Trend Trail, Trailing Stop & Buy Sell Signals [LunqFX]

An ATR trailing stop — the trend-following construction most traders know as SuperTrend — breaks in the same place every time. Price stops trending, the trailing stop gets clipped from both sides, and it prints a buy, a sell, a buy and a sell inside twenty bars. Every one of those is a false trend reversal, and the logic is not wrong: it is being asked a question the market is not answering.

This is an open-source modification of the classic ATR based SuperTrend, and it asks that question first. Before it will give you a buy or sell signal it measures whether there is a trend to trail at all. When there is not, the whole chart goes dark — the trailing stop disappears, the fill drops, the candles fall to grey, and no long entry or short entry prints.

And it does not ask you to take that on trust. A plain fixed-distance ATR trailing stop runs alongside it on the same data, and the panel shows both counts side by side with the difference worked out for you.

Included: an average true range trailing stop with adaptive distance, a self-calibrating trend regime filter, buy and sell signals with the stop level printed on every label, a dormant state that switches the chart off in ranges, a live dashboard, and alerts on every trend reversal.

❶ THE REGIME FILTER — what this adds to a SuperTrend

Trend strength is measured with the Kaufman Efficiency Ratio: the ground price actually covered, divided by the distance it travelled getting there. A clean leg scores near 1. The same distance walked back and forth scores near 0.

That raw ratio is useless as a threshold on its own, and this is where most attempts at this fail. Gold on a 30-minute chart runs an efficiency around 0.01 while the euro daily runs 0.40 — any fixed cutoff leaves the fast charts permanently asleep and the slow ones permanently awake. So the reading is scored as a PERCENTILE of the symbol's own recent history. The trail arms when efficiency reaches the top third of what this instrument normally manages, whatever that happens to be. One setting, no per-symbol tuning.

Two guards keep the state from flickering, because they catch different things. Hysteresis handles wobble around the threshold: once armed, the regime stays armed until efficiency drops clearly below the line. A minimum dwell time handles the other case — a clean spike that clears the threshold by a mile and drops straight back. Without both, a filter opens hundreds of regimes and ends up emitting more marks than the raw trail it was meant to quieten.

❷ DORMANCY — the trail does not exist in a range

This is stronger than dimming a colour. When no regime is live the trail is torn down completely, and it is rebuilt from the current price when one opens, taking its side from the move that woke it.

The reason is not cosmetic. A trail left running through a range turns over inside it, unseen, and the market then re-opens onto a direction that was decided while nobody was watching — a position with no entry behind it. Destroying and rebuilding means every segment on the chart begins with a real event, and every event gets a label.

What you see is a chart that is either lit or switched off. Grey candles, no line, no signal: there is nothing here to do, and you can read that from across the room.

❸ THE ATR TRAILING STOP AND ITS ADAPTIVE DISTANCE

The average true range sets the band width, and the trailing stop ratchets in the direction of the trend and never loosens — the same dynamic support and resistance line a SuperTrend gives you, flipping to a trend reversal when price closes through it. The stop level is printed on every buy and sell label, so the one number you need at the moment of a long entry or short entry is already on the chart.

The distance is not fixed. One multiple has to be either too tight for choppy conditions or too loose for a clean run — it cannot be right for both, so the distance widens as efficiency falls and tightens as it rises. Turn it off in the settings for a constant multiple.

❹ THE RECEIPT — a filter you can audit

A second trailing stop is computed on the same bars: fixed distance, no regime filter, nothing else — what an ordinary trailing stop would have done here. Its flip count sits in the panel next to this one's signal count, with the reduction calculated:

Signals here · plain trail 164 · 236 Noise removed −31%

Both numbers count the same thing — entries against entries. A state is not counted as a trade on either side. And when the result goes the wrong way the panel says "Noise ADDED" in red rather than quietly dropping the sign, because a panel that flatters its own script is worse than no panel.

Read it as what it is: a measure of how much less often this fires, not a claim about money. Fewer signals is not automatically better signals, and this number does not pretend otherwise.

❺ THE DASHBOARD

Direction and stop price in the header, trend strength as a 0–100 reading with a bar and the arming threshold beneath it, current stop distance in price and in ATR, and the two comparison rows. In the dormant state the header says so plainly and the stop row reads "no stop — dormant" rather than printing a number that does not exist.

HOW TO USE IT

1 — Trade the lit stretches, ignore the grey ones. That is the whole discipline the tool is built around, and it is the part most trend systems leave to you.

2 — Use the trail as the stop, not just as a signal line. The level on the label is where the stop goes; the panel keeps showing the distance in ATR as the trade runs, so you can see when the trail has tightened to the point of being one bar away.

3 — Set the arming threshold to your patience. At 65 you get the top third of this symbol's clean moves. Raise it to 75 and you will trade far less on far cleaner legs. This is the one setting worth changing.

4 — Read the comparison rows on your own instrument. If the reduction on your symbol and timeframe is small, the filter is not finding much to remove there — which is itself information about the instrument, not a reason to distrust the reading.

HOW IT WORKS

The average true range sets the band width; the mid price plus and minus that width form the raw bands, exactly as in a classic ATR trailing stop. Each band ratchets in the trend's favour and never against it, and price closing through the opposite band flips the direction — the trend reversal. Efficiency is the net move over the lookback divided by the summed absolute bar-to-bar movement, ranked as a percentile against its own history. The regime arms above the threshold with hysteresis and a minimum dwell, and outside a regime the trail is not computed at all.

Works on any symbol and any timeframe. The regime filter needs the self-calibration window to fill before it can arm, so the first stretch of a fresh chart stays dormant by design.

SETTINGS

▸ Trail — ATR length, base distance, adaptive distance and its strength. ▸ Regime Filter — on or off, efficiency lookback, self-calibration window, arming percentile, hysteresis, minimum bars per regime. ▸ Signals — buy and sell signals, labels or arrows, stop level on the label. ▸ Visuals — glow, fill, candle dimming, dashboard position.

ALERTS — buy, sell, any signal, regime opened and regime closed. All fire on closed bars.

NON-REPAINTING — the trail is built from closed-bar values and every signal fires on bar close. A printed signal never moves and never disappears.

WHY THESE PARTS ARE ONE SCRIPT

The trail alone is an ordinary trailing stop and will chop you up in a range. The regime filter alone has nothing to gate. The comparison exists only because a filter nobody can check is just a claim, and it needs both of the others to have something to measure. Take any one away and the other two stop making a point.

This indicator is an educational market-analysis tool, not financial advice. It does not predict price. The comparison figures describe how often each version of the trail changed direction on the loaded chart; they say nothing about profit or loss. Always confirm with your own analysis and manage your risk.