OPEN-SOURCE SCRIPT
AetherEdge - Smart Money Flow
🖊️ Overview
AE-SMF focuses on what smart-money price action is really about — liquidity. It maps the stop pools resting above swing highs (buyside) and below swing lows (sellside) and learns which liquidity price reaches for next (draw on liquidity). The liquidity, sweep, displacement/FVG and premium/discount concepts are re-implemented from public material, with a genuine learning core on top.
🔶 Key Features
Liquidity pool map — buyside/sellside pools, auto-removed once taken
Sweep detection — a grab of one side's liquidity followed by reclaim
ML: draw on liquidity — learns P(the buyside pool is reached before the sellside pool), turning the pull of price into a probability
RL: self-tuning bias threshold — a UCB bandit auto-tunes the firing threshold per structure regime
Sweep-origin signals — TP set to the targeted liquidity pool, SL beyond the swept extreme
Displacement/FVG and premium/discount included
Structure uses confirmed pivots; training and signals gate on bar close — no repaint
🧠 Technical Architecture
Liquidity engine: buyside/sellside pools from confirmed pivots are held in arrays; pools price trades through (taken liquidity) are removed; the nearest pools are drawn as lines.
Sweeps: piercing sellside then reclaiming up = bullish sweep; piercing buyside then closing down = bearish sweep.
ML (online logistic regression): eight features (trend, displacement tendency, premium/discount position, recent bullish/bearish sweep, distance to the buyside/sellside pools, structure direction) feed a model of the draw = P(buyside pool reached before sellside, within N bars). Each confirmed bar labels a pool-to-pool barrier — which of the nearest buyside target (above) and sellside target (below) is touched first — so the prediction uses no future data. If neither is reached in time, the sample is discarded (only clean draws train it).
RL (UCB contextual bandit): auto-tunes the bias firing threshold per structure regime (trend-strength terciles). Its reward is tied to the same draw resolution (a confident bias that proves correct = +1, wrong = −penalty, abstaining when neutral = a small reward) — so the ML training and the RL reward share one judgment loop.
Signals: long = bullish sweep + buyside draw ≥ threshold + discount; short = the mirror. TP is the targeted liquidity pool; SL is beyond the swept extreme.
Honest scope: a linear classifier + a UCB bandit over standard liquidity/structure features. Not deep learning, not a guarantee.
⚙️ Recommended Settings & Tuning Guide
Key parameters: swing lookback (pivots), pools per side, displacement multiple, barrier horizon N, learning rate, threshold search range, stop buffer.
Larger swing lookback → focus on major-structure pools; smaller → more detections
N is the window for deciding which liquidity is reached first (smaller for short-term, larger for swing)
Crypto starting points (tune on your chart):
BTC / ETH (15m–4H): defaults are the baseline (pivLen 10, N 20)
SOL / XRP and high-vol alts: a slightly longer swing lookback and a wider stop buffer to avoid wick-hunts
Scalping (1–15m): smaller pivot length and N
Swing (4H–daily): larger pivot length, N, and warmup
Bias, accuracy, and threshold are coarse until warmup plus enough draws accumulate
💡 How to Use in Practice
Read the draw bias for the liquidity price is likely to target next (buyside = up, sellside = down)
Enter on a sweep + agreeing-draw signal and use the drawn SL/TP (TP = the targeted pool)
Use premium/discount to favor advantageous pullbacks/rallies (buy discount, sell premium)
Leave the auto threshold to the learner by default
Combine with a higher-timeframe liquidity view or a precision-entry tool (AE-ACE)
⚠️ Important Notes
Needs a learning period (warmup); weights and the Q-table re-learn on input/symbol/timeframe change
Pools come from confirmed pivots, so the latest extreme confirms a few bars late (the trade-off for avoiding repaint)
A sweep does not guarantee a reversal (it can continue)
Probability, not a guarantee — always use stops and position sizing
🚨 Disclaimer
This indicator is for educational and informational purposes only and is not financial advice or a recommendation to buy or sell. No method guarantees future profits; past performance does not indicate future results, and trading carries the risk of loss. All trading decisions are your own — use proper backtesting and disciplined risk management.
AE-SMF focuses on what smart-money price action is really about — liquidity. It maps the stop pools resting above swing highs (buyside) and below swing lows (sellside) and learns which liquidity price reaches for next (draw on liquidity). The liquidity, sweep, displacement/FVG and premium/discount concepts are re-implemented from public material, with a genuine learning core on top.
🔶 Key Features
Liquidity pool map — buyside/sellside pools, auto-removed once taken
Sweep detection — a grab of one side's liquidity followed by reclaim
ML: draw on liquidity — learns P(the buyside pool is reached before the sellside pool), turning the pull of price into a probability
RL: self-tuning bias threshold — a UCB bandit auto-tunes the firing threshold per structure regime
Sweep-origin signals — TP set to the targeted liquidity pool, SL beyond the swept extreme
Displacement/FVG and premium/discount included
Structure uses confirmed pivots; training and signals gate on bar close — no repaint
🧠 Technical Architecture
Liquidity engine: buyside/sellside pools from confirmed pivots are held in arrays; pools price trades through (taken liquidity) are removed; the nearest pools are drawn as lines.
Sweeps: piercing sellside then reclaiming up = bullish sweep; piercing buyside then closing down = bearish sweep.
ML (online logistic regression): eight features (trend, displacement tendency, premium/discount position, recent bullish/bearish sweep, distance to the buyside/sellside pools, structure direction) feed a model of the draw = P(buyside pool reached before sellside, within N bars). Each confirmed bar labels a pool-to-pool barrier — which of the nearest buyside target (above) and sellside target (below) is touched first — so the prediction uses no future data. If neither is reached in time, the sample is discarded (only clean draws train it).
RL (UCB contextual bandit): auto-tunes the bias firing threshold per structure regime (trend-strength terciles). Its reward is tied to the same draw resolution (a confident bias that proves correct = +1, wrong = −penalty, abstaining when neutral = a small reward) — so the ML training and the RL reward share one judgment loop.
Signals: long = bullish sweep + buyside draw ≥ threshold + discount; short = the mirror. TP is the targeted liquidity pool; SL is beyond the swept extreme.
Honest scope: a linear classifier + a UCB bandit over standard liquidity/structure features. Not deep learning, not a guarantee.
⚙️ Recommended Settings & Tuning Guide
Key parameters: swing lookback (pivots), pools per side, displacement multiple, barrier horizon N, learning rate, threshold search range, stop buffer.
Larger swing lookback → focus on major-structure pools; smaller → more detections
N is the window for deciding which liquidity is reached first (smaller for short-term, larger for swing)
Crypto starting points (tune on your chart):
BTC / ETH (15m–4H): defaults are the baseline (pivLen 10, N 20)
SOL / XRP and high-vol alts: a slightly longer swing lookback and a wider stop buffer to avoid wick-hunts
Scalping (1–15m): smaller pivot length and N
Swing (4H–daily): larger pivot length, N, and warmup
Bias, accuracy, and threshold are coarse until warmup plus enough draws accumulate
💡 How to Use in Practice
Read the draw bias for the liquidity price is likely to target next (buyside = up, sellside = down)
Enter on a sweep + agreeing-draw signal and use the drawn SL/TP (TP = the targeted pool)
Use premium/discount to favor advantageous pullbacks/rallies (buy discount, sell premium)
Leave the auto threshold to the learner by default
Combine with a higher-timeframe liquidity view or a precision-entry tool (AE-ACE)
⚠️ Important Notes
Needs a learning period (warmup); weights and the Q-table re-learn on input/symbol/timeframe change
Pools come from confirmed pivots, so the latest extreme confirms a few bars late (the trade-off for avoiding repaint)
A sweep does not guarantee a reversal (it can continue)
Probability, not a guarantee — always use stops and position sizing
🚨 Disclaimer
This indicator is for educational and informational purposes only and is not financial advice or a recommendation to buy or sell. No method guarantees future profits; past performance does not indicate future results, and trading carries the risk of loss. All trading decisions are your own — use proper backtesting and disciplined risk management.
Open-source script
In true TradingView spirit, the creator of this script has made it open-source, so that traders can review and verify its functionality. Kudos to the author! While you can use it for free, remember that republishing the code is subject to our House Rules.
Disclaimer
The information and publications are not meant to be, and do not constitute, financial, investment, trading, or other types of advice or recommendations supplied or endorsed by TradingView. Read more in the Terms of Use.
Open-source script
In true TradingView spirit, the creator of this script has made it open-source, so that traders can review and verify its functionality. Kudos to the author! While you can use it for free, remember that republishing the code is subject to our House Rules.
Disclaimer
The information and publications are not meant to be, and do not constitute, financial, investment, trading, or other types of advice or recommendations supplied or endorsed by TradingView. Read more in the Terms of Use.