Which methods apply depends not on the approach but on what kind of market sits under the chart. Two things decide almost everything: whether the instrument has a schedule with a "trading begins" event, and whether it has real volume.
Geometry — structure, pierces, gaps, zones — transfers anywhere. Everything else depends on the data.
What each part requires
Geometry requires nothing but prices. It works on any chart.
The temporal part (killzones, anchors to the open, daily templates) requires an event: a moment before which there was no trading and after which there is. A daily rhythm is not enough — a point is needed.
The volume part (Wyckoff's legacy: effort and result, climaxes) requires real volume, meaning a recorded count of trades rather than an estimate.
Crypto
No schedule. A spot exchange runs without a break: no open, no daily close shared by everyone, no weekends. A daily rhythm does exist — volumes rise through the European and American mornings — but that is a gradual statistical property, not an event a rule can be anchored to.
Volume exists, but in pieces. Each exchange reports its own real volume in the base asset. There is no consolidated volume for the whole market: the same bitcoin trades on dozens of venues, and the figure depends on whose tape you are reading.
What remains: geometry in full. The volume part, with the caveat that volume belongs to one venue. The temporal part has nothing to attach to.
This is the extreme case, and it is exactly why "pure SMC" is so widespread in crypto: not because it works better there, but because the rest does not apply.
Forex
A schedule exists. London, New York, Tokyo — sessions with clear boundaries, weekends with a gap. The temporal part applies, and ICT largely took shape here.
There is no exchange volume at all. The currency market has no central venue, so no recorded count of trades exists. Tick volume — the number of price changes per bar — is usually substituted. That is an estimate of activity rather than turnover, and it can behave differently.
What remains: geometry and time. The volume part only as an approximation you have to keep in mind.
Equities and indices
A schedule exists, and a strict one. Open, close, pre-market, post-market, weekends and holidays. Opening gaps are genuine: price jumps overnight because no trading happened.
Volume is real. The exchange records trades, and that is not an estimate.
What remains: everything. Geometry, the temporal part and the volume part. On equities both SMC markup and the classical Wyckoff schematics with their reliance on volume are usable.
The flip side is that this is also where the most external effects live: dividend dates, earnings, opening and closing auctions. The markup knows nothing about them and quietly marks them as ordinary movement.
Precious metals
It depends on what exactly you are trading, and this is the one class where the answer is ambiguous.
Spot forex (a metal-versus-dollar pair at a broker). Behaves like forex: sessions exist, exchange volume does not, tick volume stands in for it.
Futures (an exchange contract). Behaves like an equity: there is a schedule, there is real volume, there are gaps. Both the temporal and the volume parts are available.
The same metal offers a different toolset depending on the venue. Before applying anything beyond geometry, answer which of the two instruments is on your chart.
The through-line
The closer an instrument is to an exchange with a schedule and real volume, the more methods apply. Equities and exchange futures: the full set. Forex: no volume. Crypto: geometry only.
Hence a practical conclusion usually stated the wrong way round: not "this method works poorly on crypto" but "crypto lacks the data the method requires". The method did not get worse — its foundation is missing.
A common mistake
Carrying a rule across along with the vocabulary. The term always transfers: a killzone can be drawn on any chart. But the rule behind the term rests on an event that a round-the-clock market does not have, and the drawn zone turns out to be a rectangle in an arbitrary place.
A simple check: ask what observable event the rule rests on, and whether your instrument has it.