Signals live soon

We don't tell you
what to buy.
We tell you when
to get out.

Daily regime signals for crypto portfolios. One signal per coin per day. HOLD or CASH. Every signal is timestamped to Bitcoin at publication, so it can't be edited or back-dated after the fact.

Crypto goes up 300%.
Then it drops 76%.

Most holders ride the wave up and absorb the full crash on the way down. Not for lack of discipline, but because nothing tells them when the regime has turned. By the time it feels bad enough to sell, it's usually too late.

BTC buy & hold, 2022
-64.2%

You put in $10,000. You watched it shrink to $3,580. You didn't know if it would go to zero. Many people sold near the bottom, locking in the loss.

Unhodl signals, 2022
-8.9%

The signal flipped coins to CASH as the regime broke down in early 2022. You sat out most of the crash. Following every signal across the covered coins, your $10,000 became $9,110.

One coin. One daily call.

Every day, for every covered coin, Unhodl publishes a single regime state. There are only two, and what you do with them is always your decision.

HOLD

Market conditions for that coin currently support staying exposed. HOLD is not a recommendation to buy; it means no risk-off signal is active for that coin.

CASH

The regime for that coin has turned adverse. CASH is the signal to consider stepping aside. Whether you reduce, exit, or wait before entering is up to you.

Your decision, always

Unhodl doesn't pick coins, predict prices, execute trades, or hold your funds. It publishes one regime state per coin per day. Nothing more.

Backtest results.
Costs included.

Backtested Sep 2020 – May 2026 with real transaction costs (30 basis points per side) and explicit look-ahead guards.

Metric BTC Buy & HoldBTC Unhodl SignalsSignals
Worst drawdown -76.6% -26.7%
2022 performance -64.2% -8.9%

Figures are from our reference portfolio following the signals across all covered coins (equal-weight). They are not BTC-with-signals, and not a claim that the signal improves any other strategy.

Live signal ledger: Every signal published after launch will be timestamped with OpenTimestamps and independently verifiable against Bitcoin.

Built on math, not promises.

The core idea, trend-following regime gates, is built on decades of published research. The exact parameters are ours, found and stress-tested in our own backtests. No AI, no black box: deterministic rules, evaluated the same way every day.

Downside protection, not alpha

What holds up across the backtest isn't a return edge, it's the defense: the exit logic kept the worst drawdown far below buy-and-hold. We don't claim alpha. The live ledger will let you verify whether the defense survives reality.

Bitcoin-verified signals

Every signal is anchored to Bitcoin's blockchain via OpenTimestamps. Once published, a signal can't be edited or back-dated. Anyone can verify it against Bitcoin.

We tried to destroy our own number

We put our own backtest through an adversarial stress-test, hunting for every way the number was inflated. It worked: the headline was overstated. The honest Sharpe sits closer to 0.9, inside a wide confidence interval, not the impressive figure a raw backtest suggests. We show the honest number and its caveats in full below.

See the honest number →

We'd rather be honest
than impressive.

The claim we lead with is downside protection: a -26.7% worst drawdown against Bitcoin's -76.6%. The return number people ask about, the Sharpe ratio, is far less certain. So we show it the only honest way: as a range, with its caveats, never as one confident figure.

Sharpe, full-window backtest (conservative anchor) ~ 0.9
Sharpe, full-window backtest (point estimate, 95% CI) 1.29 [0.32, 2.16]
  • Weak statistical significance. The result does not clear a 0.05 bar once you account for how many configurations were tested. We treat it as suggestive, not proven.
  • Survivorship bias. Our own estimate is that survivorship inflates the Sharpe by roughly 0.10 to 0.30. The honest number is net of that doubt.
  • Return concentration. About 63.6% of the backtest return landed in the first 282 days, an early bull run. Because so much rode on that window, we don't treat the return edge as established.

This is a pre-live backtest estimate, not a promise. The number that actually matters is the live ledger, anchored to Bitcoin, and you'll verify that yourself.

Read it yourself,
or feed it to your bot.

Investors read the daily signal and decide whether to stay exposed. Trading systems consume the same signal through a JSON API and apply their own execution policy. Same signal, same public record, same verification. One product, two ways to consume it.

For bots & agents →

Signals go live soon.

Be the first to know. No spam. Just a single email when we launch.