Is TEMA built into MetaTrader 5?

Yes. iTEMA 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 TEMA custom indicator under Indicators\Examples. The Builder calls the built-in, so a generated EA needs no extra files on the terminal that runs it.

Is TEMA just three EMAs averaged together?

No. It uses three chained EMAs but combines them as 3 × EMA − 3 × EMA(EMA) + EMA(EMA(EMA)) — two of the three terms are subtractions. Averaging them would be slower than any of them; the alternating signs cancel the lag twice over, which is what makes the line arrive early. The coefficients sum to one, so the result still lives on the price scale.

How is it different from DEMA?

One correction stage. Same period, same applied price, same smoothing constant; the difference between the two lines is EMA − 2 × EMA(EMA) + EMA(EMA(EMA)), a second difference of the same chain. In numbers: on a step from 100 to 110, TEMA reaches 95% of the new level in 4 bars against DEMA's 6, and on a random walk it sits outside the recent 14-bar range on 7.5% of bars against DEMA's 2.1%. Faster, and wrong about levels more often.

How many bars does it need before it draws?

3 × period − 3. Each EMA in the chain waits for the previous one, so a period of 14 needs 39 bars and a period of 50 needs 147. MT5 returns without calculating while the history is shorter, so the left of the chart is empty rather than wrong.

Why does the template use TEMA to close rather than to open?

Because the measured overshoot makes it an expensive trigger and a cheap exit. It sits outside the recent range on 7.5% of bars, so as an entry it mostly buys round trips — TEMA 14/28 crosses 291 times on a random walk where a plain EMA 14/28 crosses 114. On the exit the same eagerness shortens the average hold from 32.5 bars to 20.5 and raises the realised share of the peak open profit from 16.6% to 22.0% (512 trades against the opposite-cross exit's 593, over five random-walk series), and a wrong exit costs a re-entry rather than a trade.

Can I build a TEMA EA without coding?

Yes. The EMA Cross Entry x TEMA Early Exit template enters on a plain EMA 14/28 crossover and closes when TEMA breaks the fast EMA, with the close nodes ahead of the entries so a netting account reverses correctly. The three periods, the applied price, 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 read the trade list rather than only the curve. Two checks are specific to this indicator. Make sure the test contains ranges as well as trends, because the overshoot only shows its bill in the former. And before lengthening the period to calm the line down, check whether the noise you are seeing is on the entry side — if it is, the entry rule is what needs the filter, and slowing TEMA down removes the one property you selected it for.

TEMA for MT5 traders

What TEMA tells you

DEMA takes an EMA, measures how far behind it is running by taking an EMA of that EMA, and adds the difference back. TEMA does the same thing with a third stage, and the coefficients — three, minus three, plus one — are what falls out of cancelling the lag twice rather than once. The result is the fastest line in this family and the one furthest from being an average: on a step change it arrives in a third of the time a plain EMA takes, and it regularly prints values that price never traded. Whether that is useful depends entirely on which side of a trade you put it on, which is the one thing most write-ups of it leave out.

It reaches a new level in 4 bars where DEMA needs 6 and a plain EMA needs 20
It leaves the last 14 bars' range on 7.5% of bars — DEMA 2.1%, a plain EMA 0.0%
The extra stage buys speed and pays in overshoot; nothing about it is smoother

1 An EMA of price, an EMA of that, and an EMA of that again
2 Three chained lags, cancelled twice rather than once
3 TEMA = 3 × EMA − 3 × EMA(EMA) + EMA(EMA(EMA))

Show the formula & calculation detail

TEMA is one line of arithmetic, and every coefficient in it is forced:

TEMA = 3 × EMA(price) − 3 × EMA(EMA(price)) + EMA(EMA(EMA(price)))

DEMA cancels a single EMA’s lag once. Cancel it twice and this is what falls out — three terms with alternating signs whose coefficients sum to one, so the line still lives on the price scale rather than drifting off it. Subtract DEMA from TEMA of the same period and what remains is EMA − 2 × EMA(EMA) + EMA(EMA(EMA)): a second difference of the same chain, checked to within 5.7e-14 over 3,958 bars. The extra stage is not more smoothing. It is curvature.

The speed is real and so is the bill. On a step from 100 to 110 with period 14:

	95% of the new level	Peak
Plain EMA	20 bars	110.00
DEMA	6 bars	111.17
TEMA	4 bars	111.66

The peak matters as much as the timing: 111.66 is a price that never traded anywhere in the series. A real moving average cannot do that, because it is a weighted sum of its window with weights that add to one. On a random walk the line sits outside the last fourteen bars’ high-low range on 7.5% of bars — against 2.1% for DEMA and 0.0% for the plain EMA the same test was run on.

Which is why the template does not trade it. The obvious construction — a fast TEMA crossing a slow one — is the loudest thing in this family. Measured on the same random walk, TEMA 14/28 crosses 291 times, DEMA 14/28 crosses 227, and a plain EMA 14/28 crosses 114. Each of those crossings is a spread paid to find out whether the lead meant anything.

So the entries are left to the plain EMA pair and TEMA is put on the exit, where being early costs a re-entry instead of a trade. Against the same entries held until the opposite cross:

Average hold: 32.5 bars → 20.5
Share of the peak open profit realised: 16.6% → 22.0%
Trades closed within two bars of opening: 8.1% → 1.2%

Measured across five random-walk series — 512 trades under the TEMA exit against 593 under the opposite-cross exit. A random walk says nothing about whether the strategy makes money; what it can compare is how much of a move is still there when the position closes.

The period is a third of the history. Each EMA in the chain waits for the one before it, so MT5 sets PLOT_DRAW_BEGIN to 3 × period − 3 and calculates nothing before then. A TEMA(14) needs 39 bars; a TEMA(50) needs 147. Nothing errors — the left of the chart is simply empty.

Signals

The distinct, backtestable signals TEMA gives — and the regime each one suits.

TEMA Breaks the Fast EMA

Builder ready

Condition: TEMA crosses the plain EMA of the same period against the position
Best regime: Either regime
Typical use: The template's rule, and it is an exit rather than an entry — the lead is most valuable where a false signal only costs a re-entry

Price Against the Line

Builder ready

Condition: Price closes above or below TEMA
Best regime: Trending
Typical use: A direction filter that turns sooner than DEMA of the same length, and flips more often for the same reason

Fast / Slow TEMA Cross

Builder ready

Condition: A short-period TEMA crosses a longer one
Best regime: Trending
Typical use: The obvious construction, and the noisiest of the family — measured on a random walk it crosses 291 times against 227 for the same pair of DEMAs

Overshoot Past the Range

Advanced logic

Condition: The line sits outside the recent bars' high and low
Best regime: Trending
Typical use: A stretch reading in principle; comparing the line against a rolling range is not a single node

MT5 implementation

What MetaTrader 5 itself computes and draws — the reference every rule on this page is written against.

Buffers

Index	Buffer	Drawn in MT5 as	What it holds
0	MAIN_LINE	TEMA	The line, and the only series MT5 plots. The indicator declares four buffers; the other three hold the EMA, the EMA of that EMA and the EMA of that, all marked INDICATOR_CALCULATIONS — working storage rather than output. The Builder reads index 0.

Platform notes

MQL5 function

iTEMA

Applied price

In the Builder

The node exposes it and the generated EA passes it straight to iTEMA. As with DEMA and unlike FrAMA, it does what it says: all three chained EMAs run on the series you choose, so switching from Close to Typical changes every number. The Builder pins iTEMA's own ma_shift to 0 and uses the node's shift as the bar to read, which is why there is one shift on the node rather than two.

Implementation notes

The generated EA creates one iTEMA handle per node and reads buffer 0. MT5's own implementation declares four buffers — the line and the three chained EMAs — but only the first is INDICATOR_DATA; the others are INDICATOR_CALCULATIONS, working storage that happens to be reachable through the buffer API. Nothing in the Builder reads them.
The formula is one line: TEMA = 3 × EMA − 3 × EMA(EMA) + EMA(EMA(EMA)). The coefficients are not a weighting anybody chose — they are what remains after cancelling a single EMA's lag twice, and their sum is one, which is why the line still tracks price rather than drifting away from it.
Subtracting DEMA from TEMA of the same period leaves EMA − 2 × EMA(EMA) + EMA(EMA(EMA)) — a second difference of the same chain, checked to 5.7e-14 over 3,958 bars. That is the extra stage in isolation: not more smoothing, but the curvature of the cascade, which is why it leads more and why it overshoots more.
Warmup is 3 × period − 3, not period. MT5 starts each stage where the previous one produced its first value — the bundled TEMA.mq5 passes begin values of 0, period − 1 and 2 × period − 2 to ExponentialMAOnBuffer — and sets PLOT_DRAW_BEGIN to match. A period of 14 needs 39 bars, a period of 50 needs 147.
The speed and the cost were both measured rather than assumed. On a step from 100 to 110 with period 14, reaching 95% of the new level takes a plain EMA 20 bars, DEMA 6, and TEMA 4; the TEMA peak on that step is 111.66, against DEMA's 111.17 and a price high of 110. On a random walk the line sits outside the last fourteen bars' high-low range on 7.5% of bars, against 2.1% for DEMA and 0.0% for the plain EMA the test was also run on.
The template puts TEMA on the exit for that reason, and the choice was measured too. Against the same EMA 14/28 entries held until the opposite cross, closing when TEMA breaks the fast EMA shortens the average hold from 32.5 bars to 20.5 and raises the share of the peak open profit actually realised from 16.6% to 22.0%. Measured across five random-walk series: 512 trades under the TEMA exit and 593 under the opposite-cross exit, the counts differing because a shorter hold leaves room for a different number of entries. Random walks say nothing about profit; what they can compare is how much of a move is still there when the position closes.
The close nodes sit ahead of the entry nodes in the generated code, and what puts them there is the order of the connections rather than the order of the node list. The generator topologically sorts the graph, so a node reaches the output when its inputs have — meaning the branch whose inputs clear first is emitted first. In the other order a netting account nets a new order against the position it was meant to replace; a hedging account behaves correctly either way, which is what makes it easy to miss.

Works best / use carefully

No indicator is a universal edge. Here is where TEMA helps — and where it misleads.

Works best

On the exit side, where arriving early costs a re-entry rather than a losing trade
Trends steep enough that the overshoot is behind price rather than ahead of it
As a faster read of a higher timeframe, where its flips are rarer in chart-bar terms
Anywhere a DEMA has been shortened to get more lead — this is the principled version of that shortening

Use carefully

As an entry trigger: it pokes outside the recent range on 7.5% of bars, and most of those are not moves
Ranges, where two corrections keep supplying lead to a move that is not there
Short histories: nothing is drawn until three times the period minus three bars exist
Reading the line as a level — it sits at prices that never traded more often than DEMA does

Build a strategy with TEMA

Wire the signal into entry and exit rules, then export a compilable MT5 EA — no code.

One TEMA node on Close at 14, and two plain EMA nodes at 14 and 28
TEMA falls through the fast EMA → Close the long. Rises through it → Close the short
EMA 14 crosses above EMA 28 → Open Buy · SL 70 / TP 210. Below → Open Sell
All three read the last closed bar, so every cross is a settled fact
The close nodes come before the entries, and that ordering is deliberate

Combine TEMA with other indicators

One indicator rarely stands alone. These pairings cover TEMA's blind spots.

TEMA + DEMA

DEMA at 14
TEMA at 14
Compare

Why: The comparison that defines this indicator. Same period, same applied price, same smoothing constant — the only difference is that one cancels the lag once and the other twice, so every gap on screen is the third stage doing its work. It also settles the question of whether the extra stage is worth it on your instrument: on a step change TEMA arrives two bars sooner, and on a random walk it sits outside the recent range on 7.5% of bars against 2.1%. Those two numbers are the whole trade-off
Best regime: Either regime

No single-flow template pairs them. Start from EMA Cross Entry x TEMA Early Exit, add a DEMA node at 14 on Close, and put it into a Compare against the TEMA node.

Open Builder →

TEMA + ADX

ADX above 25
EMA 14 crosses above EMA 28
Buy

Why: The failure mode is specific and the answer is not more smoothing. Two correction stages keep supplying lead in a range, so the line crosses things that are not moves. A trend-strength reading does not improve the line — it decides whether the lead is worth acting on, and it belongs on the entry rather than the exit, because an early exit in a range is the behaviour you wanted anyway
Best regime: Trending

No single-flow template combines them. Add an ADX node with a Compare at 25 to EMA Cross Entry x TEMA Early Exit and join it to the entry Cross with an And gate — leave the exit path untouched.

Open Builder →

TEMA + ATR

ATR at 14
Stop loss in ATR mode
TEMA crosses the fast EMA

Why: The template exits on a line crossing and stops out at a fixed 70 pips, which are two different ideas about how much room to give a trade. Sizing the stop from measured volatility instead puts both on the same footing: the line handles the case where the move is over, the stop handles the case where it was never there. It also stops the fixed distance from being wrong in two directions at once across instruments
Best regime: Either regime

The Stop Loss node has an ATR mode, so this needs no extra condition — swap the fixed distance for the ATR setting and the exit path is unchanged.

Open Builder →

Parameters

Starting values to validate on your own pair and timeframe — not guaranteed settings.

Parameter	Default	Suggested test range	What it does
Period	14	1–999	The EMA length used three times — on price, on the result, and on that result. Longer is slower as usual, but the history it consumes is three times what you set: MT5 draws nothing until 3 × period − 3 bars exist, because the third EMA cannot start until the second has. A TEMA(50) needs 147 bars.
Applied price	Close	Close / Open / High / Low / Median / Typical / Weighted	Which series the whole calculation runs on. As with DEMA this is a real control rather than a formality — all three stages and therefore both corrections are computed from whatever you choose.
Shift	0	0–100	Which bar to read. The template sets 1 so an exit is decided on a closed bar. This is not iTEMA's own ma_shift, which slides the drawn line sideways — the Builder pins that to 0 and exposes only the bar offset.
Timeframe	Current	Current / M1–MN1	Which timeframe's bars the line is built from. Set above the chart and its flips become rarer in chart-bar terms, which is one of the few ways to keep the lead without paying the full overshoot bill.

Starting presets

Template 14 One TEMA on the exit, with a plain EMA 14/28 pair supplying the entries

Slower 21 Fewer exits and later ones; the lead survives, the flip count drops

Pair 14 / 28 Two TEMAs as a crossover — the loudest configuration, and worth measuring before trusting

Market examples

Where TEMA works, where it fails, and how a filter changes the outcome.

Works

Out near the top

A move stalls and the line breaks the fast average several bars before the entry pair would have crossed back. The position closes with most of the excursion still in it — the behaviour the 16.6% to 22.0% measurement is describing.

Fails

The exit that was early

A pullback inside a live trend breaks the line, the position closes, and the trend continues. This is the cost of the design, and it is deliberately the cheap side of the trade: the entry rule is still valid and can fire again.

Filtered

Same stretch, strength required

The identical range with a trend-strength condition on the entry. The exit path is untouched, because closing early in a range was never the problem.

Related indicators

DEMA View guide →

Moving Average View guide →

FrAMA View guide →

TRIX View guide →

FAQ

Is TEMA built into MetaTrader 5?: Yes. iTEMA 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 TEMA custom indicator under Indicators\Examples. The Builder calls the built-in, so a generated EA needs no extra files on the terminal that runs it.
Is TEMA just three EMAs averaged together?: No. It uses three chained EMAs but combines them as 3 × EMA − 3 × EMA(EMA) + EMA(EMA(EMA)) — two of the three terms are subtractions. Averaging them would be slower than any of them; the alternating signs cancel the lag twice over, which is what makes the line arrive early. The coefficients sum to one, so the result still lives on the price scale.
How is it different from DEMA?: One correction stage. Same period, same applied price, same smoothing constant; the difference between the two lines is EMA − 2 × EMA(EMA) + EMA(EMA(EMA)), a second difference of the same chain. In numbers: on a step from 100 to 110, TEMA reaches 95% of the new level in 4 bars against DEMA's 6, and on a random walk it sits outside the recent 14-bar range on 7.5% of bars against DEMA's 2.1%. Faster, and wrong about levels more often.
How many bars does it need before it draws?: 3 × period − 3. Each EMA in the chain waits for the previous one, so a period of 14 needs 39 bars and a period of 50 needs 147. MT5 returns without calculating while the history is shorter, so the left of the chart is empty rather than wrong.
Why does the template use TEMA to close rather than to open?: Because the measured overshoot makes it an expensive trigger and a cheap exit. It sits outside the recent range on 7.5% of bars, so as an entry it mostly buys round trips — TEMA 14/28 crosses 291 times on a random walk where a plain EMA 14/28 crosses 114. On the exit the same eagerness shortens the average hold from 32.5 bars to 20.5 and raises the realised share of the peak open profit from 16.6% to 22.0% (512 trades against the opposite-cross exit's 593, over five random-walk series), and a wrong exit costs a re-entry rather than a trade.
Can I build a TEMA EA without coding?: Yes. The EMA Cross Entry x TEMA Early Exit template enters on a plain EMA 14/28 crossover and closes when TEMA breaks the fast EMA, with the close nodes ahead of the entries so a netting account reverses correctly. The three periods, the applied price, 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 read the trade list rather than only the curve. Two checks are specific to this indicator. Make sure the test contains ranges as well as trends, because the overshoot only shows its bill in the former. And before lengthening the period to calm the line down, check whether the noise you are seeing is on the entry side — if it is, the entry rule is what needs the filter, and slowing TEMA down removes the one property you selected it for.

Glossary

Key terms

Exponential Moving Average Backtest Overfitting