What long averages do and do not say: their role as structure rather than signal, the regime filter, and a measurement of the well-known EMA 200 versus SMMA 99 debate.
Short averages say what state an asset is in — the first article covered the EMAs 13, 25 and 32. They say nothing, however, about where price will go when that state breaks.
That is the job of the three long averages: MA 100, EMA 200, MA 300. They are not there to decide an entry. They are there to tell you what lies underneath — and to measure how much ground separates price from the next serious reference point.
The most common confusion is to treat all averages the same way, adjusting only the period. They are in fact two distinct instruments.
Short averages describe recent momentum. They move fast, hug price, and act as a trigger: crossed or lost, they change the asset's state.
Long averages describe a position within a cycle. They move slowly — the daily EMA 200 advances by a few tenths of a percent a day — and are almost never crossed cleanly. They act as structure: they indicate where price came from and what it falls back towards.
One practical consequence: never expect a long average to give an entry signal. By the time price has crossed it “decisively”, the move is spent. Its value lies elsewhere — in the fact that it situates.
A moving average has no causal power. If price nevertheless reacts to it, two mechanisms explain why — and they must be distinguished, because they are not equally solid.
The first is accounting. A 200-day average approximates the average price paid by those who entered over that window. Around that level, the unrealised profit of an entire cohort changes sign. And the behaviour of a holder in profit differs from that of a holder at a loss: the first lets it run, the second looks to get out at break-even. That switch creates a real asymmetry of supply and demand, independent of any belief.
The second is reflexive. These three averages are among the most watched on the market. Enough participants place orders and alerts there for reactions to concentrate around them. The level works because it is observed.
This second mechanism deserves an explicit caveat: it is circular, and therefore fragile. It holds as long as collective attention stays on these markers, and weakens if that attention moves. That is a reason to treat these levels as zones of interest, never as barriers.
On a daily chart, 100, 200 and 300 sessions cover roughly a quarter, a half-year and three quarters. That is not an arbitrary split: it matches the horizons most allocations think in.
These three averages are not independent — they measure the same thing with different memories. What matters is therefore not their value in isolation, but how they are arranged relative to one another:
The number of averages crossed is in itself a measure of how mature the move is. An asset above all three is not “better” than an asset above only one — it is simply further along in its cycle, which has opposite implications depending on whether you are trying to enter or to hold.
Among the three, the reference long average — the daily EMA 200 — occupies a place of its own. Not because it is more correct, but because it acts as a regime separator.
The same technical signal is not worth the same on either side of it. Above, pullbacks tend to get bought and breakouts tend to follow through more often. Below, bounces run shorter and breakouts fail more. This is not a law: it is a shift in frequency, which moves the statistics without ever determining an individual case.
The methodological consequence is real: read the position of the leading market before looking at candidates. The same relative-strength filter does not produce the same results depending on the regime it is applied in. That is what the context banner at the top of the screener summarises.
Methods built on these averages use varying definitions of “the 200”: a classic exponential average, or a smoothed average (SMMA, also called RMA) of period 99 — the latter being common in shared scripts, where it is often loosely called “EMA 200”.
Mathematically, an SMMA of period 99 applies a smoothing coefficient of 1/99, the equivalent of an exponential of period 197. The question deserves to be settled by measurement rather than by principle.
Comparison of the two formulas over the last 1,000 daily BTCUSDT candles:
In other words, the choice has no practical consequence for this use. What matters is not which definition you pick but its consistency: two different definitions applied to the same history remain comparable with each other; a definition that changes halfway makes any history unusable.
This conclusion does not carry over to every indicator. On a short oscillator, an equivalent difference in parameters clearly changes the signals — precisely because it applies to a short window, where each candle weighs heavily. The general principle: the longer the window, the more secondary the exact choice of formula becomes.
Three dots complete those of the short averages: 100, 200, 300. Their colour indicates whether price sits above each one, in the selected timeframe.
The useful reading is not dot by dot, but as a whole:
The stretch column reads alongside: it measures how far price is from its averages, scaled by the contract's own volatility. An asset above all six averages but heavily stretched does not offer the same ratio as an asset in the same configuration that has come back into contact.
Three caveats, stated explicitly.
Insufficient history makes the calculation hollow. An MA 300 requires 300 candles. On a contract listed six months ago, the value displayed covers an incomplete window and describes no cycle at all. This is a frequent case on recent perpetuals, and a classic source of wrong conclusions.
These are zones, not lines. Exceeding a level by 0.3 % is not a crossing. Treating these levels to the tick means mistaking the precision of the calculation for the precision of the phenomenon.
Violent moves ignore them. In a liquidation cascade, price goes through all three averages with no noticeable reaction. These markers describe the behaviour of an ordinary market; they do not withstand an order imbalance that overwhelms them.
What remains is what they genuinely provide: a stable frame for situating a price, common to five hundred contracts, and dependent on no one's judgement. That is little, and it is exactly what one expects from a filter — judgement comes afterwards.
This is not investment advice. This article sets out technical-analysis concepts for educational purposes. Past performance is no guide to future performance, and trading crypto-assets carries a risk of losing the entire capital committed. Legal information.