Why is tracking the prediction_id important?

It serves as the unique key that correlates the initial entry signal with the final exit event, allowing the system to calculate the exact accuracy and error of a specific trade.

How does it handle directional noise?

It utilizes a 0.1% threshold to classify moves as up or down; anything smaller is classified as neutral to prevent tiny price fluctuations from skewing accuracy statistics.

What data is stored in the outcome record?

The skill records the actual direction (-1, 0, 1), the percentage magnitude change, the final exit price, and calculated fields like directional correctness and magnitude error.

What exit paths does this skill cover?

The pattern covers all three critical exit points: Stop-Loss/Take-Profit triggers, Signal-Based Flattening (when a signal becomes neutral), and Reverse Signal Closes.

Can this be used for both live trading and backtesting?

Yes, while specifically designed for live trading environments, the pattern is essential for backtesting to create comparable datasets between simulated and live performance.

Trading Prediction Outcome Tracker

Name: Trading Prediction Outcome Tracker
Author: smith6jt-cop

bysmith6jt-cop

Data Science & ML

Automates the recording of actual trading results against predicted signals to enable accurate performance metrics and model retraining.

About

This skill provides a standardized pattern for closing the prediction loop in algorithmic trading systems by capturing outcome data when positions exit. It ensures that every prediction—whether it ends in a stop-loss, take-profit, or manual signal reversal—is accurately mapped to its result with precise magnitude and direction calculations. By implementing this pattern, developers can generate reliable datasets for model evaluation, accuracy analysis, and automated retraining triggers, solving the common problem of 'orphaned' predictions that lack ground truth data.

Key Features

Captures outcomes across all exit points including stop-loss, take-profit, and signal reversals
0 GitHub stars
Standardizes magnitude recording as percentage-based change for consistent evaluation
Tracks prediction IDs through the entire position lifecycle to correlate entries and exits
Calculates directional accuracy using a configurable 0.1% movement threshold
Includes SQL verification queries to monitor outcome recording health and data integrity

Use Cases

Debugging trading strategy performance by correlating signals with actual exit prices
Measuring the win rate and directional accuracy of machine learning trading models
Building a verified ground truth dataset for model retraining and optimization

About

Key Features

Captures outcomes across all exit points including stop-loss, take-profit, and signal reversals
0 GitHub stars
Standardizes magnitude recording as percentage-based change for consistent evaluation
Tracks prediction IDs through the entire position lifecycle to correlate entries and exits
Calculates directional accuracy using a configurable 0.1% movement threshold
Includes SQL verification queries to monitor outcome recording health and data integrity

Use Cases

Debugging trading strategy performance by correlating signals with actual exit prices
Measuring the win rate and directional accuracy of machine learning trading models
Building a verified ground truth dataset for model retraining and optimization