Your trading plan: rules before you risk a rupee
The answer
A trading plan is a written document that answers every question a trade can ask you, before the trade exists. Its purpose is not to make you disciplined. Its purpose is to remove 1 specific failure: changing the rule after you have seen the price.
Why this costs you money
An unwritten plan is not a plan. It is a memory, and memory adjusts itself.
Here is the mechanism, and it is not about willpower.
You decide to buy when a stock breaks above its 20-day high. The stock reaches that level. You are watching. It moves 2% past the level in 4 minutes and you were not fast enough. Now you have a choice you never wrote down: buy late, or skip it.
You buy late. It works. Nothing is learned, because the outcome was good.
Three weeks later the same thing happens and you buy late again. This time it does not work. Your stop, which you placed at the level you originally intended, is now much closer than planned, and it is hit within the hour.
Read that again. The rule did not fail. There was no rule. There was an intention, and the intention was edited by the screen.
This is what an unwritten plan produces. Not a dramatic loss, but a slow accumulation of small decisions that were never tested, cannot be reviewed, and cannot be improved, because nobody knows what the rule was.
There is a second cost and it is larger. Without a written rule you cannot measure anything. You cannot say whether the method works, because "the method" was different every time. Six months of trades produce no information at all.
How it works
A trading plan has 9 parts. Missing any 1 of them creates a question you will have to answer while a position is moving, which is the worst moment available.
1. What you trade. Name the segment and the instruments. Cash equity in companies above a stated market capitalisation. Index futures. Index options. Not "the market".
2. What timeframe. The chart timeframe your signals come from, and the expected holding period. Daily charts, positions held 3 to 15 sessions.
3. Your setup. The condition that makes an instrument a candidate. This is your edge, meaning the reason you expect a positive result over many trades. Write what it is and why you believe it. If you cannot say why, you do not have an edge. You have a pattern you have seen.
4. Your entry trigger. The exact event that makes you place the order, and the order type you use. This is the part that must survive the stranger test below.
5. Your stop. The price at which the reason for the position has stopped being true, decided before entry. Not an amount of money. A price.
6. Your profit exit. How you leave a position that works. A target, a trailing rule, a time limit, or a combination. Most plans have a detailed entry and no exit, which is why most plans produce small winners and large losers.
7. Your size. The formula, not a number. Shares = (account × risk %) ÷ (entry − stop), with the caps that sit on top of it.
8. Your loss limits. Three numbers: maximum loss per position, per day, per week. When a limit is hit, you stop for that period. This is a mechanism, not a virtue. Its specific job is to prevent a bad hour from becoming a bad month.
9. When you do not trade. The most valuable section and the one nobody writes. Results day for the instrument. The first 15 minutes. Any day you have slept badly, or have less than the full attention your method needs. The session after you hit a loss limit.
The stranger test
Here is how you know whether part 4 is written or only felt.
Could a stranger, given only your written rule and a chart, place the same order you would place, without asking you a single question?
Almost every rule fails this on the first attempt. "Buy on a breakout" fails immediately. A stranger asks: a breakout of what, measured over how many days, on what timeframe, on the close or intraday, at what price, with what order type, and what if it gaps above the level at the open?
A rule that passes reads more like this. On the daily chart, when today's close is above the highest close of the previous 20 sessions, and the 20-session average volume is at least 3,00,000 shares, place a buy limit order for the next session at yesterday's close plus 0.25%, valid for that session only.
That is not a better idea than "buy on a breakout". It is the same idea, made executable. The difference is that the second version can be tested, repeated and reviewed, and the first cannot.
What it tells you, and what it does not
A written plan tells you whether your results came from your method. That is its entire analytical value. If you followed the plan on 100 trades and lost, the method is wrong and you can change it. If you did not follow it, you have learned nothing about the method and nothing about yourself except that the plan was unusable.
It does not tell you the plan is profitable. A plan can be perfectly specified and still have a negative expectancy. Precision is not an edge; it is the condition under which an edge can be detected.
It also does not survive contact with a market it was not designed for. A plan built on 20-session breakouts works in trending conditions and loses steadily in range-bound conditions. The plan should say which conditions it expects, so you know whether a losing period is a failure of the method or the method meeting the conditions it was always going to lose in.
The decision rule
A rule is written when a stranger could execute it. Until then it is a preference, and preferences change when the price moves.
Test every line of the plan against 3 questions:
- Is there a number, a price or a named condition in it?
- Could 2 different people read it and take the same action?
- Does it say what happens in the awkward case — a gap, a halt, a missed
fill, a partial fill?
Any line that fails all 3 is decoration. Delete it or fix it.
Try this now
This takes 5 minutes and it is the most uncomfortable exercise in the cluster.
- Open a blank page. Write your entry rule, exactly as you would follow it tomorrow. Do not tidy it. Write what you actually do.
- Now read it as a stranger. For each vague word, write the question the stranger would ask. "Strong" — how strong, measured how? "Near support" — how near, and which support? "Confirmation" — of what, on which candle?
- Answer every question with a number, a price, a timeframe or a named condition.
- Rewrite the rule using only those answers.
- Count the words in version 1 and version 2.
What you should see. Version 1 is usually 1 or 2 lines. Version 2 is usually 6 to 10 lines, and about half the questions in step 2 will have no answer at all, because you have been deciding them by feel each time.
Those unanswered questions are your real trading method. Every one of them is a place where the screen has been making the decision. If you have 5 unanswered questions, you do not have 1 strategy. You have 32 different strategies, chosen at random, and no amount of trade history will ever tell you which of them works.
If you have never traded, do this on the rule you were planning to use. Doing it before the first position costs 5 minutes. Doing it afterwards costs a year of unreadable records.
Three real cases
1. The Turtle programme, Richard Dennis and William Eckhardt (Chicago, recruits trained from 1983 and 1984) — written rules given to novices Dennis and Eckhardt disagreed about whether trading could be taught. To settle it, Dennis recruited people with almost no market experience, gave them a complete written rule set covering entries, exits, stops and position size, and funded them. Several of the group went on to run money professionally. The instructive part is what was given to them: not judgement, not experience, but a document. The rules were specific enough that novices could execute them, which is the exact standard in the stranger test above.
2. JPMorgan Chase, the London Whale (losses disclosed 2012) — the rule was changed to fit the position Traders in the bank's Chief Investment Office, including Bruno Iksil, built very large credit derivative positions. As losses grew, the bank's risk model was changed in a way that reported a lower risk figure for the same position. The US Senate Permanent Subcommittee on Investigations published a report on the matter in March 2013. Losses were reported in the region of $6 billion. An institution with a full risk department did the exact thing an individual does with an unwritten stop: it moved the limit rather than the position.
3. Amaranth Advisors (United States, collapsed September 2006) — no limit that could stop it The fund lost approximately $6 billion in a matter of weeks on natural gas futures positions run by Brian Hunter, and closed. Regulatory actions followed from the Commodity Futures Trading Commission and the Federal Energy Regulatory Commission. The positions were within the fund's mandate. What was missing was a limit that could halt the activity before the capital ran out, which is exactly what part 8 of a plan is for.
The question that resolves it
A novice asks: is my strategy any good?
An expert asks: is my strategy written down precisely enough that the question can be answered at all?
Almost every person asking the first question has no way to answer it, because what they executed was not a single strategy.
What would make this wrong
If traders with written plans performed no better than traders without them, this article would be wrong. That specific comparison has not been run cleanly at retail scale, and it would be difficult to run, because the people who write plans differ from those who do not in other ways as well.
The honest limits.
A plan can be precise and still lose. Precision makes results readable. It does not make them positive. Some readers will follow a written plan for 100 trades and find that the method has no edge. That is the plan working, and it is cheaper than 3 years of unreadable trades.
Over-specification has a cost. A plan with 40 conditions will produce almost no trades, and the ones it produces will be curve-fitted to the past. A usable plan has 5 to 9 conditions, not 40.
Discretion is not automatically worse. Some experienced traders use judgement well. The important detail is that they still record what they did and why, so it can be reviewed. Undocumented discretion is the problem, not discretion.
In India
Session times set the shape of the plan. The equity market has a pre-open session from 9:00 to 9:15 and continuous trading from 9:15 to 15:30, Monday to Friday.
Automatic square-off is a rule you did not write. Intraday positions taken with intraday leverage are squared off by the broker before the close, commonly in the last 15 to 30 minutes of the session. The exact time varies by broker and by segment. Your plan must state the time by which you exit, because if you do not, the broker's system decides for you at whatever price exists at that moment.
Weekly expiry sets a hard clock. Index options expire weekly. A plan that uses options must say which expiry it uses and at what point before expiry it will not open a new position. SEBI revised the expiry structure in 2024 and 2025.
Margin rules limit intraday size. SEBI's peak margin framework, phased in from 2020, requires upfront collection of margin and reports margin on intraday snapshots rather than end of day. The practical effect is that the intraday leverage available to a retail account is far lower than it was before
- A plan written from older material will assume leverage that no longer
exists.
In the United States
The session is longer than the exchange hours suggest. Regular trading runs from 9:30 to 16:00 Eastern time. Pre-market trading is widely available from about 4:00, and after-hours trading until about 20:00, at most retail brokers.
Liquidity outside regular hours is thin and spreads are wide, so a plan must state whether it permits orders in those windows at all.
A regulator writes part of your plan. Under FINRA's margin rules, an account making 4 or more day trades in 5 business days in a margin account is designated a pattern day trader and must hold at least $25,000 in equity. An American plan for a small account has to include a trade-frequency limit, because the alternative is a restricted account.
Leverage is standardised. Regulation T permits 50% initial margin on shares. Day trading buying power in a pattern day trader account is commonly 4 times maintenance margin excess.
Earnings dates are known and clustered. American companies report on a predictable calendar and the dates are published well in advance. A plan can and should include a rule about holding through results.
Where they differ, and what that tells you
The Indian plan needs 2 clauses the American plan does not, and the American plan needs 1 the Indian plan does not.
The Indian additions are the square-off clause and the expiry clause. An Indian intraday position has a compulsory end time set by a broker's system, not by you, and an Indian index option has a weekly deadline that removes its value whether or not your view was correct. Neither appears in imported material, because neither exists in the same form in the United States.
The American addition is the trade-frequency clause forced by the pattern day trader rule.
What that tells you is the direction each market pushes a beginner. The American structure limits how often a small account can trade and leaves the holding period open. The Indian structure limits how long a position can be held and leaves the frequency open. A beginner in India can therefore trade every session, in an instrument with a 1-week life, with a position that must close by the afternoon. Every part of that pushes towards more decisions per week, and more decisions per week is where an unwritten plan does its damage fastest.
Carry this
- A rule is written when a stranger could execute it without asking a question.
- Nine parts: instrument, timeframe, setup, entry, stop, profit exit, size, loss limits, and when not to trade.
- The section nobody writes is number 9, and it is the one that saves the most.
- Precision does not make a method profitable. It makes the method measurable, which is the only route to finding out.