Building a Trading Plan

Building a Trading Plan

A trading plan is the written constitution of your account. It converts vague intentions like “trade carefully” into testable rules that survive contact with markets — and survive you, the trader, on the worst day of your year. This lesson walks through the seven sections every plan needs and how to test them before live capital is at stake.

8.6.1Why a written plan beats a brilliant trader

Imagine two traders. The first reads voraciously, has memorised every Mark Minervini setup, watches CNBC at the open and follows 14 finance accounts on X. The second has a single laminated sheet beside her keyboard listing four entry conditions, three exits, a position-size formula and a stop-loss rule. Five years from now, the laminated-sheet trader is almost certainly ahead. Plans win because markets reward consistency, not cleverness.

The brain on a live P&L is not the brain that reads books. Cortisol narrows attention; dopamine biases toward action; loss aversion turns small losers into account-killers. A written plan is a pre-commitment device — Ulysses tying himself to the mast — that lets a calm version of you make decisions binding on a panicked version of you four hours later.

Empirically, Barber and Odean’s study of over 66,000 retail brokerage accounts (2000; summarized in their 2013 review) found the most active quintile of traders underperformed the market by roughly 6.5 percentage points per year net of costs, while the least active trailed by under 1 point. The gap was driven largely by trading costs and overconfident overtrading — exactly the behaviors a written plan constrains. A plan is the cheapest way to buy that discipline.

If you cannot define your edge in two paragraphs, you do not have one — you have a hobby with leverage.

8.6.2The seven mandatory sections

A complete plan answers seven questions. Skip one and the others wobble. Here is the canonical structure used across this Academy:

8.6.2.1Mission first, methodology later

The mission line forces honesty. “I want to compound my $40,000 brokerage account at 18% annualised over five years while keeping max drawdown under 15%” is a mission. “Make money trading” is not. The numbers in the mission feed everything downstream — they tell you whether you should be daytrading SPY or holding NVDA for six months.

8.6.2.2Why kill switches go last but matter most

The seventh section — kill switches — is the one most amateurs omit and most professionals consider non-negotiable. A kill switch is a pre-defined trigger that takes you out of the market: 3 consecutive losing days, an account drawdown of 8% from peak, or any single trade that violates risk rules. When triggered, you stop. You don’t argue with the rule, you stop.

8.6.3Defining your edge in writing

An edge is a measurable, repeatable expectancy advantage. It is not a feeling. It is not “I’m good at reading charts.” A defensible edge statement reads like a hypothesis a junior analyst could test:

“On large-cap US equities with a 50-day average dollar volume above $200M, breakouts above a 20-day high with relative volume >1.5x have produced an average 10-day forward return of +2.1% versus +0.4% for the benchmark across the 2015–2024 sample.”

That sentence is testable. You can pull the data, replay the trades, and either confirm or kill the idea before committing capital. Your written edge thesis must contain four things:

• Universe — the population of instruments where the edge applies (e.g. S&P 500 components, not penny stocks).
• Trigger — the precise condition that fires the signal (price, volume, time of day).
• Holding period — how long you typically remain in the trade.
• Expected statistical profile — win rate, average win, average loss, expectancy per dollar risked.

8.6.3.1Worked example: a momentum continuation edge

Consider a mid-cap pullback strategy. The plan describes it as: enter when a stock in the Russell 1000 with a 200-day rising slope pulls back to its 21-EMA on declining volume, then closes back above the prior day’s high. Stop is the swing low; target is 2R. Backtest on 2015–2024 shows a 47% win rate, average win of +8.2%, average loss of −4.1%, expectancy of +0.41R per trade (0.47 × 2R − 0.53 × 1R = +0.41R).

That single paragraph is now the foundation of the rest of the plan. Every entry, exit and risk choice traces back to it. If the live results drift more than two standard deviations from the backtest profile over 50 trades, the plan calls for a forced review — not a silent override.

8.6.4Risk and position sizing rules

This is where most plans collapse under their own weight. Keep it boring and explicit. The risk framework needs four numbers and one formula.

8.6.4.1The four numbers

• Risk per trade (R) — typical retail target is 0.5% to 1.0% of equity. On a $50,000 account at 0.75%, that is $375 per trade.
• Max open risk — sum of all live stop distances. Cap typically at 3R to 5R.
• Daily loss limit — walk away when realised + open losses hit, e.g., 2R in one session.
• Drawdown kill switch — hard stop on the strategy at, say, −8% from equity peak. Below this you go to paper-trading until rules pass review.

8.6.4.2The position-size formula

Position size in shares = (Account equity × Risk%) ÷ (Entry − Stop). A worked example: account $50,000, risk 0.75% = $375. NVDA setup at $118.40 entry, $114.10 stop. Per-share risk is $4.30. Shares = 375 ÷ 4.30 ≈ 87 shares. Capital deployed: $10,300, but risk deployed: $375. The plan binds you to that calculation; “feels right” does not.

8.6.5Entry, exit and management protocols

Once edge and risk are nailed, the execution section becomes a checklist rather than a strategy. The goal is a procedure so explicit that an honest stranger could trade your plan from your screen.

8.6.5.1The pre-entry checklist

• Setup matches one of the named patterns in §4 of the plan.
• Liquidity filter passes (e.g. $200M ADV, spread under 5 bps).
• No earnings within the holding window, unless strategy is explicitly an earnings play.
• Position size computed and within max-open-risk cap.
• Stop and first target written down before the order is sent.

8.6.5.2Exit triage

Exits split into three categories: stop hit (loss taken, no negotiation), target hit (full or partial scale), and thesis broken (the why-I-bought is no longer true even if price has not moved). The thesis-broken exit is what separates good traders from ones with great-looking entries. SPY pulling back from $723 on benign volume is not a thesis break; SPY losing the 200-day with breadth collapsing and the VIX printing 17.32 after a hot CPI is.

8.6.6The trading journal as feedback loop

A plan without a journal is a will without an executor. The journal closes the loop between intent and outcome and is where almost all genuine improvement happens. Minimum required fields per trade:

Review weekly. The single most useful number is plan-adherence rate: of your last 20 trades, how many followed the documented setup, sizing, and exit rules without intervention? Below 85%, the strategy’s stats are statistical noise — you are not running the system you wrote down.

8.6.7Common mistakes and how to fix them

Five recurring failure modes show up in plans submitted by working traders. Catching these early saves quarters of wasted effort.

8.6.7.1The plan that’s never tested

Symptom: the plan exists, but no backtest, no walk-forward, no paper period. Fix: minimum 100 historical trades on the universe, then 30 paper trades, then quarter-size live for 30 trades, then full size. Anything else is donating to the market.

8.6.7.2Setup catalogue too broad

Symptom: ten patterns named, none of which you can describe without ambiguity. Fix: cut to three. Real edges are narrow. A pullback-to-21EMA setup beats a vague “momentum trend.”

8.6.7.3Risk numbers that ignore correlation

Symptom: the plan caps per-trade risk at 1% but allows 5 open semiconductor names. Fix: a sector cap (max 2R per GICS sector) and a beta-weighted aggregate cap. Five long NVDA-correlated names is one trade, not five.

8.6.7.4No defined review cadence

Symptom: “I’ll review when I have time.” Fix: a fixed Sunday session, calendar-blocked. 45 minutes is enough. Skip three Sundays in a row and the kill switch fires.

8.6.7.5Plan inflation after wins

Symptom: a hot streak triggers an unscheduled “plan upgrade” — bigger size, looser stops, new sectors. Fix: changes only at scheduled monthly review and only with documented data justification. Wins are not data; they are samples.