Is FrAMA built into MetaTrader 5?

Yes. iFrAMA is a standard MQL5 function taking a symbol, a timeframe, the period, an ma_shift and an applied price, and the terminal also ships a FrAMA custom indicator under Indicators\Examples. The Builder calls the built-in function, so a generated EA carries no dependency on any file being present in the terminal that runs it.

What does the period actually control?

How far back the measurement looks — and it is half the window. The dimension needs the recent `period` bars, the `period` before those, and the combined stretch, so a period of 14 reads 28 bars and nothing is drawn until 28 bars exist. It is not a smoothing length. Doubling it does not halve the responsiveness the way it would on an EMA; it changes the stretch of chart the jaggedness is measured over.

How is it different from AMA?

Both replace a fixed smoothing constant with a measured one, and they measure different things. Kaufman's AMA uses an efficiency ratio — how much net distance the price covered against how much total distance it travelled. FrAMA uses the fractal dimension of the curve, which is about how jagged the path is rather than how far it got. A market can travel a long way by a very rough route, and that is exactly where the two disagree.

Why does changing applied price barely do anything?

Because it is not in the part that adapts. The fractal dimension is computed from the bars' highs and lows regardless of the setting; applied price only selects the series that the resulting weight is applied to. On an ordinary moving average applied price is one of only two controls, so the habit of treating it as significant carries over — here it is close to cosmetic.

Can I use the smoothing factor as a signal?

Not from this node. iFrAMA publishes one buffer and it is the average itself; the dimension and the weight are computed inside and discarded. The chart on this page draws the factor because it is what the indicator is about, but a Builder flow cannot read it. If a flow needs an explicit statement that the market is quiet, that has to come from a different indicator.

Can I build a FrAMA EA without coding?

Yes. The FrAMA vs Fixed EMA template puts a FrAMA node at 14 next to a Moving Average node set to EMA 28 — the same window FrAMA reads — and trades the crossings between them, closing the opposite side before it enters. The period, the fixed length and method, the shift, the stop distances, the lot and the spread limit are all EA inputs.

How should I validate settings before going live?

Backtest on quality tick data, then demo, and check two things that are specific to this indicator. First, that the test period is long enough to contain both regimes — an adaptive average tested only in a trend is being judged on the half of its job it does not need to do. Second, that the instrument does not gap much: the weighting is only well behaved while the two half-windows overlap in price, so an instrument with frequent gaps is testing the arithmetic as much as the strategy. And because the period is a window rather than a smoothing length, resist scanning it — the number that wins a scan is describing the sample, not the market.

FrAMA for MT5 traders

What FrAMA tells you

Every moving average is a decision about how much of the newest price to let in. On an EMA that decision is made once, when you pick the length, and never revisited. FrAMA remakes it on every bar. It measures how jagged the recent price curve is — its fractal dimension — and converts that straight into the weight. A market travelling in a clean line scores 1 and the average is handed the whole price, so the line sits exactly on it. A market thrashing back and forth in the same range scores 2 and the weight collapses to a hundredth, so the line goes flat. The length you set does not control the smoothing; it only controls how far back the measurement looks.

There is no fast/slow pair and no smoothing constant — the weight is measured, not chosen
The window is twice the period you set
Only the line is published; the weight it used is not

1 Measure the ranges of the recent bars, the bars before those, and both together
2 Turn the three into a fractal dimension between 1 (a line) and 2 (a scribble)
3 Weight = e^(−4.6 × (dimension − 1)), applied to the newest price

Afficher la formule et les détails de calcul

Every moving average makes one decision: how much of the newest price to let in. On an EMA you make it once, by choosing a length. FrAMA remakes it on every bar, from a measurement:

n1 = the range of the last N bars, n2 = the range of the N before those
n3 = the range of all 2N bars together
d = log₂((n1 + n2) / n3), and the weight is e^(−4.6 × (d − 1))

d is a fractal dimension: how much path the price used to get where it went. A straight line scores 1 and the weight comes out at 1 — the average takes the whole of the new price and becomes it. A pure oscillation scores 2 and the weight collapses to 0.01 — the line goes flat. Both of those are exact, not approximate.

The period you set is half the window. This is the thing worth knowing before choosing a number. Three ranges are needed and they span 2N bars, so iFrAMA with a period of 14 is reading 28. MetaTrader’s own implementation says so out loud: it sets PLOT_DRAW_BEGIN to 2 × period − 1 and refuses to calculate at all while fewer bars exist. That is why the template pairs FrAMA(14) with a fixed average of 28 — so the two lines are measuring the same stretch of chart and only the weighting differs.

Applied price does much less than it looks. The node exposes it and the generated EA passes it to iFrAMA, but the jaggedness is measured from the bars’ own highs and lows, whatever applied price is set to. It selects the series that gets smoothed; it does not change how much smoothing happens. On an ordinary moving average that setting is one of two controls. Here it is close to cosmetic.

In a strong trend the average becomes price. A weight of 1 means no history is carried at all. The line lands on price and rides it — correct behaviour, and quietly fatal to any rule built on price crossing the line, because in exactly the regime such a rule is meant to catch, the two series are identical. Compare the line against something else instead.

Two smaller things follow. The weight is not clamped: it stays inside 0.01 to 1 only while the two half-windows share a price range, and a gap between them sends d below 1 and the weight above it — on a 14-period window a 3-unit gap gives d = −0.81 and a weight near 4,080. And the weight is not published: iFrAMA has a single buffer, and it is the line. The chart above draws the factor because it is the whole point of the indicator, but no flow can branch on it.

Signaux

Les signaux distincts et testables que génère FrAMA — et le régime adapté à chacun.

Cross Against a Fixed Average

Prêt dans Builder

Condition: FrAMA crosses a fixed average of the same window
Meilleur régime: Trending
Utilisation typique: The template's rule — the two separate when the market's character changes

Price Against the Line

Prêt dans Builder

Condition: Price sits above or below FrAMA
Meilleur régime: Trending
Utilisation typique: A direction filter; note it degenerates in a strong trend, where the line is glued to price

The Line Going Flat

Prêt dans Builder

Condition: FrAMA stops sloping while price keeps moving around it
Meilleur régime: Range
Utilisation typique: A range detector — the flatness is the reading, not a side effect

Reading the Smoothing Factor

Logique avancée

Condition: The weight itself moves from its floor toward 1
Meilleur régime: Either regime
Utilisation typique: A regime signal in principle; iFrAMA does not publish the factor, so it cannot be built from this node

Implémentation MT5

Ce que MetaTrader 5 calcule et trace réellement — la référence de chaque règle de cette page.

Buffers

Index	Buffer	Tracé dans MT5 comme	Ce qu'il contient
0	MAIN_LINE	FrAMA	The average itself, and the only series iFrAMA publishes. The fractal dimension and the smoothing factor it produces are computed inside the indicator and never exposed — there is no second buffer to read them from, which is why a flow cannot branch on how adaptive the line currently is.

Notes de plateforme

Fonction MQL5

iFrAMA

Prix appliqué

Dans Builder

The node exposes it and the generated EA passes it straight to iFrAMA, but it does less than it looks. The fractal dimension is measured from the bars' own highs and lows, never from the applied price — so changing this changes which series gets smoothed, and not at all how much smoothing is applied. The Builder pins iFrAMA's own ma_shift to 0 and uses the node's shift as the bar to read instead, which is why there is one shift on the node and not two.

Notes d'implémentation

The generated EA creates one iFrAMA handle and reads buffer 0. Period and applied price both reach it as EA inputs; iFrAMA's fourth argument, ma_shift, is pinned to 0 in the generated source, and the node's own shift becomes the CopyBuffer offset instead. Two things called shift exist in MT5's signature and only one of them is a bar offset, which is why the node exposes one.
The window is 2 × period. n1 is the range of the most recent `period` bars, n2 the range of the `period` before those, n3 the range of all 2 × period together — then d = log2((n1 + n2) / n3) and the weight is e^(−4.6 × (d − 1)). MT5's own implementation sets PLOT_DRAW_BEGIN to 2 × period − 1 and returns early while fewer bars exist, so a period of 14 needs 28 bars before anything appears.
The ranges are read with iHigh and iLow — the bars' own extremes. Applied price is used only in the final smoothing line. This is why changing applied price does not make the average more or less responsive, which is the opposite of what the same setting does on an ordinary moving average.
The weight is not clamped. It stays inside 0.01 to 1 while the two half-windows overlap in price, which is the normal case: a perfectly straight line measures d = 1 and weight 1, a pure oscillation measures d = 2 and weight 0.01. Both were checked against the published formula. If the halves separate entirely — a gap, or a fast one-way move that leaves no overlap — d falls below 1 and the weight rises above it. On a 14-period window a 3-unit gap between the halves gives d = −0.81 and a weight of about 4,080, and a recursion with a weight above 1 does not converge.
The smoothing factor is not published. iFrAMA has a single buffer and it is the line, so a flow cannot test how adaptive the average currently is — only what it produced. The hero chart above draws the factor because it is the point of the indicator; the Builder cannot branch on it, and the page marks it as derived rather than as a second buffer.

Fonctionne mieux / utiliser avec prudence

Aucun indicateur n'est un avantage universel. Voici où FrAMA aide — et où il induit en erreur.

Fonctionne mieux

Markets that alternate between trending and going quiet, which is what the adaptation is for
As the trend side of a pair, where a fixed average supplies the reference it is measured against
Higher timeframes, where a 28-bar window covers a meaningful stretch
Anywhere a fixed length has been re-tuned repeatedly — this replaces that tuning with a measurement

Utiliser avec prudence

As a price-cross trigger in a strong trend: the weight reaches 1, the line lands on price, and crossings become noise
Short histories — nothing is drawn at all until twice the period has passed
Assuming applied price changes the responsiveness; it does not, it only changes what gets smoothed
Instruments that gap: the dimension is only bounded while the two half-windows overlap in price

Créer une stratégie FrAMA

Connectez le signal aux règles d'entrée et de sortie, puis exportez un EA MT5 compilable — sans code.

One FrAMA node at 14 — so the window it reads is 28 bars
One Moving Average node, EMA, set to that same 28
FrAMA crosses above the fixed average → Close the short, then Open Buy · SL 70 / TP 210
FrAMA crosses below it → Close the long, then Open Sell
Both read the last closed bar, so a cross is a settled fact

Combiner FrAMA avec d'autres indicateurs

Un indicateur fonctionne rarement seul. Ces associations couvrent les angles morts de FrAMA.

FrAMA + AMA

AMA rising
FrAMA crosses above the fixed average
Buy

Pourquoi: The other adaptive average on this site, and the pairing is instructive because they adapt to different measurements rather than to different degrees. Kaufman's version reads an efficiency ratio — net travel divided by total travel — so it responds to how directional the move was. This one reads the fractal dimension of the curve, so it responds to how jagged the path was. A market can be directional and jagged at once, which is exactly where the two disagree, and that disagreement is more informative than either line alone
Meilleur régime: Either regime

No single-flow template pairs them. Start from FrAMA vs Fixed EMA and add an AMA node with a Compare against a second AMA at a further shift, joined to the Cross with an And gate.

Ouvrir Builder →

FrAMA + ADX

ADX above 25
FrAMA crosses above the fixed average
Buy

Pourquoi: FrAMA already knows when the market is quiet — that is what the flat line means — but it has no way to say so to the rest of a flow, because the smoothing factor is not published. A trend-strength reading supplies the missing output: it states in a number what the flat line only shows visually, and it is the practical answer to this indicator's one structural limitation
Meilleur régime: Trending

No single-flow template combines them. Add an ADX node with a Compare at 25 to FrAMA vs Fixed EMA and join it to the Cross with an And gate.

Ouvrir Builder →

FrAMA + Moving Average

A fixed EMA over 28 bars
FrAMA over the same 28 bars
Cross

Pourquoi: The comparison the template is built on, and worth understanding before anything is added to it. A plain average weights every bar in its window identically whatever the market is doing; FrAMA reweights on each bar from a measurement. Running them over the same window length isolates the adaptation as the only difference between the two lines, which is the cleanest way to see what this indicator actually contributes
Meilleur régime: Trending

This is the FrAMA vs Fixed EMA template itself. Change the fixed side's method from EMA to SMA to see how much of the signal came from the weighting rather than the adaptation.

Ouvrir Builder →

FrAMA + DEMA

FrAMA above the fixed average
Fast DEMA crosses above slow
Buy

Pourquoi: The two ways of attacking lag, and they are genuinely opposite. This page's line varies how much smoothing to apply, easing off when the market measures as directional; DEMA leaves the smoothing untouched and subtracts an estimate of the lag instead. The difference shows in ranges — an adaptive line goes flat and stops saying anything, while an extrapolated one keeps leading and keeps crossing. Requiring both is asking for a move that is directional and clean at once
Meilleur régime: Trending

No single-flow template pairs them. Add two DEMA nodes into a Cross on top of FrAMA vs Fixed EMA and join the two conditions with an And gate.

Ouvrir Builder →

Paramètres

Valeurs de départ à valider sur votre propre paire et timeframe — pas des réglages garantis.

Paramètre	Défaut	Plage de test suggérée	Ce que ça fait
Period	14	1–999	How far the measurement looks — and the number to be careful with, because it is half the story. The three ranges the dimension is built from span 2 × period bars in total, so a period of 14 reads 28. MT5 draws nothing until that many bars exist. This does not make the line smoother or faster in the way an EMA length does; it changes the stretch of chart the jaggedness is measured over.
Applied price	Close	Close / Open / High / Low / Median / Typical / Weighted	Which series gets smoothed. It does not affect the adaptation at all: the fractal dimension is measured from the bars' highs and lows regardless of what is set here. That makes it a far smaller lever than on an ordinary moving average, where applied price and length are the only two controls there are.
Shift	0	0–100	Which bar to read. The template sets it to 1 so a cross is decided on a closed bar. This is not the same as iFrAMA's own ma_shift, which slides the whole line sideways on the chart — the Builder pins that to 0 and exposes only the bar offset, so there is one shift here rather than two.
Timeframe	Current	Current / M1–MN1	Which timeframe's bars the measurement runs on. Set it above the chart and the line becomes a slower regime backdrop for entries timed on the chart's own bars — the usual way to use an adaptive average as a filter rather than a trigger.

Préréglages de départ

Default 14 (28-bar window) MT5's own default and the template's setting

Slower 30 (60-bar window) Adapts less often; the dimension is measured over a longer stretch

Faster 8 (16-bar window) Reacts sooner and reaches the 1.0 weight more readily — expect the line to sit on price more of the time

Exemples de marché

Où FrAMA fonctionne, où il échoue et comment un filtre change le résultat.

Fonctionne

The line woke up

A quiet stretch keeps the average flat while price crosses it repeatedly; then a genuine move starts, the weight climbs toward 1 and the line lifts off the fixed average it had been sitting on. The separation, not the price move, is what the template trades.

Échoue

A cross with nothing behind it

The two averages touch and part again inside a range where neither is really sloping. The dimension never leaves the top of its scale, so the adaptive line never actually accelerated — the crossing came from the fixed side drifting.

Filtré

The same range, with a strength test attached

The identical stretch with a trend-strength condition joined to the cross. The lines still touch; no order is sent, because the flow was asked for a number the flat line cannot supply on its own.

Indicateurs associés

AMA Voir le guide →

VIDyA Voir le guide →

Moving Average Voir le guide →

ADX Voir le guide →

FAQ

Is FrAMA built into MetaTrader 5?: Yes. iFrAMA is a standard MQL5 function taking a symbol, a timeframe, the period, an ma_shift and an applied price, and the terminal also ships a FrAMA custom indicator under Indicators\Examples. The Builder calls the built-in function, so a generated EA carries no dependency on any file being present in the terminal that runs it.
What does the period actually control?: How far back the measurement looks — and it is half the window. The dimension needs the recent `period` bars, the `period` before those, and the combined stretch, so a period of 14 reads 28 bars and nothing is drawn until 28 bars exist. It is not a smoothing length. Doubling it does not halve the responsiveness the way it would on an EMA; it changes the stretch of chart the jaggedness is measured over.
How is it different from AMA?: Both replace a fixed smoothing constant with a measured one, and they measure different things. Kaufman's AMA uses an efficiency ratio — how much net distance the price covered against how much total distance it travelled. FrAMA uses the fractal dimension of the curve, which is about how jagged the path is rather than how far it got. A market can travel a long way by a very rough route, and that is exactly where the two disagree.
Why does changing applied price barely do anything?: Because it is not in the part that adapts. The fractal dimension is computed from the bars' highs and lows regardless of the setting; applied price only selects the series that the resulting weight is applied to. On an ordinary moving average applied price is one of only two controls, so the habit of treating it as significant carries over — here it is close to cosmetic.
Can I use the smoothing factor as a signal?: Not from this node. iFrAMA publishes one buffer and it is the average itself; the dimension and the weight are computed inside and discarded. The chart on this page draws the factor because it is what the indicator is about, but a Builder flow cannot read it. If a flow needs an explicit statement that the market is quiet, that has to come from a different indicator.
Can I build a FrAMA EA without coding?: Yes. The FrAMA vs Fixed EMA template puts a FrAMA node at 14 next to a Moving Average node set to EMA 28 — the same window FrAMA reads — and trades the crossings between them, closing the opposite side before it enters. The period, the fixed length and method, the shift, the stop distances, the lot and the spread limit are all EA inputs.
How should I validate settings before going live?: Backtest on quality tick data, then demo, and check two things that are specific to this indicator. First, that the test period is long enough to contain both regimes — an adaptive average tested only in a trend is being judged on the half of its job it does not need to do. Second, that the instrument does not gap much: the weighting is only well behaved while the two half-windows overlap in price, so an instrument with frequent gaps is testing the arithmetic as much as the strategy. And because the period is a window rather than a smoothing length, resist scanning it — the number that wins a scan is describing the sample, not the market.

Glossary

Termes clés

Backtest Surajustement Drawdown maximal