How to Build a Trading Plan
To build a trading plan, fix one instrument and one timeframe, write the three prices for every trade, set the size from the stop, state the round-trip cost, and fix the sample size before you are allowed to judge the result. The plan is those numbers written down.
Most published plans are statements of intent. A plan that works is a short list of numbers you decided while calm, so that nothing has to be decided while you are not.
Before you start
One instrument you will actually trade. Not a watchlist. One.
Your round-trip cost in money — spread plus commission plus slippage, for that instrument, on the size you will use.
An account balance you can state, because two of the four numbers below are percentages of it.
Nothing here needs software you do not have. The whole plan fits on one page.
The steps
1. Fix one instrument and one timeframe
Write both down. Every number after this depends on them, and changing either invalidates the rest.
2. Write the three prices for every trade
Entry, stop, target — decided before the position exists. The entry and exit page is the long version of why the exit is the harder half.
3. Set the size from the stop
Money you will risk, divided by the distance to the stop. That gives the position, and whatever leverage it implies is the leverage you are using.
4. Write down the round trip
As a share of a typical bar, not as a fee. On this history 0.02 against a 1.17 bar is under 2% — and across 123 trades it still took 67% of the gross.
5. Fix the sample size before you start
A hundred trades, no changes, then review. Twenty-five leaves a win rate uncertain by about twenty percentage points either way.
6. Decide the worst run you will sit through
Five losses in a row is ordinary — one appears inside a record of 64 wins and 59 losses. At 2% risked per trade that leaves 90% of the account; at 20% it leaves 33%.
How to tell it worked
The test is not whether the plan sounds sensible. It is whether it answers a question you have not had yet.
Take four situations and read your own page for the answer.
Price is 2% from your stop and you still like the read — does the plan say move it? You have had three losses this week — does it say trade smaller? A setup appears on an instrument that is not your one — does it say skip it? You are 40 trades in and down — does it say review, or wait?
If any of those four sends you back to thinking, that part is not written yet.
And one arithmetic check. Multiply your round trip by the trades you expect in a year. If the answer is a large share of what you expect to make, the frequency is the problem and no other part of the plan matters until it changes.
The plan is four numbers
Everything above compresses to this, and it is worth having on one line.
Cost per trade. Stated, not estimated. Sample size. How many before you judge. Worst run. What you will sit through. Net. Gross minus costs, which is the only line that pays anybody.
They combine into one figure — expectancy. Win rate times the average win, minus loss rate times the average loss, minus the cost. On the worked record: 0.52 × 0.74 − 0.48 × 0.74 − 0.02 = +0.010 per trade, which across 123 trades is the +1.23 those trades actually made. The arithmetic closing is the test that the four numbers are the right four.
What to leave out
A plan gets abandoned because it is too long, not because it is too short. Four things are routinely written into one and should not be.
Market commentary. A view about where price is going is not a rule, and it expires.
Indicator settings you have not tested. They belong in the rule the backtest covers, not in the plan — the plan says how much and how often, not which oscillator.
Goals in money. “£500 a month” is not something you control, and a target in money is the most reliable way to make a plan produce trades that were not in it.
Anything you cannot check. If you cannot tell at a glance whether you followed it, it is not a rule you can follow.
What is left is short enough to read in the moment it matters, which is the only moment a plan has a job.
The original data
Across our study of 24,971 trading videos, 146 cover trading plans. The median one gets 9,100 views, 79% never pass 50,000, and the median length is 12.5 minutes.
The corpus carries description text for only 17 of those 146, which is too thin to say anything about how the topic is written, and this page does not.
The field size is the readable part. 146 videos on writing a plan, against 1,465 on day trading and 812 on backtesting — roughly one video about the plan for every ten about the activity.
When it fails
It is a philosophy, not a page of numbers
“Trade with discipline” is not a plan. If nothing on the page can be checked against a chart or an account balance, nothing on it will survive a bad fortnight.
The review trigger was missing
A plan without a stated review point gets reviewed after every loss, which is the selection problem the trading journal page describes. Fix the number of trades in advance and the argument disappears.
The size was set by the setup
Taking more because one looks better is how an ordinary losing run becomes permanent. The size comes from the stop, and the stop comes from the chart.
You wrote it and never opened it again
A plan is read at the moment it is least welcome. That is its whole function, and a plan you have not opened in a month is a document rather than a tool.
Related
How to backtest a strategy is where the rule inside the plan gets tested before it costs anything.
How to start trading is the shorter version for a first account.
And trading as a business turns these four numbers into one expectancy figure and shows the arithmetic closing.
My plan fits on one page and most of it is numbers rather than rules. The part I got wrong for years was leaving the review trigger out, so every bad fortnight became a live argument about whether to change something. Deciding that in advance is the whole difference.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.