A Trading Journal That Actually Teaches You Something
A trading journal teaches you something when it stops being a diary and becomes a dataset: one row per trade, thirty-three columns, the plan written down at entry before you know how the trade ends, and the whole record graded on process — never on profit. I call mine the DataMine, because that's what you do with it. You mine it.
Most journals never get there. They're feelings with timestamps — "felt good about this one, market was choppy, need to be more patient" — and after a year of faithful entries they can't answer a single question about your trading with a number. This article is the structure that fixes that: what to log, why the plan columns are the load-bearing part, how the grading works, and the honesty rules that keep the whole thing from quietly turning into a story you tell yourself.
Why do most trading journals fail?
Because they have no structure. That's the entire diagnosis. All journaling is not the same — most of it, honestly, is a waste of time and energy, because a paragraph of prose about how a trade felt cannot be sorted, filtered, counted, or compared. Ten prose entries are ten anecdotes. Ten structured rows are the start of a dataset.
I'll own my part of this: I didn't keep a journal when I started trading. I wanted to make money quickly and the habit felt like homework. It took losing — enough that I couldn't pretend the problem was bad luck — before I accepted that the record has to come first. The journal I describe here is the one I built out of that lesson and the one I use for my own trading today.
The difference shows up the first time you ask your record a question. "Do my losses cluster on days I skipped my checklist?" A diary shrugs. A dataset answers in thirty seconds with a filter. That answering ability is the whole product, and structure is the only way to get it.
What should a journal actually log?
One trade, one row. The columns come in six groups, and every column has a one-line job:
| Group | What it holds | When you fill it |
|---|---|---|
| Identity | Trade number, dates, ticker, direction, setup name — which trade this is | At entry |
| Execution | Entry, exit, size, times, fees, realized result — the plain facts from the broker | At entry and close; much of it can import from broker fills |
| Plan + Discipline | Planned stop, planned target, planned risk, expected reward-to-risk, gate and checklist checks, one line on why you took it | At entry, before the outcome exists |
| Factors | The readings your setup was built on — volume versus normal, momentum, gap size, sector strength, whatever your system actually uses | Auto-joined from a scan, or noted by hand |
| Outcome | A one-word resolution label, realized reward-to-risk, whether the setup was still valid at the close, hold time | At close |
| Notes | Everything the columns didn't catch — where a pattern first shows up as a hunch | Whenever |
The full log runs thirty-three columns plus a Notes column at the end. If that sounds like a lot on day one, don't let it stop you: start with the starred minimum — trade number, date, ticker, direction, setup name, planned stop, planned target, realized result, the outcome label, and realized R — log ten trades, and add the rest when the habit holds. A short honest log beats a perfect empty one.
Three small spreadsheet formulas do all the arithmetic, and you set them once:
- Planned risk — the distance from your entry to your stop, times your size.
- Planned R — the distance to your target divided by the distance to your stop: the reward-to-risk you expected going in.
- Realized R — your actual result divided by your planned risk.
R is the currency the whole log runs on. It lets a small trade and a large one sit on the same scale, so your record compares decisions instead of position sizes. The full four-part picture behind those plan columns — size, stop, target, and the time dimension almost nobody logs — gets its own treatment in risk in four dimensions.
The long version of this is in the book — Become a Cyclitecnical Trader: the cycle ladder, the FLD, and the eight interactions, written out end to end. It's free. Send me a copy. We email it to you. No card, and you can unsubscribe any time.
Why does the plan have to be written at entry?
Because a plan written after the outcome is fiction, and discipline measured against fiction is flattery.
The Plan + Discipline columns are only honest if you fill them the moment you enter — planned stop, planned target, whether the setup actually cleared your rules, and one line, in your own words, on why you took it. Written before the outcome, those entries are a real test of your judgment. Written after, they are you drawing the target around the arrow. Every trader who backfills a journal produces a record that says their process was excellent, because hindsight quietly edits the plan until it matches what happened.
So the rule is blunt: write the plan first, or leave the columns blank. A blank cell is honest — it says "I entered without a written plan," which is real information about your process. A backfilled cell is a small lie that compounds, because every summary statistic built on it inherits the flattery.
This same discipline — the record is only valid if it's logged before the outcome — is exactly how a public track record has to work too. I've written up the scoring rules I hold myself to in how a trading educator should score their own calls; the journal is the private version of the same contract.
What does it mean to grade process instead of profit?
It means the scorecard asks one question — "did I follow my own rules?" — and never asks "did I make money?" Every trade gets six plain yes-or-no checks, one point each:
| Check | The question |
|---|---|
| Followed my plan | Did I take the trade my rules said to take, and manage it the way I said I would? |
| Waited for the signal | Did I wait for the actual trigger, or jump early on a feeling? |
| Right size | Did I size from my risk rule, not from how confident I felt? |
| Honored my stop | Did I exit where I planned to be wrong, or slide it and hope? |
| Respected the expiry | When the setup went invalid, did I drop it, or hold a dead trade? |
| Cut the loser near −1R | Did my losses come in around the risk I planned, instead of blowing past it? |
Here's why this works. A trade can lose money and score a perfect six — you read it well, sized it right, took the signal, and the market didn't cooperate. That's a good trade with a bad outcome, and you want more of them. A trade can also make money and score a two — you broke your rules, oversized, ignored your stop, and got bailed out. That's a bad trade with a good outcome, and it's the most dangerous kind you can have, because it pays you to repeat the wrong thing. The scorecard exists to tell those two apart. Nothing else on your desk does — your account balance certainly doesn't.
The result of any single trade is mostly out of your hands. The process is entirely in them. So the process is the only thing worth grading — and all the grading can honestly claim to measure is whether you're becoming more disciplined. It says nothing, and promises nothing, about money. I want that limit stated out loud.
Here's what the honest-record habit looks like against my own public calls — logged before the outcome, scored after, misses kept:
2026-07-09 — AMD, graded a miss. On the 4PM show I flagged AMD as a long: the setup read aligned to me across the stock, its sector, and the broad market, and I said so on air. It worked for exactly one day — up about 2% the next session — then gave all of it back and kept going, down roughly 9% from the call by July 17. The ledger scored it a miss. Memory, left alone, would have kept the good first day and quietly dropped the rest. The row keeps both. That's the entire argument for the row.
How does a journal become the only backtest of you that exists?
You can backtest a strategy against decades of price data. You cannot backtest yourself against anything — except the record of what you actually did. Your hesitations, your oversizes, the setups you claim to trade versus the ones you actually take: none of that exists in any dataset on earth until you write it down. Thirty honest rows are the only evidence about your own trading that will ever exist. That's why I call the practice a DataMine — the record is the mine, and the review is the digging.
Two things make the mine productive. The first is the setup-name column: tagging every trade with which of your own setups it was lets you group later and discover which setups you actually run — which is routinely a surprise. The second is indexing rows to market context, because it lets you put today's situation next to every similar past situation you've logged and read them side by side. The market repeats its rhythms often enough that a well-indexed record stops being a rear-view mirror and becomes preparation.
One hard warning, because this is where journals go wrong in the other direction: every pattern the log surfaces is a question, never an answer. The record can show you that your losses clustered on choppy days. It cannot tell you choppy days are the cause — thirty trades can hand you a coincidence dressed as a law. When something jumps out, write it down as a question and investigate it deliberately, on more data, before it changes a single trade. The record surfaces patterns. It does not certify them. Sometimes what it surfaces is that your best decision on certain days was no trade at all — the null trade is a position too, and a good journal is where you first see it earning its keep.
How often should I review it?
Three passes, three speeds:
- Daily — fill it. Log every trade the same day, plan columns at entry, outcome columns at the close. Done live it's under a minute a trade. Done from memory a week later, it's fiction.
- Weekly — skim it. Read the week's rows and check one thing only: is the record complete and honest? Every trade in, plans written at entry, nothing quietly omitted because it embarrasses you. Fix the record, not the trading — that comes later.
- Monthly — mine it. Once the sample is real, read the whole record and ask what it's trying to tell you. Where does the process hold? Where does it leak? Which setups do you actually run?
And one hard floor under all of it: read nothing into the numbers before roughly thirty trades. Below that, the statistics are noise wearing a suit — five wins prove nothing, five losses prove nothing, and a ten-trade scorecard will tell you a confident story that the next ten trades flatly contradict. "Not enough data yet" is not the tool failing. It's the tool being honest, and it will be the correct answer more often than you'd like.
What are the honesty rules?
Five. They're not decoration — break any one and the journal turns from a mirror into a story.
- Record what happened, not what you wish had happened. The trade that embarrasses you is exactly the one to log in full.
- Fill the plan at entry, not after. Written first, it tests your judgment. Written after, it's the target drawn around the arrow.
- Descriptive, not predictive. The log describes your past process. It does not predict your future results, and a pattern that held before is evidence, not a promise.
- It scores process, not profit. A high process score is not permission to size up. There is no recommendation hiding anywhere in this practice.
- Your data is yours. The log lives in your spreadsheet, on your machine, answerable to nobody but you — which is precisely why it can afford to be honest.
Questions traders ask
Do I need journaling software to do this?
No. A blank spreadsheet and the habit are the entire toolkit — headers across row one, one trade per row, three formulas dragged down a column. Software can automate the broker-import parts, but the load-bearing columns are the ones only you can fill, at entry, in your own words.
How many trades before the journal tells me anything real?
About thirty, as a floor — and treat everything before that as provisional. Small samples mislead confidently: they'll hand you a "pattern" that's pure coincidence and a scorecard that flatters or slanders you at random. Keep logging through the silence. The discipline of the logging is doing its work long before the statistics are.
Isn't thirty-three columns overkill?
For day one, yes — which is why you don't start there. Start with the starred minimum and let the log grow as the habit does. But the full width is what makes the record mineable later: every column is a question you'll eventually want to ask, logged in advance. You can't retroactively ask your first hundred trades what the volume looked like if you never wrote it down.
Should the journal tell me what to trade?
No — and if you ever catch it nudging you toward a trade, you're using it wrong. The record is descriptive: it's a witness to what you did, never a fortune-teller about what you'll get. Every pattern it surfaces is a hypothesis to investigate, not a rule to deploy. You keep every decision. The journal just makes sure you can no longer lie to yourself about the decisions you've already made.
Where does the journal meet the method?
Everything above works for any trader with any system — the structure doesn't care what your setups are. My own journal goes one step further: it's indexed to the market's recurring cycles, so each day's row can sit next to the same point in every past repetition of the rhythm, and the expectation for tomorrow gets built against everything those past days actually did. If you want the method that indexing comes from, start with the 80-day cycle — the rhythm the whole practice is organized around — and the FLD, the displaced price line whose interactions give each day of the cycle its character, in what is a Future Line of Demarcation. How the full framework fits together, and how we score our own public reads against it, is laid out in the methodology hub. The journal is where that method stops being theory and starts being your own evidence.
Keep reading
Get the Cycle Pass — from $297