Learning path / Trading methodology
Planning a Trading Method
Learn how to define trading rules, document assumptions, evaluate costs and risk, and test whether a strategy idea is supported by evidence.
A repeatable process begins with clear rules. A chart pattern, a confident prediction or a successful example trade is not yet a method. This path keeps the ten existing lessons, the browser marks, and the hypothetical rulebook.
A method is a documented set of rules and assumptions. It is not proof of an edge, a recommendation to trade, or a promise of future results. Finishing the pathway is not evidence of trading competence.
- Research hypothesis
- Define rules
- Model execution and risk
- Test against data
- Review evidence
- Revise or reject
Your methodology learning journey
Progress loads in this browser. Progress is stored in this browser where supported. It is not synchronized to an account and does not represent a qualification.
Stored marks for other paths: …. Completing a step here does not mean the method has an edge.
Ten steps, in a useful order
The order is a recommendation. Any step can be opened. A later step is easier once the earlier mechanism is clear. Marking a step complete is optional, and it records a click, not an understanding.
- 1
What a strategy has to specify
Name the market, the condition, the size and the exit before any result exists.
Not started. Opening the lesson does not mark it.
- 2
Technical analysis
Read a chart as a record of past prices, and keep that record separate from a rule.
Not started. Opening the lesson does not mark it.
- 3
Trend, momentum and a written plan
A description of past movement still needs a size and an exit.
Not started. Opening the lesson does not mark it.
- 4
A price area and a live bid
A prior high or low is history. A bid is a price that might actually be traded.
Not started. Opening the lesson does not mark it.
- 5
Fundamental analysis
Filings, earnings and the balance sheet answer a different question from a chart.
Not started. Opening the lesson does not mark it.
- 6
Order execution
An instruction and a fill are different events.
Not started. Opening the lesson does not mark it.
- 7
Slippage and liquidity
The price in a plan and the price that trades can differ, and the difference is a cost.
Not started. Opening the lesson does not mark it.
- 8
Write the rules
Write the hypothetical rules on this page, before looking for a result.
Not started. Opening the lesson does not mark it.
- 9
Testing and the limits of a test
A test describes one history under the assumptions you stated.
Not started. Opening the lesson does not mark it.
- 10
When the rule is changed
Changing a rule after the result is a new decision, not a correction of the old one.
Not started. Opening the lesson does not mark it.
Before you begin: understand the foundations
Foundations and a risk budget. Chart patterns are not a prerequisite for the rulebook. Chart patterns are not a substitute for a risk budget, funding, costs, or the legal terms of an account.
Broker
The firm on the account contract. Eligibility, custody and complaints follow that legal entity and your country.
Trading software
The screen used to look at markets or send an instruction. It can be separate from the firm. The directory is at trading software.
Execution venue
The exchange, trading facility or dealer arrangement for that product. A contract with a firm is not itself an exchange. Over-the-counter products do not share a public book with listed shares.
One firm can hold more than one of these roles. This site does not stream prices and does not rank firms.
From an idea to a testable hypothesis
An observation is something the data showed. An opinion is a view about what happens next. A hypothesis names a condition you could mark true or false. A complete rule adds the order, the exit, the invalidation and the cost. A test result is what that rule did on one history. Evidence of a persistent edge would have to survive other histories, other costs and the ways a test can flatter itself. This page does not supply that evidence.
Vague statement
“Buy when the market looks strong.” Two people can disagree on every bar, so there is nothing to test.
Defined hypothesis
Under a named market, session and indicator condition, evaluate a written entry, a written exit and a written risk assumption on a named history. That can be challenged. It is not a recommendation.
What a written method may have to include
- Instrument or market. Name the thing the rule refers to.
- Data source and session. Say which prices, and which hours, the rule is allowed to see.
- Timeframe. A daily bar and an hourly bar are different objects.
- Entry. The condition that makes an order eligible.
- Exit and invalidation. How the idea ends, and what would show it was wrong.
- Order type. Market, limit, or stop, and what that instruction does not promise.
- Size and risk. The assumption about exposure. A line on a chart is not a size.
- Costs. Spread, commission, slippage, and financing where the product charges it.
- Evaluation. What you will count, and what would make you reject the test.
- Version rule. How a change is recorded so the old test can still be read.
Build a hypothetical trading rulebook
Document your assumptions before examining a result. A written rulebook makes a strategy easier to challenge, test and revise.
Hypothetical research documentation — not investment advice, a trade recommendation or proof of profitability. Do not type account numbers, passwords or API keys. This form does not connect to a broker.
A method must survive the realities of execution
The spread, the commission, slippage, the size the book can take, a partial fill, a gap, and an order that expires all sit between a planned price and a result. A stop submits an instruction. It does not reserve the trigger price. Leverage and financing, where the product charges them, change the account result again. The order types page explains the instruction itself: market, limit and stop orders.
A long sketch: result equals exit minus entry, then minus the spread, commission and slippage once each. These are not live prices. A historical mid price is not a fill.
- Planned price difference
- 2.00
- Difference at the illustrative fills
- 1.50
- Costs, counted once
- 0.15
- Net
- 1.35
Understanding risk budgets and position exposure
Account exposure is how much of the account the idea uses. Notional value is price times size times multiplier. Margin is collateral the firm asks for, and it is not a cap on loss. A planned loss is the cash you assigned to the distance between entry and invalidation. The actual loss can be larger if the market gaps or the fill is worse. Volatility describes how wide recent moves have been. None of these is a personal recommendation.
Hypothetical defaults. Planned cash risk is capital times the percent. Theoretical size is that cash divided by the price distance times the multiplier. Whole units round down. Fees are excluded. This is not a position you are being told to hold.
- Planned risk amount
- 100.00
- Price distance
- 5.00
- Theoretical size
- 20.0000
- Whole units
- 20
- Notional at entry
- 2000.00
A stop trigger is not a maximum loss. Markets gap, and the fill can be worse. Leverage changes how large that difference is in account terms. The theoretical size may not be tradable or permitted.
The same arithmetic, with more fields, is on the position size calculator and the risk and reward calculator.
How to evaluate a strategy without hindsight
Write the rule, then freeze it, then gather the data, then test, then include costs, then read the result, then try to break it, then revise or reject. Defining the rule after the outcome is how a test flatters its author. Look-ahead bias uses a fact you could not have known. Survivorship bias keeps the names that still exist. Overfitting tunes the rule to one sample. A later sample, a walk-forward split, and a benchmark are ways to argue with the result. Liquidity asks whether the size in the test could have traded. Uncertainty remains after all of that.
- Define
- Freeze the rules
- Gather data
- Test
- Include costs
- Evaluate
- Challenge
- Revise or reject
Counts worth writing down include the number of trades, the share that gained, the average gain and average loss, the gross and the net, the drawdown, the variation of results, and how long the idea was exposed. No one of those figures proves a method works.
Does the idea survive its costs?
Each cost is the amount you assign to one completed trade. Put both sides in the figure if that is how you count, and do not enter the spread again as commission. A method that trades often pays the sum often.
- Gross
- 20.00
- Total costs
- 6.00
- Net
- 14.00
- Break-even gross per trade
- 0.30
A useful method explains when it is wrong
A losing trade can sit inside a rule that is still the rule you wrote. A short run of gains can sit inside a rule that was never specified. Separate a run of outcomes from a broken assumption, a bad data set, an execution constraint, a market that has changed its structure, and ordinary statistical noise. This page does not offer a universal profit or loss threshold.
- What does the hypothesis assume? Market, session, data and the condition.
- What would weaken it? Costs that consume the gross difference, or a sample that is mostly one regime.
- What would contradict it? An observation you wrote in advance as invalidation.
- What data could be missing? Names that delisted, revised filings, or a venue whose volume you cannot see.
- When is the method reviewed? On a date or a version rule, not after every outcome.
- When is a test stopped? When the frozen rule cannot be applied, or when you would be changing it to fit the result.
Document decisions before the outcome
A useful entry has a version, a date, the hypothesis, the entry and the exit, the cost assumption, and space for the outcome and any deviation. Fill the outcome after it exists. The template can be copied, printed, or saved in this browser.
This note stays in this browser if you save it. It is not a live trade blotter and it is not uploaded.
Learn, then calculate, then review
- Learn. Read the step. The mechanism is in that guide.
- Calculate. Type hypothetical inputs in the labs on this page, or in the profit and loss calculator. Nothing here sends an order.
- Review. Put the written rule next to the result. A match with the hypothesis is still one history.
Knowledge check
Mistakes that weaken a trading method
Vague entries
“Looks strong” cannot be applied the same way twice. A test needs a condition another person could mark true or false.
Rules rewritten after the result
A rule that changes once the outcome is known was not the rule that was tested.
Costs left out
Spread, commission and slippage sit between a price difference and a result.
Parameters fitted to one path
Changing a lookback until the picture improves is a fitted story, not a found rule.
Too few observations
A handful of trades can be a run of luck. The count belongs in the note.
One friendly regime
A rule examined only in a rising market has not been examined in the other ones.
An indicator treated as an edge
An indicator is arithmetic on past prices. The edge, if any, is the rule after costs, and it still may not persist.
Liquidity ignored
A rule that needs size the book cannot take is a different rule from the one on paper.
A stop treated as a cap
A stop submits an order. A gap can fill it worse, or not at all.
Gross mixed with net
The price difference and the result after costs are different numbers. Report the one you mean.
Uneven risk in the comparison
Two results are hard to compare when one used more of the account than the other.
A history treated as the future
A path that already happened does not set the next one.
Before considering a strategy hypothesis test complete
Questions worth answering slowly
What is a trading method?
A documented set of rules and assumptions: which market, which condition, which order, which exit, which risk, and which costs. It is not proof of an edge, a recommendation to trade, or a promise about future results.
What makes a trading strategy testable?
Someone else can apply the same rules to the same data and get the same marks. A mood, a chart that “looks strong”, or a list of winning examples chosen afterwards cannot be tested in that way.
Is a trading strategy the same as a trading plan?
The words overlap. On this page a method is the rule. A plan also has to say how large the position is, what would invalidate it, and what the costs are assumed to be. A slogan is neither.
Do you need technical analysis to define a method?
No. A method can be written from filings, from a calendar, or from a rebalancing rule. A chart can be an input. It does not replace the risk, the funding, or the legal account the order would sit in.
What should entry and exit rules include?
The condition that makes an order eligible, the order type, the condition that closes it, and the condition that means you do not send the order. Leave those blank and the later test is a story about whatever you decided in the moment.
What is an invalidation condition?
The observation that would show the hypothesis, as you wrote it, no longer fits. It is not a universal loss limit. A single losing trade can be inside a valid rule, and a short winning run can be inside a rule that is still poorly specified.
Why are transaction costs important?
Each trade can pay a spread, a commission and some slippage, and some products add financing. A small gross difference, repeated often, can become a net loss. The cost lab on this page is arithmetic on numbers you type.
What is backtesting?
Applying a rule that was written in advance to past data, including costs, and reading what that rule would have done on that path. It describes that path. It does not set the next one.
What is look-ahead bias?
Using a fact in the test that you could not have known when the decision would have been made. A close that had not printed yet, a revised filing, or a constituent list from the future are typical forms.
What is overfitting?
Adjusting the rule until it fits the sample you already saw. The fit can be tight and still say nothing about a later sample. Freezing the rule before you look at the held-out data is the ordinary defence.
Why is out-of-sample testing useful?
It asks the frozen rule to face data that was not used to invent it. Agreement with that data is still one history. Disagreement is a reason to doubt the rule, not a reason to keep tuning until the disagreement goes away.
What is a trading journal?
A record written so the plan and the later outcome can be compared. The template on this page stays in this browser if you save it. It is not a brokerage blotter and it is not uploaded.
How should a strategy be revised?
Give the change a version, write why it changed, and treat the new version as a new hypothesis. Editing the old version after you know the result makes the test impossible to read fairly.
Does a profitable backtest mean a strategy will work?
No. The test can omit costs, use a friendly period, fit the parameters to that period, or assume fills the market did not offer. A flattering history is a reason to inspect the assumptions, not a forecast.
Does completing the learning path mean I am ready to trade?
No. Marks in this browser record that you pressed them. They are not a qualification, a licence, or evidence that a method has an edge.
Sources and review
Written by: WiT Markets Editorial. Writes and structures WiT Markets pages. Separates a sourced fact from a cell that is Not confirmed.
Review: Editorially reviewed. This is not a qualified specialist review and it is not a legal sign-off.
Last reviewed: . Next review due . The date is an editorial pass. It is not a live check of every third-party fee.
Regulatory and official sources
- Investor.gov — checked 13 October 2026
- FINRA — checked 13 October 2026
- U.S. Securities and Exchange Commission — checked 13 October 2026
Other sources
- WiT Markets methodology — checked 13 October 2026
Methodology. Editorial policy. Corrections. Report an error.