DEMA vs TEMA: which fast average lags less?

Your moving average finally rolled over. By the time it did, the move was most of the way done, and the candle that gave you permission to act was already several bars behind you.

DEMA vs TEMA is a choice between two fixes for exactly that lag. Each one smooths price, measures how far the smoothing fell behind, then hands that gap back so the line sits nearer to where price is. TEMA hands back more of it than DEMA does, which makes it the faster of the two and the twitchier.

Here is what the correction does to your chart, which of the two suits a trend versus a range, and the check that tells you whether either one earns its place.

What makes DEMA and TEMA faster than an EMA?

A correction, not more smoothing. The indicator smooths price once, smooths that smoothed line a second time, and treats the distance between the two passes as a rough reading of how far the first pass fell behind. Adding that distance back drags the result forward toward price.

Patrick Mulloy published both designs in Technical Analysis of Stocks and Commodities in 1994, and his stated goal was reducing lag rather than removing noise, as the summaries of his double exponential moving average work still describe. The name is the part that misleads people. “Double” and “triple” describe how many smoothing passes go into the arithmetic, not how heavily the output is smoothed. The output is less smoothed than an ordinary exponential average, not more. If you are still choosing between that ordinary EMA and a simple average, start with how an EMA and an SMA weigh candles.

Read the green line against the red one at the turn: it commits to the reversal roughly a bar sooner, because the lag it measured has been handed back.

DEMA vs TEMA: what separates the two?

The dose. DEMA applies one lag correction. The triple exponential moving average stacks a second, larger one on top of it, so at an identical setting the TEMA line sits closer to price, turns first at a reversal, and gives back more of the smoothing you originally wanted.

Both sit on the same line of reasoning as the other lag-reduction designs. If you have already compared them with a zero lag EMA or a Hull moving average, you will recognise the trade being made: every one of them buys earlier turns with the noise filtering it removes.

Nothing on that scale is free. Sliding right buys timing and sells patience, and the instrument decides which of the two you can afford.

Which one belongs on your chart?

Match the correction to the market, not to a preference. Strong directional instruments and higher timeframes reward TEMA, because most bars carry real information and an early turn is usually a true one. Ranges, thin sessions and low timeframes punish it, and that is where DEMA’s smaller correction earns its keep. If even DEMA flips too often for your instrument, the T3 average and its volume factor let you set the correction somewhere between a plain EMA and a DEMA.

Answer the question at the top honestly and the rest of the chart setup follows. Answer it by hope and you will spend the week fighting your own line.

Do DEMA and TEMA repaint?

No. Every value on both lines is calculated from prices at or before the bar it is drawn on, so once a candle closes its reading is final and no amount of reloading will move it. Overshoot is not repainting, and a line that swings past price and comes back has not rewritten anything.

The live bar is a different matter, and it catches people out. The current bar’s value tracks price tick by tick, so a line that appears to have turned can straighten out before the close. That is true of every moving average ever drawn, and it shows up more on a fast one because the correction magnifies whatever the last tick did. For the wider version of this problem, the non-repaint indicator explainer covers what does and does not count.

How do you test one before you trade it?

Run the comparison on your own instrument, because nobody else’s chart settles this. Three steps, and they take an evening:

1. Plot the exponential average, the DEMA and the TEMA at one shared period on the pair and timeframe you trade.
2. Scroll back through a trending stretch and a dead range, and mark every point where each line changed direction.
3. Count the turns that led somewhere against the turns that reversed within a couple of bars, per line, per regime.

You are not looking for a winner. You are looking for the period at which your chosen line stops reacting to bars you would have ignored by eye, and that number will differ from the one a video told you to use. Log it. Then leave it alone for a month.

Where does a faster average stop being enough?

At the point where you need a decision rather than a slope. A corrected average tells you the trend has probably bent. It does not tell you where the level was, what invalidates the idea, or where the stop belongs, and a faster line does not make any of those questions easier.

That is the gap RelicusRoad Pro is built to close. Its entries come from price structure and settle when the bar closes, so what you tested is what you trade, and it is meant to sit alongside a trend filter like a DEMA rather than replace your judgement about risk. Nobody is claiming an indicator fixes position sizing; it sharpens timing, and timing is the part a lag correction was already trying to buy you.

Pick one of the two lines, learn its false turns on your own instrument, and let something with structure behind it handle the entry.

Frequently asked questions

What is the difference between DEMA and TEMA? The size of the lag correction. Both start by smoothing price with an exponential average, then smoothing that result again, and the gap between the two passes is a rough measure of how far the smoothing fell behind. DEMA adds that gap back once. TEMA carries the idea a step further and adds a second, larger correction on top of it. At the same period setting TEMA therefore sits nearer to price, turns sooner when price turns, and wobbles more when price goes nowhere.

Is TEMA better than DEMA? Not better, just faster, and speed is only useful if the market is paying for it. TEMA earns its extra correction on strongly trending instruments and on higher timeframes where each bar carries real information. On a quiet range or a low timeframe full of noise, that same correction turns the line early and often, and every early turn is a trade you have to sit through or pay to exit. If you cannot tell which regime you are in, DEMA is the more forgiving default.

What period should I use for DEMA or TEMA? Start from whatever period you already trust on a plain exponential average and change nothing else. Because DEMA and TEMA remove lag rather than add smoothing, a 21 setting on either behaves more like a much shorter conventional average, which is why traders who drop them in at their old settings usually complain about noise. Compare the three lines on the same chart first, then lengthen the period until the fast line stops reacting to bars you would have ignored by eye.

Do DEMA and TEMA repaint? No, in the sense that matters. Every value comes from prices at or before the bar being drawn, so once a candle closes its reading is fixed and reloading the chart will not move it. What does change is the live bar. Its value tracks the current price tick by tick, so a line that looks like it has turned can flatten again before the candle ends. Judge the turn at the close, not while it is forming.

Can DEMA or TEMA replace an EMA crossover? You can build a crossover from either, and the signals will arrive earlier than the exponential version. You will also get more of them. A faster line crosses more often, which means more entries, more spread paid, and more of your risk budget spent on trades that reverse. Test the same rules on all three average types over the same period before deciding the earlier fill was worth the extra count.

Want entries that settle at the close instead of a line you have to second-guess? See how RelicusRoad Pro builds its signals.

Written for RelicusRoad by RelicusDigital.com.