The dial is a description of market state computed from public prices with the fixed formula below. It is not a timing service and not advice to enter or exit anything. Regime filters reduce some risks and create others, notably whipsaw costs, and both sides are documented here. Decisions are yours.
If crypto has one honest lesson, it is that the market has weather, and that most portfolio destruction happens to people caught fully exposed in the wrong season. The Regime Dial is our published, fixed, recomputable answer to a single question: is this market currently in a condition where trend-following long exposure has historically been survivable, or not?
The formula, exactly
Two ingredients, both computed from 00:00 UTC daily closes, both displayed with their live values so you always see the inputs, not just the verdict. Ingredient one, the BTC trend stack: is BTC's close above its 50-day simple moving average, and is the 20-day average above the 50-day? Ingredient two, breadth: what percentage of the monitor's top-100 universe closed above its own 50-day average? The dial then reads as follows. RISK-ON: both trend-stack conditions true and breadth at or above 40%. RISK-OFF: both trend-stack conditions false, or breadth below 25%. NEUTRAL: everything else. Three states, two moving averages, one participation ratio, thresholds fixed and stated. There is nothing else in it, and that is the point: a dial that gets re-tuned whenever it embarrasses its author is not an instrument, it is a press office.
Why BTC anchors the whole market's regime
Because in stress, crypto is one trade. Altcoin correlations to BTC rise toward one in every drawdown episode, which means an alt-heavy portfolio's true regime exposure is BTC's regime whatever the holdings claim. That is also why position-count diversification failed in our founding experiment, per the transplant write-up. Breadth is the second ingredient because a BTC-only reading misses one specific and dangerous state: majors holding firm while the rest of the board quietly deteriorates underneath, historically the anteroom of the worst months for momentum exposure. Direction from the reference asset, participation from the whole board.
The evidence, honestly framed
The strongest evidence we can offer is the experiment we ran before this site existed. An untuned equity breakout system lost 22% over five years of crypto data taken whole, but the identical system permitted to enter only while the BTC trend stack was aligned upward returned 2.09x with a drawdown reduced from -74% to -59%. One filter, chosen in advance, never iterated, contributed more than every tuned parameter in the donor system combined. Public history rhymes: a disciplined 50/200-day framework kept a mechanical follower out of most of 2022's collapse, the year BTC fell roughly three quarters from its 2021 peak, at the cost of some false exits in the choppy recovery that followed. We attribute most of the gated experiment's RETURN to the regimes it kept rather than to any cleverness, a distinction unpacked in the regime-versus-skill argument; what the gate demonstrably provided was survivability.
The cost side: whipsaw, quantified conceptually
Every regime filter pays for its protection in false alarms, and the shape of the cost is predictable. In trending years the dial flips rarely and earns its keep; in ranging years the trend stack can cross and re-cross monthly, and a mechanical follower pays a spread-and-slippage toll each time while missing whatever the chop eventually resolves into. In our gated experiment the filter sat out roughly 40% of days, and some of those days were up days; that forgone upside IS the premium. Anyone who markets a regime signal without stating this cost is selling insurance while hiding the premium. The dial's design accepts the premium consciously, because in this asset class the catastrophic tail it protects against is not a hypothetical: it arrives roughly once per cycle.
Against the alternatives
Why not the plain 200-day average? It is slower to false-alarm but historically surrenders a third of a bear market before reacting; the 20/50 stack trades earlier warning for more noise, and adding breadth filters part of that noise back out. Why not BTC dominance, funding rates, or on-chain flows? Each encodes real information, and each is either structurally ambiguous (dominance rises in both panic and BTC-led bulls), data-fragile, or unavailable at consistent quality across our whole window. The dial's ingredients were chosen for being boring, universal, and impossible to misreport, which is what a published instrument needs more than sophistication. Richer regime models are a research question for the derivation program, not a reason to make the public dial less checkable.
Publication and proof
The dial updates once daily at 00:05 UTC from closed bars, displays all three input readings alongside the state, and is committed each day to the public attestation chain with the monitor snapshot, so claims like "the dial read RISK-OFF through month X" are verifiable against hashes rather than against our memory, per the attestation explainer. The formula on this page is versioned with a change policy of rarely, loudly, forward-only. The daily email carries the same state with the same timestamp discipline.
How to read it as a human
The dial is a barometer, not a butler. RISK-ON does not mean any particular coin rises; it means conditions under which trend exposure has historically been survivable are present. RISK-OFF predicts nothing; it observes that trend and participation have already deteriorated, and that historically similar readings preceded environments where cash and patience outperformed courage. People who treat weather reports as certainties get wet; people who refuse to look at them get wetter. The dial exists so one fixed, dated, unchangeable formula can stand between you and the loudest voice on your feed. What you do downstream of it is, by design and by conviction, your decision.