The trader's daily process and best practices
The answer
A trading process is a fixed sequence of steps performed at fixed times. Its purpose is not to make you feel organised. It is to make your decisions comparable to each other, because a decision that cannot be compared to your other decisions cannot be improved.
Why this costs you money
Without a written record you have no data about yourself, and every conclusion you reach about your trading is a memory of the trades you happen to remember.
Memory is not neutral about this. It keeps the large win and the dramatic loss. It discards the 40 ordinary trades where the actual pattern lives.
Here is the specific cost, and it is not abstract.
A trader loses money for 8 months. They conclude the method is wrong and change it. They now have 8 months of results from method 1 and a new method 2, and no record that could tell them whether method 1's problem was the entry rule, the exit rule, the position size, the market conditions, or the fact that they overrode the plan on 30% of trades.
So the change is a guess. Method 2 gets 8 months, produces mixed results, and the process repeats. Three years pass in which the trader has taken perhaps 500 positions and learned almost nothing, because no 2 of those positions can be compared.
The loss is not the money in those 500 trades. It is that the 500 trades produced no information. That is a very expensive way to buy nothing.
There is a second cost, and Indian readers should note it particularly. Trading income is commonly treated as business income, which brings record-keeping and, above certain thresholds, audit obligations. A trader with no records is reconstructing a year of activity in June.
How it works
The process has 4 layers, and each one runs at a fixed time.
Before the market opens, 20 to 30 minutes
1. Review what you already hold. For every open position: the stop is where it should be, and it exists as a live order rather than as an intention.
2. Check the calendar for what you hold. Results dates, dividend and other corporate action dates, scheduled economic announcements. Your plan says whether you hold through those. Apply it now, not at 14:00 when the announcement is 1 hour away.
3. Calculate your current heat. The sum of money at risk across all open positions, as a percentage of the account. This is the number from article 4. It tells you whether you may take anything new today.
4. Prepare candidates in full. For each candidate, write the trigger price, the stop price, the resulting position size, and the order type. Written before the open, when nothing is moving. If a candidate cannot be fully specified, it is not a candidate.
5. Write today's no-trade conditions. Any that apply today, from your plan and from your own state. Slept badly, travelling, a family matter, an important meeting at 11:00. Write it down, because at 10:45 you will not.
During the session
Execute what you prepared. Do not research.
The only new decisions permitted during the session are the ones your plan already contains: your trigger fired, or your exit condition occurred. Anything that requires you to form a new opinion while the market is open belongs to the next morning's preparation.
This is a mechanism with a specific purpose. Almost every failure in article 6 occurs during the session, when a decision is made in a compressed period with money moving. The process does not ask you to make those decisions well. It removes the occasions on which they arise.
After the market closes, 15 minutes
Log every trade, the same day. A journal entry written 3 days later is a reconstruction. The fields that matter most are the ones memory rewrites first.
The 10 fields:
| Field | Why it is there |
|---|---|
| Date, instrument, long or short | Identifies the trade |
| Setup name | Lets you group trades and measure each setup separately |
| Entry price and time | Shows whether you entered at your trigger or chased |
| Planned stop, and planned target | The commitment made before the outcome |
| Size, and risk as a percentage | Shows whether size was consistent |
| Exit price, time and reason | Distinguishes a plan exit from a decision exit |
| Result in R | Makes trades of different sizes comparable |
| Plan followed: yes or no | The load-bearing field |
| If no, exactly what differed | Turns an override into data |
| One line on your state | Sleep, distraction, whether you were behind for the week |
The 8th field is the one that makes the journal work. Without it you have a record of outcomes, which the broker already gives you. With it you have a record of behaviour, which nobody gives you.
Once a week, 30 to 45 minutes
This is where the journal becomes useful. Four calculations.
1. Expectancy by setup. Group trades by setup name and calculate expectancy in R for each. Most traders find that 1 of their setups produces all the profit and 1 produces most of the loss, and they had no idea which was which.
2. The plan-followed percentage. Count the trades marked yes, divide by total. This number is the honest measure of the week, and it is the only measure you control. Below 80%, no conclusion about the method is available, because what you executed was not the method.
3. Every override, listed. Read the "what differed" field on each one. A repeated override is not an error; it is an unwritten rule you already follow. Either add it to the plan and test it, or add a mechanism that prevents it.
4. Costs. Total brokerage, taxes and fees for the week, as a percentage of gross profit. Most retail traders have never calculated this and are surprised.
Once a month, 30 minutes
Heat and correlation groups, from article 4. Phase progress, from article 5: trade count, expectancy and drawdown against your ladder. And the drawdown itself, measured from the peak.
When to step away
The step-away triggers are written in advance, like every other limit here, and they are not about how you feel.
- The weekly or monthly loss limit is hit.
- Three or more overrides in a week.
- A period when your available attention is reduced: illness, a family matter, a demanding period at work.
- Any week when the pre-market preparation was skipped more than once. The skipping is the signal, not the results.
What it tells you, and what it does not
A journal tells you what you did, and grouping it tells you which parts of what you did produced the results. That is its whole function, and it is enough.
It does not tell you whether the method will work in the next market regime. A setup with a strong expectancy over 100 trades in a rising market has been measured in a rising market only.
And it does not, by itself, change behaviour. A journal that is written and never grouped is a diary. The weekly review is the part that does the work, and it is the part most often abandoned, because writing entries feels productive and reading them does not.
The decision rule
Every trade is graded twice: once on the result, which you do not control, and once on whether the plan was followed, which you do.
The weekly number that matters is the second one.
- Plan-followed above 80%: the results are readable, and you may draw
conclusions about the method.
- Plan-followed below 80%: no conclusion about the method is available. The
work this week is on the mechanisms, not on the strategy.
A trader looking for a better strategy while executing the current one 60% of the time is solving the wrong problem, and will solve it again next year.
Try this now
Five minutes. This is a diagnostic, not a setup task.
- Open your broker's tradebook and find your last 10 closed trades.
- On a page, draw 10 rows and the 10 columns from the table above.
- Fill in what you can for each trade, from records and memory.
- Count the cells you cannot fill.
What you should see. The first 6 columns fill easily, because the broker records them. The last 4 usually cannot be filled at all.
You will not know what setup most of them were, because you did not name your setups. You will not know the planned stop, because it was a level you were watching. You will not know whether the plan was followed, because the plan and the trade were the same thought.
Those empty cells are the reason your results are unreadable. Not the strategy. Not the market. Ten trades, and 4 of the 10 columns are blank, which means 500 trades would produce the same blank columns 500 times.
The fix takes 2 minutes. Copy the 10 columns into a spreadsheet. Fill it in for the next trade you take, before you place the order, not after. The first 6 fields take 40 seconds. Fields 7 to 10 take 30 seconds after the exit.
If you have no closed trades yet, this is the most valuable minute in the cluster. Build the sheet now. Traders who start a journal in year 3 have year 3 onwards. Traders who start it on trade 1 have everything.
Three real cases
1. Emkay Global Financial Services, 5 October 2012 (India) — an order that skipped a check An erroneous set of orders in a basket of Nifty constituents was entered, causing a very large intraday fall in the Nifty 50 index before trading was halted. NSE annulled a number of the trades and regulatory action followed against the broker. The failure was not analytical. An order was entered without the verification step that would have caught it, in a firm where that step existed.
2. Knight Capital Americas, 1 August 2012 (United States) — a deployment without a checklist A software deployment left old code active on 1 of the firm's servers, causing its systems to send a very large number of unintended orders. Losses of approximately $440 million accumulated in roughly 45 minutes. The firm was recapitalised days later and merged the following year. The SEC brought an administrative proceeding in 2013. The relevant detail for an individual is that the error was not detected during the deployment, because the step that would have detected it was not part of a required sequence.
3. Mizuho Securities, 8 December 2005 (Japan) — the confirmation that was dismissed A dealer entering an order for the newly listed shares of J-Com transposed the price and the quantity, entering an order to sell 610,000 shares at 1 yen rather than 1 share at 610,000 yen. A warning message was reportedly dismissed, and the firm's attempts to cancel were unsuccessful. The reported loss was in the region of 40 billion yen. A confirmation step that can be dismissed with 1 action is not a control. This is the same failure mode described for every self-imposed rule in this cluster.
The question that resolves it
A novice asks: how did I do this week?
An expert asks: what percentage of this week's trades followed the plan, and which setup produced the profit and loss?
The first question is answered by the account balance and it tells you almost nothing, because a week is far too short for the result to carry information. The second is answered by the journal and it tells you what to change.
What would make this wrong
If traders who keep journals performed no better than traders who do not, the central claim here would be wrong. That comparison has not been run cleanly, and it would be hard to run, because people who keep records differ in other ways as well.
The honest limits.
A journal has a real time cost. Fifteen minutes a day and 45 minutes a week is about 6 hours a month. For a trader taking 4 positions a month, that cost is larger than the benefit, and a simpler record is correct.
Records do not improve a method that has no edge. A perfectly kept journal of a losing system is a precise measurement of a loss. It is still worth having, because it identifies the loss quickly rather than in year 3.
Over-measurement is possible. A trader who reviews performance every day will react to noise, and reacting to noise is the behaviour in article 3 that takes size down before recoveries. Weekly is frequent enough. Daily review of process is fine; daily review of results is not.
Checklists fail when they can be dismissed. The Mizuho case is the general version of this. A step that can be skipped with 1 action will be skipped on the day it matters, which is the day you are in a hurry.
In India
The session shapes the timing. The pre-open session runs from 9:00 to 9:15 and continuous trading from 9:15 to 15:30, so preparation has to be complete before 9:00.
The square-off deadline is part of the process. Intraday positions taken with intraday leverage are squared off by the broker before the close, commonly in the last 15 to 30 minutes. Your process needs its own exit time, earlier than the broker's, written down.
Records have a second purpose. Trading income is commonly treated as business income rather than capital gains, with intraday equity trading as speculative business income and derivatives as non-speculative business income. This brings bookkeeping obligations and, above certain turnover thresholds, an audit requirement. The practical effect is that an Indian trader needs records regardless of whether they want to improve, and the journal described here already contains most of what is required.
Broker documents you should collect monthly. The contract notes, the tradebook, the profit and loss statement, and the monthly holding statement from CDSL or NSDL. The last one is worth reading rather than filing.
The holiday calendar matters more than in the United States, because Indian market holidays are numerous and fall on varying dates. A weekly process should include checking the coming week's trading days.
In the United States
The session is longer and the process has to state its own hours. Regular trading runs from 9:30 to 16:00 Eastern time, with pre-market from about 4:00 and after-hours to about 20:00 at most retail brokers. A process that does not name its hours will expand into all of them.
Scheduled events are published far in advance. Federal Reserve meeting dates are published for the year ahead, and company results dates are announced weeks in advance. The pre-market calendar check is therefore straightforward and there is little excuse for being surprised by a scheduled event.
Part of the record is created for you. Brokers issue Form 1099-B, which reports proceeds and, for covered securities, the cost basis to the Internal Revenue Service. This is not a substitute for a journal, because it contains no setup names, no planned stops and no plan-followed field, but the raw transaction record exists without effort.
The wash sale rule forces a specific record. A loss is disallowed if a substantially identical security is bought within 30 days before or after the sale, which means an American trader must track re-entries across a 61-day window. Brokers report this for identical securities in the same account, but not across accounts.
Where they differ, and what that tells you
In the United States, the tax record is largely produced for you. In India, you produce it.
An American trader receives a 1099-B with proceeds and cost basis reported to the tax authority. The trading record therefore exists in a usable form whether or not the trader keeps one.
An Indian trader treating trading as business income is responsible for the books, and above the turnover thresholds for the audit. The record has to be built, and it is built either from a journal kept through the year or from a frantic reconstruction in the following June.
What that tells you is that the Indian journal has 2 purposes and the American journal has 1. The improvement purpose is identical in both countries. The compliance purpose exists only in India, and it means the cost of keeping records has already been paid by any Indian trader who intends to file correctly. Adding the setup name, the planned stop and the plan-followed field to a record you must keep anyway costs about 30 seconds per trade.
Most Indian traders do neither, and then meet both problems in the same month.
Carry this
- Prepare before the open. Execute during the session. Log after the close. Review weekly.
- The load-bearing field is "plan followed: yes or no". Everything else is already in your broker's report.
- Below 80% plan-followed, no conclusion about your method is available.
- A repeated override is an unwritten rule. Write it and test it, or build a mechanism against it.
- Calculate costs as a percentage of gross profit once a week. It is usually the largest single line.