Algorithm Library

59 ML-based and quantitative trading algorithms. Each runs in real time on live market data across multiple timeframes. Assign any combination to any asset — all algorithms can run simultaneously.

ARIMA Forecast Engine

Uses autoregressive integrated moving average modeling to forecast the next-period price via Yule-Walker equations solved with Levinson-Durbin recursion.

▸Estimates AR(p) coefficients via Yule-Walker equations
▸Forecasts next-bar price
▸Entry when forecast deviation exceeds residual SE

HMM Regime Switching

A 3-state Hidden Markov Model classifies the market into Bull, Bear, and Range regimes using Gaussian emissions and a learned transition matrix.

▸3-state HMM with Gaussian emissions
▸Forward algorithm computes posterior state probabilities
▸Enters when directional regime probability > 70%

Logistic Regression ML

A true ML binary classifier that predicts next-bar direction using online gradient descent with 8 engineered features.

▸Online logistic regression with 8 features
▸Trained via SGD on logistic loss
▸Enters when probability > 0.65 or < 0.35

Adaptive Perceptron

An online perceptron classifier with the kernel trick — updates weights sharply on misclassification for fast regime adaptation.

▸Classic perceptron with online updates
▸8 features: momentum, volatility, volume, RSI
▸Sharp adaptation to regime shifts

PCA Eigenvector Momentum

Performs PCA on multi-horizon return vectors to extract the dominant signal direction via power iteration.

▸Covariance matrix of multi-horizon returns
▸Power iteration extracts principal eigenvector
▸Filters false signals by requiring alignment

News Sentiment Enhanced

Combines quantitative price momentum with LLM-powered news sentiment analysis. Entry requires alignment between technical and AI sentiment signals.

▸Uses InvokeLLM for market sentiment analysis
▸Sentiment cached for 5 minutes
▸Entry requires tech + sentiment alignment

Opening Range Breakout

Establishes the opening range as the high/low of the first 15 bars, then enters on breakouts. Uses ATR-based trailing stop.

▸Opening range from first 15 bars
▸ATR-based trailing stop (2× ATR)
▸Exits on trailing stop or range midpoint cross

Dual EMA Neural Gate

9/21 EMA crossover enhanced with an online logistic regression confidence gate. Only enters when ML confidence > 0.65.

▸9/21 EMA crossover detection
▸Online logistic regression gate
▸Filters ~40% of false signals

RSI Divergence ML

Detects bullish/bearish RSI divergences using linear regression on both price and RSI to statistically confirm the pattern.

▸Detects price/RSI divergence via pivots
▸Linear regression confirms divergence
▸High win-rate mean-reversion entries

ATR Supertrend Adaptive

ATR-based supertrend with adaptive multiplier (1.5x-3.5x based on volatility percentile). Entries on supertrend flips with volume confirmation.

▸Adaptive multiplier based on ATR percentile
▸Supertrend line ratchets
▸Volume confirmation required

Ornstein-Uhlenbeck Reversion

Models price as an Ornstein-Uhlenbeck process. Estimates parameters via MLE and computes mean-reversion half-life.

▸Fits OU process via MLE
▸Computes mean-reversion half-life
▸Only trades when half-life < 20 bars

Multi-Factor Momentum

Combines three orthogonal alpha factors — price momentum, volume-weighted momentum, and trend slope — into a composite z-scored signal.

▸Three orthogonal factors
▸Each factor z-scored
▸Composite signal must exceed 1.5σ

Bollinger Squeeze Breakout

Detects Bollinger Band squeezes and enters on band breakouts with ML confirmation.

▸Bandwidth percentile over 60 bars
▸Squeeze when bandwidth < 20th percentile
▸Entry on band breakout after squeeze

Volume Profile Nodes

Constructs a volume profile histogram to identify High Volume Nodes (HVN). Enters when price deviates from HVN.

▸Builds volume profile (20 bins)
▸Identifies HVN
▸Entry when price deviates >0.4% from HVN

MACD Histogram Classifier

Enhances MACD with a gradient-boosted ensemble of decision stumps on histogram features.

▸MACD (12,26,9) with histogram
▸Ensemble of 5 decision stumps
▸Filters noise from standard MACD

Keltner Channel Breakout

Uses Keltner Channels (EMA ± 2×ATR) with an ADX directional movement filter for trend breakouts.

▸Keltner Channel: 20-period EMA ± 2×ATR
▸ADX-like directional movement filter
▸Entry on channel breakout when ADX > 25

Choppiness Regime Filter

Uses the Choppiness Index to dynamically switch between trend-following and mean-reverting logic based on market regime.

▸CI < 38.2 = trending → trade breakouts
▸CI > 61.8 = choppy → fade extremes
▸38.2 < CI < 61.8 = neutral → no trades

Kalman Filter Reversion

Uses a Kalman filter to adaptively track fair-value price. Trades entered when price deviates >2σ from the Kalman estimate.

▸Adaptive Kalman filter for price tracking
▸Entry at >2σ deviation
▸Exit when price reverts to estimate

Hurst Regime Momentum

Classifies market regime using the Hurst exponent. H > 0.55 → trending (breakout entries), H < 0.45 → mean-reverting (fade entries).

▸Hurst exponent via R/S analysis
▸Trending regime: breakout entries
▸Mean-reverting regime: fade entries

Z-Score Bollinger

Computes a rolling z-score of price against its 20-period mean and SD. Entries at ±2σ, exits when z crosses 0.

▸True statistical z-score
▸Entry at ±2σ (95% CI)
▸Exit when z-score crosses 0

Wavelet Denoised Trend

Applies Haar wavelet decomposition to remove noise from the price series, revealing the underlying trend.

▸Haar wavelet decomposition
▸Soft-thresholds detail coefficients
▸Entry on denoised trend slope flip

VWAP Deviation

Computes Volume-Weighted Average Price and enters reversion trades when price deviates significantly from VWAP.

▸VWAP = Σ(price × volume) / Σ(volume)
▸Entry long when price >0.3% below VWAP
▸Exit when price crosses VWAP

Donchian Channel Breakout

Classic turtle trading system. Enters on breakouts above/below the 20-period Donchian channel.

▸20-period Donchian channel
▸Entry on channel breakout
▸Exit on 10-period channel reversal

Stochastic Reversion

Stochastic Oscillator (%K/%D) identifies overbought/oversold conditions. Enters on %K/%D crossovers in extreme zones.

▸14-period %K with 3-period %D
▸Entry on crossover in <20 or >80 zones
▸Exit when %K crosses 50

Williams %R Reversion

Williams %R momentum oscillator. Enters mean-reversion trades at extreme levels (<-80 oversold, >-20 overbought).

▸14-period Williams %R
▸Entry at %R < -80 or > -20
▸Exit when %R crosses -50

CCI Mean Reversion

Commodity Channel Index measures deviation from moving average. Enters at ±100 thresholds for mean reversion.

▸20-period CCI
▸Entry at CCI < -100 or > +100
▸Exit when CCI crosses 0

Ichimoku Cloud Breakout

Ichimoku Kinko Hyo system. Uses Tenkan-sen, Kijun-sen, and Senkou Span to identify trend direction and cloud breakouts.

▸Tenkan-sen (9), Kijun-sen (26)
▸Cloud (Senkou Span A/B)
▸Entry on price breakout above/below cloud

Parabolic SAR Reversal

Parabolic Stop and Reverse. Trails price with accelerating SAR dots. Enters on SAR flip.

▸PSAR with AF starting at 0.02, max 0.2
▸Entry on SAR flip
▸Trailing stop mechanism

Money Flow Index Reversion

Volume-weighted RSI variant. Uses price × volume to identify overbought/oversold with volume confirmation.

▸14-period MFI
▸Entry at MFI < 20 or > 80
▸Volume-weighted momentum

ADX Trend Strength

Pure ADX trend filter. Only enters when ADX > 25 (strong trend) and DI+/DI- crossover confirms direction.

▸14-period ADX
▸DI+ and DI- crossover
▸Entry only when ADX > 25

Linear Regression Channel

Fits a linear regression line and computes ±2SE channel. Trades when price breaks the channel.

▸20-period linear regression
▸±2 SE channel bands
▸Entry on channel breakout

Chandelier Exit Trend

ATR-based trailing stop from highest high/lowest low. Follows trends with a 3×ATR chandelier exit.

▸22-period highest high/lowest low
▸3×ATR chandelier exit
▸Trend-following with trailing stop

Elder Ray Bull/Bear Power

Alexander Elder's Bull Power (high - EMA) and Bear Power (low - EMA). Combines trend (EMA) with momentum (power).

▸13-period EMA
▸Bull Power = High - EMA
▸Bear Power = Low - EMA
▸Entry when both powers align

Double EMA Crossover

Double EMA (DEMA) crossover system. DEMA reduces lag compared to standard EMA for faster signal detection.

▸DEMA = 2*EMA - EMA(EMA)
▸20/50 period crossover
▸Reduced lag vs standard EMA

TRIX Momentum

Triple-smoothed EMA rate of change. Filters noise by triple smoothing, revealing underlying momentum.

▸12-period triple-smoothed EMA
▸Signal line (9-period EMA of TRIX)
▸Entry on TRIX/signal crossover

Vortex Indicator

Vortex Indicator (VI) uses directional movement to identify trend starts. VI+ > VI- = bullish, VI- > VI+ = bearish.

▸14-period VI+ and VI-
▸Entry on VI crossover
▸True range normalization

Awesome Oscillator

Bill Williams Awesome Oscillator. Difference between 5-period and 34-period SMA of median price. Detects momentum shifts.

▸AO = SMA(median,5) - SMA(median,34)
▸Entry on AO zero cross
▸Momentum direction detection

Know Sure Thing

KST momentum oscillator. Combines four different rate-of-change periods into a single momentum signal.

▸4 ROC periods (10,15,20,30)
▸Weighted sum with smoothing
▸Signal line crossover

On-Balance Volume Trend

OBV cumulative indicator. Uses volume flow to confirm price trends. Divergence between OBV and price signals reversal.

▸Cumulative volume-based indicator
▸OBV trend via linear regression
▸Divergence detection

Fibonacci Retracement

Identifies swing high/low and enters on Fibonacci retracement levels (38.2%, 50%, 61.8%) with trend confirmation.

▸Swing high/low detection
▸38.2%, 50%, 61.8% retracement levels
▸Entry on bounce from 61.8% level

Chaikin Money Flow

Chaikin Money Flow measures accumulation/distribution over N periods. Positive CMF = accumulation, negative = distribution.

▸20-period CMF
▸Money Flow Multiplier × Volume
▸Entry on CMF sign change with price confirmation

Aroon Trend System

Aroon Up/Down indicator identifies trend strength and direction. Aroon Up > 70 and Aroon Down < 30 = strong uptrend.

▸25-period Aroon Up/Down
▸Oscillator = Aroon Up - Aroon Down
▸Entry on crossover + oscillator > 0

Micro Kalman Scalper

High-frequency Kalman filter scalper. Tracks fair value adaptively and enters on tight 1σ deviations — half the threshold of standard mean-reversion. An online logistic regression gate filters low-probability entries. Exits lock profit at the first 0.1% favorable move, ensuring consistent small wins.

▸Adaptive Kalman filter for micro fair-value tracking
▸Entry at 1σ deviation (vs 2σ standard)
▸ML logistic regression confidence gate > 0.62
▸Profit-lock exit at +0.1% favorable move
▸Tight 0.4% stop, 0.15% target

Tick Momentum ML Scalper

Captures micro-momentum bursts using an online perceptron classifier trained on 1-3 bar returns. The perceptron makes sharp updates on misclassification, rapidly adapting to short-term direction. Only enters when classifier confidence is high. Profit-lock exit triggers at +0.12% favorable.

▸Online perceptron on 1/3/5-bar micro-momentum features
▸Sharp weight updates for fast regime adaptation
▸Entry only when |score| > 0.2 and ML aligned
▸Profit-lock exit at +0.12% favorable move
▸Tight 0.35% stop, 0.15% target

VWAP Micro Scalper

Scalps tiny deviations from Volume-Weighted Average Price. Enters when price deviates just 0.05% from VWAP — far tighter than standard VWAP strategies. An ML logistic regression gate confirms the direction. Exits immediately when price touches VWAP, locking in micro-profits on nearly every trade.

▸Entry at 0.05% VWAP deviation (vs 0.3% standard)
▸ML logistic regression gate > 0.6
▸Exit at VWAP touch (profit lock)
▸Very high win rate — VWAP acts as magnet
▸Tight 0.3% stop

Order Flow Imbalance Scalper

Detects micro buy/sell pressure imbalance using volume distribution within candle bodies. An ML perceptron classifies whether the imbalance predicts continuation. Enters on strong imbalance with ML confirmation. Profit-lock exit at +0.1%.

▸Volume-weighted body position analysis
▸Buy/sell pressure ratio from candle anatomy
▸ML perceptron classifier on imbalance features
▸Entry when imbalance > 60% and ML aligned
▸Profit-lock exit at +0.1% favorable

Z-Score Micro Scalper

Statistical z-score scalper with tight 1σ entries — half the standard threshold. Enters when price deviates 1 standard deviation from its 15-period mean. ML logistic regression gate filters entries. Exits when z crosses 0 or profit-lock triggers at +0.1%.

▸Rolling 15-period z-score
▸Entry at ±1σ (vs ±2σ standard)
▸ML logistic regression gate > 0.6
▸Exit when z crosses 0 or +0.1% profit lock
▸Tight 0.35% stop, 0.12% target

Wick Rejection Scalper

Detects price rejection via candle wick analysis. When a candle shows a long wick on one side, it signals institutional rejection of that price level. An ML perceptron classifies whether the rejection will lead to reversal. Enters on confirmed rejection. Profit-lock exit at +0.12%.

▸Wick-to-body ratio analysis for rejection detection
▸Requires wick > 2× body for signal
▸ML perceptron classifier on wick pattern features
▸Entry on confirmed rejection with ML alignment
▸Profit-lock exit at +0.12% favorable

Micro Mean Reversion Scalper

Adaptive EMA-based mean reversion scalper. Computes a fast 8-period EMA as the fair value and enters when price deviates by 0.15% from it — much tighter than standard. ML logistic regression gate confirms reversion probability. Exits on return to EMA or profit-lock at +0.1%.

▸8-period EMA as adaptive fair value
▸Entry at 0.15% deviation (vs 0.3%+ standard)
▸ML logistic regression gate > 0.6
▸Exit when price returns to EMA
▸Profit-lock exit at +0.1% favorable, tight 0.3% stop

Hyper Kalman Flash Scalper

Ultra-fast Kalman filter scalper with 0.5σ entries — half the threshold of the micro Kalman scalper. An 8-feature online logistic regression gate filters entries. Exits lock profit at the first 0.04% favorable tick, sniping micro-moves before any pullback.

▸Adaptive Kalman filter with 0.5σ entry (vs 1σ micro)
▸8-feature online logistic regression ML gate > 0.65
▸Hyper profit-lock exit at +0.04% favorable
▸Tightest 0.25% stop, 0.08% target
▸Designed for 15s–1m timeframes

Tick Snipe ML Scalper

Snipes micro-momentum bursts with an 8-feature online perceptron. Sharp weight updates adapt in real-time to tick-level direction. Exits at the first 0.04% favorable move — pure sniping with no patience for pullbacks.

▸8-feature online perceptron classifier
▸Score-based entry when |score| > 0.15
▸Hyper profit-lock exit at +0.04% favorable
▸Tight 0.22% stop, 0.07% target
▸Fastest adaptation to tick direction changes

Micro VWAP Flash Scalper

Ultra-tight VWAP scalper entering at just 0.02% deviation — 2.5× tighter than the micro VWAP scalper. Gaussian Naive Bayes classifies the reversion probability. Exits the instant price touches VWAP, locking micro-profits on nearly every trade.

▸Entry at 0.02% VWAP deviation (vs 0.05% micro)
▸Gaussian Naive Bayes ML classifier
▸Exit at VWAP touch or +0.05% profit lock
▸VWAP acts as ultra-strong magnet
▸Tight 0.2% stop

Z-Score Nano Scalper

Statistical z-score scalper with ultra-tight 0.5σ entries — half the threshold of the micro z-score scalper. K-nearest-neighbor ML classifies the next-bar direction from the 8-feature vector. Exits when z crosses 0 or profit locks at +0.04%.

▸Rolling 12-period z-score
▸Entry at ±0.5σ (vs ±1σ micro)
▸K-nearest-neighbor ML classifier (k=5)
▸Exit when z crosses 0 or +0.04% profit lock
▸Tight 0.22% stop, 0.07% target

Order Flow Nano Scalper

Detects micro buy/sell pressure from 3-candle volume-weighted body analysis. Ridge regression predicts next-bar direction from 8 micro-features. Enters on strong imbalance with ML confirmation. Profit locks at +0.05%.

▸3-candle volume-weighted body position analysis
▸Online ridge regression ML classifier
▸Entry when imbalance > 65% and ML aligned
▸Hyper profit-lock exit at +0.05% favorable
▸Tight 0.25% stop

Wick Sniper ML Scalper

Snipes candle wick rejections with ultra-fast response. Requires only a 1.5× wick-to-body ratio (vs 2× standard) for faster signal generation. Decision stump ensemble classifies the rejection. Profit locks at +0.04%.

▸1.5× wick-to-body ratio (vs 2× standard)
▸Decision stump ensemble ML classifier (3 stumps)
▸Entry on confirmed micro-rejection with ML
▸Hyper profit-lock exit at +0.04% favorable
▸Tight 0.22% stop

Micro Range Break Scalper

Detects ultra-tight 5-candle ranges and snipes the breakout. Logistic regression with a 60-bar training window confirms the breakout direction. Profit locks at +0.05% — exits the instant the breakout extends.

▸5-candle micro-range detection
▸8-feature logistic regression with 60-bar window
▸Entry on breakout with ML confirmation
▸Hyper profit-lock exit at +0.05% favorable
▸Tight 0.25% stop

Tick Velocity ML Scalper

Measures price velocity (rate of change acceleration) across 1-3 bar windows. An 8-feature perceptron with a 60-bar window classifies whether velocity will continue. Enters on strong velocity with ML confirmation. Profit locks at +0.04%.

▸1/2/3-bar velocity (acceleration) features
▸8-feature perceptron with 60-bar window
▸Entry when velocity > threshold and ML aligned
▸Hyper profit-lock exit at +0.04% favorable
▸Tight 0.22% stop

Micro Bollinger Squeeze Scalper

Detects Bollinger Band squeezes (band width contraction) and snipes the expansion. Gaussian Naive Bayes classifies the breakout direction. Enters on band touch with ML confirmation. Profit locks at +0.05%.

▸10-period Bollinger Band squeeze detection
▸Band width contraction ratio < 0.6
▸Gaussian Naive Bayes ML classifier
▸Entry on band touch with ML confirmation
▸Hyper profit-lock exit at +0.05% favorable

Hyper Mean Reversion Nano Scalper

Ultra-tight EMA mean reversion scalper. Uses a 5-period EMA as fair value and enters at just 0.08% deviation — nearly half the micro mean reversion scalper. K-nearest-neighbor (k=7) classifies the reversion. Exits on return to EMA or +0.04% profit lock.

▸5-period EMA as adaptive fair value
▸Entry at 0.08% deviation (vs 0.15% micro)
▸K-nearest-neighbor ML classifier (k=7, 50-bar)
▸Exit when price returns to EMA5
▸Hyper profit-lock exit at +0.04% favorable, tight 0.2% stop

Ready to put these algorithms to work?

Create an account and start running any of these strategies on live market data in minutes.

Get Started