.5%+0.2%+67.4%
2022-02-07 – 2022-05-04-4.3%-22.4%-47.6%
2022-01-27 – 2022-01-30+5.5%+5.9%-42.8%
2021-11-22 – 2021-11-26-13.6%-31.6%-47.8%
2021-09-07 – 2021-10-05+14.8%+7.9%-18.0%
2021-08-17 – 2021-08-20+6.9%+42.6%-5.6%
2020-08-15 – 2020-11-08-10.1%+37.6%+303.6%

Every period is listed here, including the ones where the base rate didn't hold.