ianzeng123
EMA, ATR, MOMENTUM, EFFICIENCY, BREAKOUT
This Isn’t Your Average Triple EMA Strategy - It’s a Multi-Dimensional Trend Machine
Don’t be fooled by the surface-level 18/50/120 EMA setup. The core of this strategy is an 8-dimensional trend validation system, each with precise numerical standards. It’s not simple golden/death crosses, but a weighted scoring mechanism (minimum 5.0 for entry, below 2.5 for exit) that judges trend quality.
Traditional EMA strategies suffer from excessive false signals. This system elevates entry success rates through multiple filters: path efficiency (minimum 33%), momentum persistence (57%+ bullish bar ratio), volatility regime (95%+ ATR ratio), and more.
Breakout Quality Detection: 0.15x ATR Minimum for Real Breakouts
90% of market breakouts are false. This strategy requires breakout strength of at least 0.15x ATR, meaning the breakout magnitude must exceed 15% of recent average volatility to qualify as valid.
The pullback reclaim mechanism is even more sophisticated: requiring price to pull back at least 0.9x ATR from the fast EMA, then reclaim with 0.15x ATR strength. This design effectively filters shallow fake breakouts, capturing only genuinely fund-driven trend initiations.
2x Leverage + 2.8x ATR Trailing Stop Risk Logic
The 2x leverage appears aggressive but paired with 2% hard stop and 2.8x ATR dynamic trailing, actual risk remains controlled. More critically, the 20.8x ATR profit lock mechanism automatically raises stop levels when floating profits reach this threshold, ensuring major trend gains don’t evaporate.
The mandatory 1-bar hold design prevents high-frequency churn, while the 5-bar cooldown period avoids emotional consecutive trades. This rhythm control matters more than pure technical indicators.
Three Entry Modes Cover Different Market States
Trend continuation entry: for established strong trends, requiring breakout + slope + efficiency + momentum all meeting standards. Pullback re-entry: targeting healthy trend corrections with sufficient depth and powerful reclaim. Early trend entry: capturing the golden 14-bar window of trend transitions.
This multi-mode design ensures no trend opportunity gets missed while maintaining strict quality standards for each mode. It’s precision targeting, not casting wide nets.
Efficiency Indicator: The Most Undervalued Trend Quality Tool
Path efficiency calculates the ratio of net displacement to cumulative displacement over 18 periods, rejecting any trend below 33%. This indicator effectively identifies false trends in choppy markets, avoiding repeated stops in sideways action.
Momentum persistence requires 57%+ bullish bars combined with positive 12-period momentum. This dual verification ensures internal trend strength beyond mere surface price breakouts.
Applicability Analysis: Not Universal, But Exceptional in Trending Markets
This strategy clearly favors trending markets, frequently triggering exits in choppy conditions. The forced exit when ATR regime drops below 80% and efficiency below 25% shows the strategy has clear environmental requirements.
The biggest risk lies in lag during trend reversals. Despite early warnings like fast EMA crosses, sharp reversals can still cause significant drawdowns. Recommended for high-volatility growth stocks or crypto markets, avoid low-volatility blue chips.
Risk Warning: Historical backtests don’t guarantee future returns. Strategy carries consecutive loss risks requiring strict risk management. Performance varies significantly across different market environments.
/*backtest
start: 2026-01-07 15:30:00
end: 2026-03-15 08:00:00
period: 1h
basePeriod: 1h
exchanges: [{"eid":"Futures_Binance","currency":"XAG_USDT","balance":500000}]
*/
//@version=5
strategy("Quant Trend Engine Long Only v2 - Manual Leverage Fixed", overlay=true)
// === Core lengths ===
fastLen = input.int(18, "Fast EMA")
midLen = input.int(50, "Mid EMA")
slowLen = input.int(120, "Slow EMA")
smoothLen = input.int(3, "EMA Smoothing")
pullbackLen = input.int(8, "Pullback Lookback")
breakoutLen = input.int(20, "Breakout Length")
effLen = input.int(18, "Efficiency Length")
persistLen = input.int(7, "Persistence Length")
momLen = input.int(12, "Momentum Length")
slopeLen = input.int(10, "Slope Length")
atrLen = input.int(14, "ATR Length")
atrBaseLen = input.int(40, "ATR Baseline Length")
// === Thresholds ===
minScore = input.float(5.0, "Minimum Entry Score", step=0.25)
exitScore = input.float(2.5, "Weak Trend Exit Score", step=0.25)
minSepPerc = input.float(0.30, "Min EMA Separation %", step=0.05)
minSlowSlopePerc = input.float(0.03, "Min Slow Slope %", step=0.01)
minEff = input.float(0.33, "Min Efficiency", step=0.01)
minAtrRegime = input.float(0.95, "Min ATR Regime", step=0.05)
minBreakoutAtr = input.float(0.15, "Min Breakout ATR Strength", step=0.05)
pullbackAtrMult = input.float(0.90, "Pullback Distance ATR", step=0.05)
reclaimAtrMult = input.float(0.15, "Reclaim Distance ATR", step=0.05)
cooldownBars = input.int(5, "Cooldown Bars After Exit")
// === Risk ===
leverage = input.float(2.0, "Leverage", step=0.1, minval=0.1)
hardStopPerc = input.float(2.0, "Hard Stop %", step=0.1)
trailAtrMult = input.float(2.8, "ATR Trail Mult", step=0.1)
profitLockAtrMult = input.float(20.8, "Profit Lock ATR Mult", step=0.1)
// === Smoothed EMAs ===
fast = ta.ema(ta.ema(close, fastLen), smoothLen)
mid = ta.ema(ta.ema(close, midLen), smoothLen)
slow = ta.ema(ta.ema(close, slowLen), smoothLen)
// === Regime structure ===
bullStack = fast > mid and mid > slow
sepPerc = slow != 0 ? math.abs(fast - slow) / slow * 100 : 0.0
sepOk = sepPerc >= minSepPerc
fastSlope = fast[slopeLen] != 0 ? (fast - fast[slopeLen]) / fast[slopeLen] * 100 : 0.0
midSlope = mid[slopeLen] != 0 ? (mid - mid[slopeLen]) / mid[slopeLen] * 100 : 0.0
slowSlope = slow[slopeLen] != 0 ? (slow - slow[slopeLen]) / slow[slopeLen] * 100 : 0.0
slopeOk = slowSlope >= minSlowSlopePerc and midSlope > 0 and fastSlope > 0
// === Path efficiency ===
netMove = math.abs(close - close[effLen])
stepMove = 0.0
for i = 1 to effLen
stepMove += math.abs(close[i - 1] - close[i])
efficiency = stepMove != 0 ? netMove / stepMove : 0.0
effOk = efficiency >= minEff
// === Momentum persistence ===
upBars = 0.0
for i = 0 to persistLen - 1
upBars += close[i] > close[i + 1] ? 1 : 0
persistRatio = persistLen > 0 ? upBars / persistLen : 0.0
momRaw = close[momLen] != 0 ? (close - close[momLen]) / close[momLen] * 100 : 0.0
momOk = momRaw > 0 and persistRatio >= 0.57
// === Volatility regime ===
atr = ta.atr(atrLen)
atrBase = ta.sma(atr, atrBaseLen)
atrRegime = atrBase != 0 ? atr / atrBase : 0.0
atrOk = atrRegime >= minAtrRegime
// === Breakout quality ===
hh = ta.highest(high, breakoutLen)[1]
breakoutDist = close - hh
breakoutStrength = atr != 0 ? breakoutDist / atr : 0.0
breakoutOk = close > hh and breakoutStrength >= minBreakoutAtr
// === Pullback / reclaim logic ===
pullbackLow = ta.lowest(low, pullbackLen)
distFromFastAtr = atr != 0 ? (fast - pullbackLow) / atr : 0.0
deepEnoughPullback = distFromFastAtr >= pullbackAtrMult
reclaimFast = close > fast and close[1] <= fast[1]
reclaimMid = close > mid and close[1] <= mid[1]
reclaimStrength = atr != 0 ? (close - fast) / atr : 0.0
reclaimOk = (reclaimFast or reclaimMid) and reclaimStrength >= reclaimAtrMult
// === Transition memory ===
bullCross = ta.crossover(fast, mid) or ta.crossover(fast, slow) or ta.crossover(mid, slow)
barsSinceBullCross = ta.barssince(bullCross)
recentTrendBirth = barsSinceBullCross >= 0 and barsSinceBullCross <= 14
// === Weighted score ===
trendScore = 0.0
trendScore += bullStack ? 1.50 : 0.0
trendScore += sepOk ? 0.90 : 0.0
trendScore += slopeOk ? 1.10 : 0.0
trendScore += effOk ? 1.00 : 0.0
trendScore += atrOk ? 0.80 : 0.0
trendScore += momOk ? 1.00 : 0.0
trendScore += breakoutOk ? 1.25 : 0.0
trendScore += reclaimOk ? 1.10 : 0.0
// === Entry models ===
trendContinuationEntry = bullStack and breakoutOk and slopeOk and effOk and momOk
pullbackReentry = bullStack and sepOk and slopeOk and deepEnoughPullback and reclaimOk and effOk
earlyTrendEntry = recentTrendBirth and bullStack and sepOk and slopeOk and atrOk and momOk
// === Cooldown ===
var int lastExitBar = na
cooldownOk = na(lastExitBar) or bar_index - lastExitBar > cooldownBars
// === Final entry ===
enterLong = strategy.position_size == 0 and cooldownOk and trendScore >= minScore and close > slow and (trendContinuationEntry or pullbackReentry or earlyTrendEntry)
// === Manual leveraged sizing only ===
equity = math.max(strategy.equity, 0)
positionValue = equity * leverage
qty = positionValue > 0 ? positionValue / (close * syminfo.pointvalue) : 0.0
// === Entry tracking / mandatory 1 full candle hold ===
var int entryBarIndex = na
justOpened = strategy.position_size > 0 and strategy.position_size[1] == 0
if justOpened
entryBarIndex := bar_index
canExitNow = strategy.position_size > 0 and not na(entryBarIndex) and bar_index > entryBarIndex
// === Entry order ===
if enterLong and qty > 0
strategy.entry("Long", strategy.long, qty=qty)
// === Risk logic ===
hardStopPrice = strategy.position_size > 0 ? strategy.position_avg_price * (1 - hardStopPerc / 100) : na
var float trailStop = na
var float highSinceEntry = na
highSinceEntry := strategy.position_size > 0 ? (na(highSinceEntry) ? high : math.max(highSinceEntry, high)) : na
rawTrail = strategy.position_size > 0 ? close - atr * trailAtrMult : na
profitLock = strategy.position_size > 0 ? highSinceEntry - atr * profitLockAtrMult : na
combinedTrail = strategy.position_size > 0 ? math.max(rawTrail, profitLock) : na
trailStop := strategy.position_size > 0 ? (na(trailStop) ? combinedTrail : math.max(trailStop, combinedTrail)) : na
// === Exit logic ===
bearCross = ta.crossunder(fast, mid) or ta.crossunder(fast, slow)
structureBreak = close < mid and fast < mid
scoreWeak = trendScore <= exitScore
momentumFailure = persistRatio < 0.40 and momRaw < 0
regimeFailure = atrRegime < 0.80 and efficiency < 0.25
exitLong = strategy.position_size > 0 and canExitNow and (bearCross or structureBreak or scoreWeak or momentumFailure or regimeFailure)
// Only allow stop/trailing exits after 1 full candle has passed
if strategy.position_size > 0 and canExitNow
strategy.exit("Risk Exit", from_entry="Long", stop=math.max(hardStopPrice, trailStop))
justClosed = strategy.position_size[1] > 0 and strategy.position_size == 0
if justClosed
lastExitBar := bar_index
trailStop := na
highSinceEntry := na
entryBarIndex := na
if exitLong
strategy.close("Long")
// === Plots ===
plot(fast, color=color.green, linewidth=2, title="Fast EMA")
plot(mid, color=color.orange, linewidth=2, title="Mid EMA")
plot(slow, color=color.red, linewidth=2, title="Slow EMA")
plot(strategy.position_size > 0 ? trailStop : na, color=color.blue, linewidth=2, title="Adaptive Trail")
plot(trendScore, title="Trend Score", color=color.aqua)
plot(efficiency, title="Efficiency", color=color.fuchsia)
plot(atrRegime, title="ATR Regime", color=color.yellow)
plot(breakoutStrength, title="Breakout Strength", color=color.lime)
plot(persistRatio, title="Persistence Ratio", color=color.white)