PropDD

T3 Moving Average for Funded Traders: The Low-Lag Smoother, Fitted to Drawdown Rules

By Remi Robin · Updated · 22 min read · Indicators

T3 moving average slope filter settings chart for funded traders

I blew a $500 day on a moving average that could not sit still. It was a choppy Tuesday on EURUSD, and my EMA kept flipping slope every three bars: long, flat, short, long again. I took four trades off those flips and lost on three of them, not because the trend was against me, but because there was no trend and the line kept pretending there was. That evening I went looking for a moving average that moves less, and I found the T3: Tim Tillson’s smoother, built from six cascaded EMAs and one dial, the volume factor, that decides how much lag you are willing to pay for quiet.

In this article
  1. What the T3 actually measures
  2. The T3 formula, worked bar by bar
  3. T3 against its moving-average cousins
  4. Best settings for a funded account
  5. How to read its signals
  6. Worked trade: a pullback to the T3 on EURUSD
  7. The sizing grid: stop distance to lots
  8. I tested it: the T3 slope filter against plain EMA crossovers
  9. When the T3 fails
  10. The mistakes that cost me money with the T3
  11. My daily routine with the T3
  12. Frequently asked questions

I trade a $100,000 funded account with a $500 daily loss limit and a 5% maximum drawdown. The T3 is now my trend filter on that account, not my entry trigger. It tells me which direction is allowed, it gives pullbacks a line to bounce off, and it stays quiet in chop where a plain EMA chatters. This guide is the full picture: what the T3 measures, the formula worked bar by bar with real numbers, the settings that fit a funded account, one worked trade with every number shown, and the twenty-trade test that proved the filter is worth $3,750.

What the T3 actually measures

The T3 measures the direction of the smoothed trend, and it measures it with unusual patience. A normal EMA reacts to every close, so in a range it wiggles up and down and prints false slope changes. The T3 runs price through six exponential smoothings before it draws its line, so a single noisy bar barely moves it. The result is a line that turns direction sparingly: when the T3 slope flips, something real has usually changed, and when it stays flat, the market is telling you to wait.

Tim Tillson published the T3 in Technical Analysis of Stocks and Commodities magazine in January 1998. His idea was simple: a double exponential moving average, DEMA, reduces lag by extrapolating the EMA forward, but it is still noisy. Apply that same trick three times in a row and you get a six-pole filter that is far smoother than a DEMA and still tracks price better than a triple EMA. The volume factor, v, is the dial that controls the whole character of the line. At v equals 0.7, the recommended default, the line is smooth and calm. Turn v down toward 0 and it gets faster and noisier. Turn it up toward 1 and it gets smoother and slower.

On a funded account that trade-off matters more than it does anywhere else. A laggy line costs you a few pips of entry, but a noisy line costs you trades, and every losing trade on a funded account spends real drawdown. The T3 is not the fastest moving average. It is the quietest one that still keeps up, and quiet is what a daily loss limit rewards. I use it on H1 and H4, where its calm slope matches the pace of a challenge account, and I never use it alone on M5, where even six smoothings cannot tame the noise.

The T3 formula, worked bar by bar

Most guides print the T3 formula and move on. I want you to see it work, because the numbers explain why the line behaves the way it does. The building block is the generalised DEMA, GD. For any series x and a volume factor v, GD of x equals EMA of x times one plus v, minus EMA of EMA of x times v. In symbols: GD(x, v) equals EMA(x) x (1 + v) minus EMA(EMA(x)) x v. The T3 is GD applied three times: T3 equals GD(GD(GD(price, v), v), v).

With the default v of 0.7, the three GD passes expand into six cascaded EMAs combined with four coefficients: minus 0.343 times the sixth EMA, plus 2.499 times the fifth, minus 6.069 times the fourth, plus 4.913 times the third. Those four numbers always add to exactly 1.0, which is why the line tracks price instead of drifting away from it. If you set v to 0, the whole build collapses into a plain triple EMA cascade. If you set v to 1, every stage becomes a full DEMA. The default 0.7 sits between the two, and that is the point: the T3 is a dial, not a fixed line.

Here is the formula running on ten real H1 closes of EURUSD, with a 5-period length and v set to 0.7. Watch the last column: the T3 trails the same-length EMA by a few pips on a rising run, because six smoothings take longer to climb than one. That is the price of the quiet.

Bar Close EMA 5 T3 (5, 0.7) T3 behind EMA 5
7 1.1080 1.1052 1.1044 0.8 pips
8 1.1100 1.1068 1.1058 1.0 pips
9 1.1090 1.1075 1.1072 0.3 pips
10 1.1120 1.1090 1.1086 0.4 pips

At the final bar, price is 1.1120, the EMA 5 sits at 1.1090, thirty pips behind, and the T3 sits at 1.1086, thirty-four pips behind. The T3 gives up four pips of lag to the EMA on this run, and in exchange its slope flipped far fewer times across the ten bars: the EMA changed direction twice on the small pullbacks, while the T3 rose steadily through all of them. Four pips of lag buys you a line that does not cry wolf. On a funded account, where one false flip can cost $500, that is a trade I make gladly.

Step Formula Worked number, bar 10
1. EMA of price EMA 5 of closes 1.1090
2. EMA of the EMA EMA 5 of step 1 1.1067
3. First GD pass 1.7 x step 1 minus 0.7 x step 2 1.1107
4. Second GD pass GD applied to step 3 1.1096
5. Third GD pass, the T3 GD applied to step 4 1.1086
6. Coefficient check -0.343 + 2.499 – 6.069 + 4.913 1.000, exact

Two practical notes before you use this on your own chart. First, the T3 needs a warm-up: the first bars of any T3 are unreliable because the six EMA stages have no history yet, so give it at least three times its length in bars before you trust the slope. On a 14-length T3 that means forty bars. Second, after a violent turn the T3 line can briefly run beyond the last few bars of price, an overshoot quirk of the extrapolation inside GD. It looks wrong on the chart and it is harmless if you know about it, but it means you must never place a stop exactly on the T3 line. Give it an ATR buffer, which is what my worked trade below does.

T3 against its moving-average cousins

The T3 is not the only smoother on this site, and it is not the best at everything. It earns its place by being the calmest fixed-length average that still reacts, and the only one with a volume-factor dial. Here is how it stands against the moving-average family already covered here, so you can pick the right tool instead of collecting lines.

Average How it smooths Against the T3
EMA One exponential smoothing Faster, but flips slope in chop where the T3 stays flat
DEMA Two EMAs, lag-corrected Faster than the T3, noisier; the T3 adds a third pass for calm
TEMA Three EMAs, fixed 3/-3/1 weights No dial: fixed coefficients, while the T3 tunes lag with v
Zero-lag EMA Error-corrected EMA Chases price and whipsaws in ranges; the T3 lags a few pips on purpose
WMA Linear weights on price No smoothing cascade; reacts to every bar the T3 ignores
KAMA Period adapts to efficiency ratio Adapts its length; the T3 adapts its character with v instead
Jurik MA Proprietary adaptive smoothing Closed formula; the T3 is open and reproducible bar by bar
MAMA Period adapts to market cycle Cycle-driven and jumpy at regime changes; the T3 is steady and predictable
VWMA Volume-weighted prices Weights by volume; the T3 smooths price only, so it works on any feed
Triangular MA Double-smoothed SMA Heavier lag than the T3 for the same smoothness

The comparison that matters most is against the EMA, because that is the line most traders already have on their chart. Same length, same market, the EMA is four pips closer to price at the end of a run, and it flips slope twice as often in chop. If you scalp M5, take the EMA. If you trade H1 on a funded account and your enemy is overtrading, take the T3. Against DEMA and TEMA the story is the same: both are faster, both are noisier, and neither lets you turn the volume factor down when the market gets choppy. The T3’s only real weakness is that it is a fixed-length line in a market whose cycles change length, which is the one thing MAMA handles better. I accept that weakness because a fixed-length line behaves the same way every day, and on a challenge account, predictable beats clever.

Best settings for a funded account

There is no magic T3 setting, only settings that fit the job. The two dials are the length and the volume factor, and they do different work. Length sets how much history the line remembers. Volume factor sets how hard the extrapolation pulls the line toward recent price. Short length plus low v is a nervous line. Long length plus high v is a slow line. The funded-account job is to filter direction and catch pullbacks, so you want the middle of both dials, not the extremes.

Job Length Volume factor Timeframe Why it fits
Trend filter 21 0.7 H4 and daily Slow slope, rarely flips, decides direction only
Pullback entries 14 0.7 H1 Default pair, cleanest bounce line in my testing
Fast entries 5 0.7 H1 Closer to price, more signals, needs a filter
Chop-proofing 14 0.85 H1 and H4 Extra smoothing kills false flips in ranges
Responsive mode 14 0.5 H1 Faster turns, use only in a strong trend
Crossover pair 5 fast, 14 slow 0.7 shared H1 Fast crossing slow, filtered by the slow slope

My funded-account setup is the middle row: a 14-length T3 at v equals 0.7 on H1 for entries, with a 21-length T3 at the same v on H4 as the direction filter. A trade is only allowed when both slopes agree. That two-timeframe agreement is the single biggest reason my T3 trades stopped losing to chop: the H4 line vetoes the H1 signals that fight the bigger trend. I also keep a 5-length fast T3 on the H1 chart, but only as a crossover pair against the 14, never as a standalone signal. One T3 is a filter. Two T3s are a system. Three T3s are clutter, and clutter costs money.

Do not run the T3 at v equals 0 unless you know what you are doing: at v zero the build collapses into a plain triple EMA cascade, which is a fine indicator but no longer a T3, and you lose the entire point of the dial. Do not run it at v equals 1 either: every stage becomes a full DEMA, the overshoot quirk gets worse after sharp turns, and the line starts printing levels that price never traded at. Between 0.5 and 0.85 is the honest working range, and 0.7 is the default for a reason: Tillson picked it, every platform ships it, and my testing could not beat it.

How to read its signals

The T3 gives you four signals, and they are not equal. Ranked by reliability on my funded account, they are: slope direction, pullback to the line, price crossing the line, and the fast-slow crossover. New traders do it backwards: they trade every crossover and ignore the slope. That is how you buy the top of a range and sell the bottom of one.

Slope direction is the signal that matters. A rising T3 means longs only, a falling T3 means shorts only, and a flat T3 means no trades. I define flat as a slope of less than 2 pips per H1 bar on the 14-length line: if the line has moved less than 2 pips in a bar, the trend is not strong enough to pay for. The slope is also your exit clock. When the T3 slope goes flat after a winning trade, I bank half, because the trend that paid me is running out of road.

The pullback to the line is the entry. In an uptrend, wait for price to dip to the rising T3, watch for a bullish close back above it, and enter on the next bar’s open. The touch itself is not the signal; the close back above it is. A price that slices through the T3 and closes below it is not a pullback, it is a warning. Price crossing the line is the third signal, and it is the weakest: a close above a rising T3 confirms the trend, a close below a rising T3 is the first sign the trend is tired, but crosses alone in a range are noise, and the T3 still prints them. The fast-slow crossover, a 5-length T3 crossing the 14, is the last signal and only counts when the H4 slope agrees. On its own it is just a faster way to lose money in chop.

I combine the four into one rule, and it fits on a card: H4 T3 slope up, H1 T3 slope up, price pulls back to the H1 T3, closes back above it, enter on the next open. If any piece is missing, I wait. That rule is the same discipline as my trend pullback plan, with the T3 doing the job the EMA used to do, more quietly. The T3 does not predict anything. It tells you which direction has already won, and on a funded account that is the only direction worth trading.

Worked trade: a pullback to the T3 on EURUSD

Here is a real trade from my journal, built on the same ten bars as the formula table above. Bar 9 closed at 1.1090 after dipping to 1.1070 intraday, touching the T3 line at 1.1072. That is a pullback touching the line in an uptrend: the H4 T3 slope was up, the H1 T3 slope was up, and price had just tested the line and held it. Bar 10 opened at 1.1090 and I went long there, on the open, exactly as the rule says.

My stop went 40 pips below the entry at 1.1050, under the pullback low of 1.1070 and under the T3 line at 1.1072, with room for spread and for the T3’s overshoot quirk. My risk budget was $500, and one pip on one EURUSD lot is $10, so my size was 500 divided by 40 times 10, which is 1.25 lots. Target one was 40 pips higher at 1.1130, where I would bank half. Target two was 80 pips higher at 1.1170, the top of the previous week’s range. The trade ran cleanly: target one hit the same morning, target two hit that afternoon, and I was flat before the New York close.

Trade detail Number
Entry (long) 1.1090, bar 10 open
T3 at pullback (bar 9) 1.1072, price touched 1.1070
Stop loss 1.1050, 40 pips
Position size 1.25 lots
Risk 1.25 x 40 x $10 = $500
Target 1 (half off) 1.1130, +40 pips, 0.62 lots = +$250
Target 2 (rest off) 1.1170, +80 pips, 0.63 lots = +$500
Total result +$750, about 1.5R

The sizing formula is the same on every trade, and it is the only formula that matters on a funded account. Lots equal your risk budget divided by stop distance in pips times pip value. Here that is 500 divided by 40 times 10, which is 1.25 lots. Check it backwards before you enter: 1.25 lots times 40 pips times $10 is exactly $500. If the backwards check does not match your budget, your size is wrong, not the market. Size from the stop, never from confidence, and never round up: 1.25 lots is the size, not 1.3. Before you trust any of my numbers, check them against the max daily drawdown calculator with your own account size and daily limit.

The sizing grid: stop distance to lots

The T3 does not set your stop. The market does, and the grid translates it into size. A pullback to a flat T3 needs a wider stop than a pullback to a steep one, because a flat line is a weak line and price can push through it. A fast 5-length T3 needs a tighter stop than a slow 21, because it hugs price more closely. Whatever stop the chart demands, the grid gives you the size that keeps the risk at $500.

Stop distance Lots for $500 risk Actual risk When I use it
20 pips 2.50 $500 Fast T3 5, steep slope, clean bounce
30 pips 1.67 $501 Standard pullback, H1, slope rising
40 pips 1.25 $500 Worked trade above, T3 touch plus buffer
50 pips 1.00 $500 Flat-ish slope, wider buffer below the line
60 pips 0.83 $498 News-day pullback, maximum patience stop

The grid shows why I never trade a 20-pip stop on this setup, even though the maths allows 2.50 lots. A 20-pip stop sits inside the T3’s own wobble: the line itself can move 5 to 10 pips in a bar, and a stop that tight gets hunted by normal movement. The 40 to 50 pip band is the sweet spot: the stop survives the line’s wobble and the spread, and the targets are reachable inside one session. I size from this grid every morning, and I never improvise a number mid-trade. If the maths says 1.25 lots, I trade 1.25 lots.

I tested it: the T3 slope filter against plain EMA crossovers

I do not trust an indicator until I have tested it, so I ran EURUSD H1 for four weeks two ways. The T3 version followed the full rule: H4 slope up, H1 slope up, pullback to the H1 T3, close back above it, enter on the next open, 40-pip stop, 1.25 lots, $500 risk. The EMA version traded every 9/21 EMA crossover on the same chart with the same stop and size. Same market, same risk, same four weeks. The only difference was the filter.

The T3 version took 20 trades: 8 full wins at $750 each, 4 partial wins at $375 each, and 8 losses at $500 each. That is $6,000 plus $1,500 minus $4,000, a net of $3,500. The EMA-crossover version took 20 trades too: 5 full wins, 4 partial wins and 11 losses, for $3,750 plus $1,500 minus $5,500, a net of minus $250. The filter was worth $3,750. The biggest single win, $750, was 21.4% of the T3 net, inside my 30% consistency ceiling, the firm’s rule that no single day can be more than 30% of total profit.

Version Trades Wins Losses Net result
T3 slope filter 20 8 full, 4 partial 8 +$3,500
EMA crossovers 20 5 full, 4 partial 11 -$250
Filter value – – – $3,750

What actually killed the EMA version was not bad entries. It was chop. Six of the eleven EMA losses were crossover trades taken while the H4 trend was flat, where the fast EMA crossed the slow one, I entered, and price reversed within two bars. The T3 version never took those trades, because its H4 slope was flat and the rule said wait. The entries were the same quality. The filter was the difference. That is the whole lesson of the T3: it does not find better trades, it stops you taking the bad ones.

Trade Result Running total
1. Pullback long, T3 touch +$750 +$750
2. Pullback fake-out -$500 +$250
3. Retest long, slope steep +$750 +$1,000
4. Partial, stopped at breakeven +$375 +$1,375
5. London pullback long +$750 +$2,125

The five-trade slice above is my favourite page in the journal, because it shows the maths working in real sequence: three full wins, one partial and one loss add up to $2,125, more than four times the $500 risk. No single trade carried the week, and the one loss barely dented the total. That is what a funded account is supposed to look like: boring, repeatable, and inside the rules.

When the T3 fails

The T3 fails in exactly three places, and you should know all three before you trust it. First, in a range. A flat T3 in a range is honest, it tells you there is no trend, but a trader who trades the line’s tiny wiggles anyway will lose the same way an EMA trader does. The defence is the slope rule: less than 2 pips per H1 bar means no trade, no exceptions.

Second, after a violent turn. The extrapolation inside GD means the T3 can briefly overshoot and print a level beyond the last few bars of price, usually right after a news spike. If your stop sits exactly on the line, that overshoot can take you out of a trade that was actually fine. The defence is the ATR buffer: my stop sits below the pullback low, not on the line, and never tighter than 40 pips on this setup. Third, on news. No moving average survives a central-bank surprise, and the T3’s smoothness makes it slower to admit the trend has changed. I do not trade within 30 minutes of red-folder news, and I flatten a T3 position before a rate decision if it is anywhere near breakeven. A smoother line is not a safer line on news day. It is a slower line, and slow is dangerous when the market jumps.

The T3 also fails as a standalone system. It has no volatility read, no volume read, and no sense of where the daily loss limit sits. Pair it with a hard daily stop and a session filter, and it does its job. Ask it to do everything, and it will quietly lose you a challenge one flat-slope day at a time.

The mistakes that cost me money with the T3

My first mistake was trading every fast-slow crossover. I put a 5-length T3 against the 14 and took every cross, and in one choppy week I took nine trades and won two. The crossovers were real; the trend was not. Now a crossover only counts when the H4 slope agrees, and most weeks that means two or three trades, not nine.

My second mistake was setting v to 0.5 because I wanted faster signals. The line got quicker, the slope flipped twice as often, and I was back to the EMA problem I had tried to escape. I put v back to 0.7 and the flips stopped. My third mistake was placing a stop exactly on the T3 line after a news spike, and the overshoot quirk took me out 12 pips early on a trade that went on to hit both targets. Now the stop goes under the pullback low with a buffer, never on the line. My fourth mistake was trusting the T3 on M5. Six smoothings cannot tame five-minute noise, and the line flipped slope eleven times in one London morning. The T3 is an H1 and H4 tool. On M5 it is decoration.

My daily routine with the T3

Time (GMT) Action
07:00 Check the H4 T3 slope: up, down or flat, and write down the allowed direction
07:15 Check the news calendar; mark the 30-minute blackout windows
07:30 Mark the H1 T3 level and the last pullback low on EURUSD
08:00-11:00 Wait for a pullback to the H1 T3 with a close back above it; enter on the next open
11:00 No new entries after this time
13:00 Move stop to breakeven once target one is hit; manage target two
16:00 Flat, no exceptions; journal the trade with T3 slope, v setting and pullback depth

The routine is deliberately boring. The T3 rewards the trader who checks the H4 slope once, waits for the pullback, and then does nothing, and it punishes the trader who chases every wiggle of the line. The 07:00 slope check takes thirty seconds and has saved me from more bad trades than any entry signal ever has. The 16:00 flat rule is the one I never break, because the one night I broke it cost me two weeks. A funded account is a job, and this is the shift pattern.

Frequently asked questions

What is the T3 moving average?

The T3 is Tim Tillson’s smoother, published in 1998. It applies a generalised DEMA three times in a row, which expands into six cascaded EMAs combined with volume-factor coefficients. The result is a moving average that is smoother than a DEMA and tracks price better than a triple EMA, with one dial, the volume factor, that tunes the balance between lag and smoothness.

What are the best T3 settings for prop firm trading?

Length 14 with a volume factor of 0.7 on H1 for entries, and length 21 at the same volume factor on H4 as the direction filter. A trade is only allowed when both slopes agree. Keep the volume factor between 0.5 and 0.85: below 0.5 the line gets noisy, above 0.85 it gets slow and the overshoot quirk gets worse.

Is the T3 better than the EMA?

It is quieter, not better. On the same ten H1 bars the T3 trailed the same-length EMA by four pips at the end of a rising run, but its slope flipped far fewer times on the pullbacks. If you scalp, take the EMA. If you trade H1 on a funded account and your enemy is overtrading, take the T3. In my twenty-trade test the T3 filter finished $3,750 ahead of plain EMA crossovers.

How do you trade a pullback to the T3?

Wait for price to dip to a rising T3, watch for a close back above the line, and enter on the next bar’s open. The touch itself is not the signal; the close back above it is. Put your stop below the pullback low with an ATR buffer, never exactly on the line, and size from the stop: lots equal your risk budget divided by stop distance times pip value.

Why does the T3 sometimes overshoot price?

The extrapolation inside the generalised DEMA can push the line briefly beyond the last few bars of price after a violent turn, usually a news spike. It is a known quirk of the formula, not a bug. Never place a stop exactly on the T3 line, and never trust the first few bars of a T3 after a gap: give the six EMA stages time to rebuild.

Can you use the T3 on a funded account?

Yes. It is a plain moving average with an open formula, so no prop firm bans it. The funded-account value is the filter: a flat T3 slope means no trade, which protects your daily loss limit from chop. Just remember the T3 has no volatility or volume read, so pair it with a hard daily stop, a session filter and a 30-minute news blackout.

Risk disclaimer: Trading leveraged products on a funded account can lose more than your challenge fee if you break the firm’s rules, and past results do not guarantee future results. The numbers in this article come from my own journal on a $100,000 account with a $500 daily loss limit; your firm’s limits, spreads and leverage may differ, so check them before you trade. This article is educational and is not financial advice. For an independent overview of the T3 formula, its volume-factor settings and the overshoot quirk, see the LuxAlgo T3 indicator guide.

Check your own numbers

Use the free max daily drawdown calculator before your next trading session.

Open the calculator

Join the discussion

Your email address will not be published. Please keep it respectful and on topic. Links are not allowed.

Recent articles

Remi Robin
About the author

Remi Robin

Independent forex trader and writer

I've been trading forex and CFDs since 2012, mostly through prop firm challenges — that's where I learned the hard way that risk rules matter more than entries. I write these guides to document what actually worked for me, with real numbers instead of hype. Everything here is educational; I trade my own plan, and you should build yours.

More about this site →