Technical Analysis

Building a Trading System

The capstone: turn everything you've learned into one rule-based system you can actually run. Define your market and timeframe, your entry, exit and size, then backtest it, forward-test it, and learn when to trust the edge — and when it's broken.

0 / 11 done

StartLesson 0 · System toolkit smoke test (dev only)

What you’ll leave with11 lessons · 10-line playbook

A system is five written rules — market, entry, exit, size, review — with a positive expectancy you repeat across many trades. The edge does the work, not any one trade.
Trade the timeframe you can actually watch, in liquid markets that trend in clean steps — and only enter when the higher timeframe's trend and the lower timeframe's signal point the same way.
A real entry is three things at once — a trigger, a condition (with the trend), and a location (at a level or the rising average) — and it only fires when all four written rules are true together.
The exit is half the system: stop just below the pullback swing low (where the setup is wrong), first target at +2R or the prior swing high, then trail under each new higher-low once +1R is banked — and never move a stop away from price.
Confluence is independent signals pointing the same way: one is noise, three or more is conviction — and correlated signals don't double-count. Score the present factors and let the tier (skip / watch / take) make the call.
Backtesting applies your written rules to past bars until you have a number: expectancy = (win rate × avg win) − (loss rate × avg loss). Log every signal — winners and losers, left to right — over a big enough sample, and read the equity curve. A backtest is a hypothesis, not a promise.
A backtest is a hypothesis; forward-testing proves it. Paper-trade the rules live, journal every trade the same way (entry, stop, target, exit, R), and after 20 trades compare live expectancy to the backtest. Live adds slippage, missed fills and emotion — the journal is how you catch them.
Optimizing sweeps a parameter and reads the performance; overfitting tunes to the noise. Robustness is a broad PLATEAU where the result barely changes — trust that, never the lone SPIKE. Prove it out-of-sample, and remember simpler beats complex: fewer parameters, fewer ways to fool yourself.
Every edge has losing runs. A drawdown is only a broken edge if it's DEEPER or a losing streak LONGER than your backtest ever produced, or the regime has changed (trending → range). On a normal dip, continue; when the tells flash, reduce and pause; halt and rebuild only when the numbers say the edge is gone. Review every 20 trades.
A complete trading system is five components run as one loop: a market & timeframe you can trade, written entry rules, a confluence filter, exit & sizing rules, and a review. Run it the same way every time — market → entry → confluence → exit → log → expectancy — and let the edge do the work across many trades. That's the whole craft.

Building a Trading System

All lessons