How to Keep a Trading Journal That Improves Your Process

Imagine two traders after 100 trades. One followed a defined setup and took several stopped-out losses to a news spike. The other traded impulsively and got lucky. A journal that records only entry, exit and profit makes the second trader look better. The fix is to record the decision separately from the result.

Use the free trading journal sheet to record the same fields on every trade. This guide focuses on reviewing those records and choosing one change to test.

Why outcome-only journals mislead

One trade's result contains luck. A sound decision can lose and a poor one can win, and over a small sample the noise swamps the signal. If you want the journal to teach you something, it has to ask a different question: given what I knew at entry, was this the trade my plan called for? That question can be answered before the outcome is known, which is why most of the fields below are filled in at entry.

Fields to fill in at entry

1. Setup name. If you can't name it, you're improvising. Write that down.
2. Confirmations. The two or three specific things you saw, such as a volume change, a reaction at a reference level, or a close back through a level.
3. Invalidation. Where the stop is and why that price proves the idea wrong. A stop chosen because it is the loss you can tolerate is an account rule, not an invalidation point, and the two should be recorded separately.
4. Thesis. One sentence on what needs to happen.
5. Emotional state. One word, written honestly: calm, bored, annoyed, rushed.
6. Context. Range or trend, session time in your time zone, scheduled news, and any regime measure you actually use.

Fields to fill in at exit

1. Exit reason. Stop, target, time, discretionary exit, or trail. Mark whether the discretionary exit followed the plan.
2. Process grade from 1 to 5. Define the scale ahead of time. A 5 means you followed the written plan. It should be possible to grade a loser 5 and a winner 1.
3. One improvement. One sentence. Not a post-mortem.

The weekly review

Daily logging collects data and the weekly review is where it gets used. Pick a fixed time and do it even after a bad week.

• Process grades. What share of trades graded 4 or 5? Choose your own target and track the trend. I don't have a published benchmark for what a good share looks like.
• Emotional state. Group trades by the word you wrote and compare average process grade and result. With a small sample, treat any difference as a lead to watch, not a conclusion.
• Setup by type. Compare setups, but keep the sample size in view. Thirty trades per setup is a reasonable minimum before you form an opinion, and it still isn't proof of an edge.
• Exit quality. Compare planned exits to actual exits.
• Context. Compare range days and trend days, event weeks and quiet weeks, to see whether your setups behave differently.

A worked exit-quality example

Suppose a setup risks 8 ticks to make 8 ticks. On ES that is $100 each way per contract ($10 on MES). Before costs, you break even at a 50% win rate. Now suppose you keep taking full 8-tick losses but cut winners at an average of 4 ticks ($50). The breakeven win rate becomes 8 / (8 + 4) = 66.7%. Add commissions and slippage and it climbs further. That is a hypothetical, but the mechanism is general: leaving winners early raises the win rate you need. A capture rate (actual points divided by planned points) puts a number on the habit. You can model different win rates and payoffs with the edge calculator.

Where to keep it

A spreadsheet is enough. Suggested columns:

You own the data, it costs nothing, and you can add columns as you learn. The weakness is friction: it's easiest to skip on the days you most need it. Dedicated journal tools such as Tradervue, TraderSync and Edgewonk import fills from brokers and build charts, which saves typing. Their built-in stats focus on results, so you'll probably still want your own process grade column.

The minimum version

If the full list is too much, start with four things: setup name and a one-word emotional state at entry, then process grade and exit reason at exit. Keep that going for a month. It's more useful than a perfect system you abandon in week two.

Keep sim and live separate

If you trade both, tag each trade. Simulated fills don't capture queue position, slippage or the feeling of real money at risk, and lumping them together can hide differences in your own behavior.

What a journal can and can't do

It can show you which of your own rules you break, and what that costs. It won't create an edge or tell you what to trade, and it can't turn a small sample into proof. Treat each finding as a hypothesis, change one thing at a time, and check it again after enough new trades to see whether it held.