A trading plan is a written set of rules for choosing trades, controlling exposure, placing orders and reviewing results. A strategy describes the opportunity you want to trade. The plan also decides whether you should take it, how much you can put at risk and when to stop.
For active trading, build the rules before testing them, test them against realistic costs, then check whether you can follow them as prices unfold. Treat profitability as a question to investigate, not an assumption. A clear plan makes decisions testable; it does not make a losing strategy profitable.
Define the purpose and boundaries
Start with the conditions under which you can trade. Record your available capital, markets, permitted instruments, trading hours and maximum holding period. Keep money needed for living expenses or emergencies outside the trading budget.
Match the plan to your schedule. If you cannot monitor positions during the session, do not build a method that requires immediate decisions every few minutes. Include time for preparation and review, not just order placement.
Choose process objectives rather than compulsory income targets. “Record every qualifying setup and follow the exit rule” gives you something to assess. “Make $200 every day” says nothing about which trades deserve your money.
Set separate boundaries for individual trades and total account exposure. Your trade-plan risk parameters should address the planned loss per trade, simultaneous positions and the loss threshold that ends a session. Write the required action beside each threshold so there is no argument with yourself later.
Turn a trading idea into repeatable rules
A testable rule lets you decide whether a trade qualifies using only information available at that moment. “Buy a strong breakout” leaves too much unresolved. What counts as strong? Must the price close above the level? How long does the signal remain valid?
Use the following trading plan template as a worksheet. Complete each field before collecting results.
| Plan field | Decision to record |
|---|---|
| Market and schedule | Eligible instruments, chart interval, trading window and time zone. |
| Setup | The observable conditions that must exist before an entry is considered. |
| Entry | The trigger, order type, order expiry and conditions for canceling an unfilled order. |
| Exit | Initial stop, profit exit, time exit and any permitted adjustments. |
| Position size | The sizing calculation, cash requirement and total exposure cap. |
| Exclusions | Conditions that prohibit trading, such as excessive spreads or missing data. |
| Loss controls | Session and account loss thresholds, including what happens to open positions. |
| Validation and review | Test periods, cost assumptions, acceptance criteria and review dates. |
For a rule-writing example, replace “buy above resistance” with “after a five-minute bar closes above the previous session’s high, submit a market buy order at the next bar’s opening.” That is an entry rule to investigate, not a recommended strategy. It still needs eligibility filters, sizing and exits.
Specify whether positions must close before the session ends. If holding beyond the close is permitted, add separate rules for overnight and weekend risk, including how to handle scheduled events.
Calculate position size before placing an order
Start with the planned exit and an affordable loss budget, then calculate the position. Do not choose a large position first and squeeze the stop closer simply to make the arithmetic fit.
For a simplified share trade:
Position size = trade risk budget ÷ (entry-to-stop distance per share + estimated costs per share).
Suppose a hypothetical account contains $20,000 and the chosen trade budget is 0.25%, or $50. With a $0.50 stop distance and a $0.05 per-share allowance for round-trip costs and slippage, the calculation is $50 ÷ $0.55. Rounding down gives 90 shares and an estimated loss of $49.50 if those assumptions hold.
These figures demonstrate the calculation, not a suitable risk percentage for every trader. Check the cash required as well. Futures and forex positions need the appropriate contract or pip value and, where relevant, currency conversion.
Distinguish planned risk from guaranteed maximum loss. A stock stop order becomes a market order when triggered, so the execution price can differ from the stop price. Build that uncertainty into the plan rather than assuming perfect exits; the SEC’s order-type guidance sets out that distinction.
Also define how several positions share the account risk budget. Include open losses and costs in session controls, rather than counting closed trades alone.
Backtest the rules with realistic data and costs
Backtesting applies the written rules to historical data. Before starting, save a dated version of the plan and record the data provider, date range, session settings and cost model. Audit a small sample manually, even if software runs the full test.
For manual testing, advance through the chart without seeing future bars. Log every qualifying signal, including losing trades and unfilled orders. Selecting only attractive chart examples is not a useful test.
Check that the data represents what could have been known and traded at the time. Using information published later creates look-ahead bias. Testing a stock-selection method only on companies that survive today creates survivorship bias. These are central checks in CFA Institute’s backtesting framework.
Model commissions, spreads and slippage. Include financing, borrowing or other holding costs where they apply. If simulated execution prices already include the spread, do not subtract it again. Keep data and software subscriptions in a separate business-cost calculation.
Write down how the test handles uncertain execution. If a bar reaches both the stop and target, its high and low alone do not reveal which came first. Use more detailed data or apply a declared conservative assumption consistently. Do not award yourself the profitable outcome.
Run a second calculation with worse execution assumptions. As an illustrative stress test, increase estimated trading costs by 50%. Check whether the result remains useful, and whether that stress is severe enough for the market and order size being tested.
Keep unseen data separate from development
Separate the history used to develop the rules from the history used to evaluate them. The development sample is where you resolve definitions and compare initial ideas. The evaluation sample should remain untouched until the rules are frozen.
A simple arrangement uses an earlier period for development and a later period for evaluation. For a longer study, consider rolling windows: develop using an earlier window, test on the next window, then move forward. Keep future observations out of each earlier decision.
Testing many indicator combinations and reporting only the winner can produce impressive results without genuine predictive value. NBER research on strategies built from multiple signals examines this overfitting problem. Keep a record of rejected variations, not just the successful version.
If you change the rules after seeing evaluation results, that period has become part of development. Do not continue calling it an independent test. Reserve fresh data or wait for new observations.
Check nearby settings too. If a method works with a 20-bar lookback but collapses at 19 or 21, investigate its dependence on that choice. The aim is not to find the prettiest historical chart; it is to test whether the reasoning survives reasonable changes.
Set acceptance criteria beyond total profit
Write the acceptance criteria before examining the final results. A proposed screening rule might require positive results after costs on untouched data, losses within a declared budget and no dependence on one exceptional trade. Passing these checks supports further investigation, not a promise of future returns.
Measure average net outcome per trade, maximum drawdown, losing streaks and how long losses take to recover. Drawdown measures the decline from an equity peak to a subsequent trough. Include open-position values where possible; closed-trade results alone can conceal uncomfortable exposure.
For comparisons, define one R as the initial entry-to-stop dollar risk before costs. A hypothetical test with 45% winning trades, an average gross win of 1.8R, an average gross loss of 1R and average costs of 0.1R per trade gives:
Estimated net expectancy = (0.45 × 1.8R) − (0.55 × 1R) − 0.1R = 0.16R per trade.
That is a sample average, not an amount the next trade should earn. If win and loss figures already include costs, do not subtract costs twice.
Keep the calculations consistent with your trading journal and performance measurements. Break results into periods and market conditions to check where gains and losses occurred.
Do not treat 30 or 100 trades as an automatic pass. Consider the date span, dependence between trades and range of conditions covered. Many trades during one market move offer less varied evidence than the trade count suggests. Treat the worst historical drawdown as an observation, not a ceiling.
Forward test before considering real money
Forward testing means applying frozen rules as new prices arrive. Use a simulation account or a timestamped paper record, and follow the same schedule, sizing method and order instructions that the live plan would require.
Record missed signals, delayed entries, rejected orders and occasions when the rules were ambiguous. Keep these separate from the theoretical strategy result. You need to know both whether the method has promise and whether it can be executed as written.
Simulation cannot fully reproduce financial pressure or actual execution. Hypothetical results may misrepresent liquidity and slippage, and they cannot fully capture a trader’s response to real losses. The NFA notice on hypothetical performance identifies these gaps.
Set the forward-test review point in advance using both elapsed time and qualifying opportunities. Do not end the test early because the first few trades look good.
If the evidence supports proceeding and losses are affordable, consider the smallest practical live exposure that fits the plan. Compare actual fills and costs with the assumptions. If the minimum tradable size exceeds your budget, remain in simulation or reject that instrument. Real-money trading is not a compulsory graduation ceremony.
Review the plan without rewriting it after every loss
Give every plan version a date and keep results attached to that version. Review execution regularly, but reserve strategy changes for scheduled assessments unless a safety issue requires an immediate pause.
Separate three problems: a losing trade that followed the rules, a trade that broke the rules and evidence that the method no longer meets its acceptance criteria. Each calls for a different response.
Predefine suspension triggers, such as the account loss threshold, repeated execution failures or persistent rule breaches. State how open positions will be handled and what must happen before restarting. This gives practical force to your controls against overtrading and loss-chasing.
When changing a rule, record the reason and test the revised version separately. Keep the operating document short enough to use during a session, with research stored alongside it. Before the next order, it should answer three questions plainly: does this trade qualify, what can go wrong and what action follows?