TradeBot

AI-Powered Trading
Loading... demo
Authentication required
Auth required
Account Equity
Open Positions
0
Total Trades
0

Broker Connection

Loading...

Quick Status

Loading...
EMA
Volume
Forecast
Loading strategies...

Open Positions

Symbol Side Units Entry Stop Loss Take Profit Trailing P&L
No open positions

Front-Month Contracts

Loading...

Instrument Specifications

Symbol Description Exchange Tick Size Tick Value Point Value Class Months
Loading instruments...

Position Sizing Calculator

0 = use account equity
Models Trained
Symbols Active
Total Predictions
Avg Accuracy
Avg Profit Factor
Models Valid
Avg F1 Score
Top Performer

Live Performance Dashboard

Active
Loading AI metrics...

Model Validation & Health

Loading AI models...
Updated —
Loading forecast...

Narrative Summary

Awaiting forecast data.

Trading Configuration

Data & Lookback

Adjust for scalping (lower) or swing trading (higher)
Candles for signal analysis (Scalping: 100-200, Swing: 300-500)
Historical candles for AI training (Scalping: 500-1000, Swing: 1500-3000)

Risk Management

Maximum allowed drawdown before kill switch triggers
Maximum position size cap (0 = no cap)
Minimum ATR to allow trades (0 = disabled)
Kill Switch
Stop trading if drawdown exceeds max

Exit Parameters

Stop loss distance (ATR x multiplier)
Take profit as multiple of SL distance
Trailing stop distance (ATR x multiplier)
Partial Exit
Portion to exit early (0.5 = 50%)
Early TP as multiple of SL distance

Position Sizing

Percent-Based Sizing
Use % of equity instead of fixed units
Percent of equity per trade (0.10 - 0.20)
Fixed units (when % sizing disabled)
Volatility Sizing
Adjust position size based on volatility
Size Bounds

AI Configuration

AI Filter
Use AI probability to filter entries
Enhanced AI Model
Use Optuna optimization, calibration, regime detection
Require AI Prediction
Block trades on symbols with no trained AI model
AI Prediction Logging
Log AI predictions to file for analysis (A/B test mode)
Prophet Forecasting
Use Facebook Prophet for price forecasting
Prefer Prophet
Prioritize Prophet over sklearn models
Auto Thresholding
Use validation folds to set AI thresholds
Choose how validation selects thresholds
Sliding window size for training (0 = full lookback)
Higher = fewer but higher-quality signals (0 = disabled)

Enhanced AI Settings

Advanced AI validation and signal filtering
Model Validation
Models older than this are considered stale (168 = 1 week)
Z-score for drift detection (higher = less sensitive)
Minimum live accuracy before warning
Enforce Min Accuracy
Reject models below minimum accuracy
Choose how drift is detected
Recent samples used for drift checks
Only used when drift method is PSI
AI Calibration
Calibrate probabilities for AI classifier
Forecast Calibration
Calibrate forecast classifier probabilities
Feature Selection
Auto Feature Selection
Automatically select best features (reduces overfitting)
Maximum features to select (0 = use all)
Choose how features are ranked when auto-selection is enabled
Signal Filtering
Minimum signal strength (0 = disabled)
Minimum expected price move (0 = disabled)
Consensus Filter
Require multi-timeframe agreement
Minimum timeframe alignment score (0-1)

Deep Learning Models

LSTM/Transformer model configuration
Deep Models Enabled
Include LSTM/Transformer models during training
Extended Features
Use extended feature set for AI training
Input sequence length for LSTM/Transformer
Hidden layer size for deep models
Training epochs for deep models
Optimizer learning rate

AI Validation & Staleness

Model retraining, sample weighting, and stale model handling
Sample Weighting
Weight recent samples higher during training
Higher = more weight on recent samples
Auto-Retrain on Stale
Automatically retrain stale models
Minimum wait between retrain attempts
Block on Stale
Block trades when model is stale
Minimum samples needed for validation
Sample size for drift detection
Position scale-down on validation warning

Shadow Training & Promotion

Shadow model training and automatic model promotion
Shadow Training
Train shadow models alongside primary
How shadow models are promoted to primary
Min predictions before promotion eligible
Shadow must beat primary by this amount
Minimum shadow accuracy for promotion

Market Data Quality

Gap tolerance and degraded data handling
Max consecutive missing candles to tolerate
Max missing-candle ratio allowed
Max ratio for degraded data acceptance
Allow Degraded on Gap
Use degraded data when gap detected
Last-Resort Degraded
Accept degraded data as last resort

Scheduler

Automatic AI training and forecast intervals
How often to auto-retrain AI models (1440 = daily)
How often to auto-train forecast models

Reconciliation & Risk Limits

Position sync, circuit breakers, and portfolio exposure
Auto-Fix Mismatches
Auto-reconcile position mismatches with broker
How often to check positions vs broker
Maximum total exposure as multiple of equity
Consecutive sync failures before halting

Strategy Selection

Strategies that must agree (consensus mode only)

Session Filter

Session Filter
Only trade during active sessions
News Filter
Skip signals around major news events
Minutes before news to stop trading
Minutes after news to resume trading
Comma-separated: london, newyork, tokyo, sydney
Session Times (UTC - 24-hour format)

Enhanced Trading Filters

Multi-timeframe confluence, S/R detection, signal scoring
Multi-Timeframe Confluence
Confluence Filter
Analyze H4, H1, M15 for signal alignment
Minimum score to allow trade (0-100)
Weights should sum to 100
Support/Resistance Detection
S/R Filter
Prevent trades near S/R levels
Periods for swing high/low detection
Touches to confirm level
Min distance to S/R in ATR units
Signal Quality Scoring
Signal Scoring
Score signals on multiple factors
Minimum score to allow trade (0-100)
Session Quality
Minimum quality score (0-1)
Require Overlap
Only trade during session overlaps

Set up authentication

Capital.com Account

DEMO Not connected

Enter your Capital.com credentials to trade. Each user has their own separate account. The bot will automatically switch to the correct server (demo or live) based on your selection.

Get your API key from Capital.com platform settings
Leave empty for auto-detection (demo/live)

NinjaTrader Gateway

SIM Not connected

NinjaTrader connects locally via the ATI Gateway — no credentials required. Click Auto Connect to connect with defaults (localhost:36973, Simulation).

Not connected
Advanced: Gateway Settings
NinjaTrader Gateway host address
Default NinjaTrader ATI port (36973)
NinjaTrader account name (leave blank for auto-detect)
Available Futures Instruments
Loading instruments...

Account

Not logged in
No events yet