Here is a result that shows up over and over when traders finally tag their history. Overall they are roughly flat. Split by setup, two of their three patterns are modestly profitable and the third loses enough to eat both. The third one is usually the trade they find most exciting. Without tags, that structure is invisible, because the aggregate number averages it away.
Tagging is the step that turns a trade log into something you can query. It costs about fifteen seconds per trade and it is the highest return administrative work in trading.
Start with four tag groups, not forty
The failure mode is enthusiasm. Traders build a taxonomy with thirty labels, use it inconsistently for three weeks, and abandon it. Every tag you add splits the sample, and a sample split thirty ways has nothing to say.
Four groups cover most of what you need. Setup: the pattern you claim to trade, three to six names, no more. Context: the market condition, something like trending, ranging, news-driven. Session: London, New York, Asia, or the overlap. Execution: what you actually did, entered early, chased, moved the stop, closed at target, cut before target.
That last group is where the money hides. Setup tags tell you what you look for. Execution tags tell you what you do, and the gap between them is the real report. A trader whose "moved stop" tag appears on 20 percent of trades and carries the worst average result does not have a strategy problem.
Tag facts, not feelings dressed as facts
Emotional tags are worth keeping, but they have to be recorded at entry, not at review. "Revenge trade" written after a loss is a confession. "Took this within 10 minutes of a loss" recorded automatically from timestamps is data, and it produces the same insight without relying on honest hindsight. Anything you can derive mechanically, time since last trade, whether the previous trade lost, position size relative to average, should be derived rather than typed. That covers most of what people mean by revenge trading and overtrading, and it removes the bias.
Keep one free-text field for the note that does not fit a tag. Do not let it replace the tags. Free text cannot be counted.
A tag is only useful if it is applied the same way every time. Write a one-line definition for each tag and keep it visible. If you cannot state in a sentence what makes a trade a "breakout", the tag will drift and the analysis built on it will be false precision.
Reading the output
Once you have a few hundred tagged trades, three cuts do most of the work.
Expectancy by setup, measured in R rather than currency, tells you which patterns pay. Use R so that position size differences do not distort the comparison. This is the number described in expectancy, and per-setup expectancy is where it becomes actionable rather than academic.
Win rate and average R by session tells you when to trade. Plenty of traders discover they lose consistently in one session and keep trading it out of habit or timezone convenience. The session breakdown in trading sessions explains why liquidity differs enough for the same setup to behave differently at different hours.
Result by execution tag tells you what to fix first. If your average trade is positive when you leave the stop alone and negative when you move it, you have found a rule, not a preference.
Be careful with sample size. Ten trades in a bucket is a story, not evidence. Thirty is thin. A hundred starts to mean something. If a tag has too few trades to judge, the honest conclusion is that you do not know yet.
Keeping it consistent
Tags decay because tagging is done later. Tag at the moment of exit, or better, set the setup and context tags at entry when the reasoning is present. Most journal software supports templates or hotkeys for this, and the practical differences between packages are covered in trading journal apps. A spreadsheet works too, as long as the tag values come from a dropdown rather than free typing, because "breakout", "Breakout" and "b/o" are three different tags to a filter.
Review on a schedule, not on emotion. Monthly is enough for setup-level conclusions, weekly for execution ones. Reviewing after a bad day produces conclusions about the bad day. The wider discipline of what to record and how to sit with it is in the trading journal guide.
What to do with what you find
The output of tagging is a set of removals, more often than additions. Drop the setup that has lost across a hundred trades. Stop trading the session where your edge disappears. Add a rule that blocks the execution behaviour that costs you.
Resist the opposite move, which is to over-optimise on your own history. If a tag combination shows a superb result across twenty trades, that is noise until it is not. Treat the finding as a hypothesis and track it forward, the same way you would treat any other fitted result. Trading is high risk and your own sample is small, so the discipline that applies to system testing applies to your journal too.
Firms see the same structure from the other side. A prop firm reviewing which behaviours precede a blown account is tagging too, just at portfolio scale, and the account-level analysis in a Prop Firm CRM is built on exactly this idea: attach labels to trades, then count.
"Every trader I know has one setup they love and one that pays them. Tagging is how you find out they are not the same setup."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- Keep the tag set small, four groups and a handful of values each, or every bucket becomes too thin to read.
- Derive behavioural tags from timestamps and sizes instead of typing them after the fact.
- Measure expectancy in R by setup, session and execution tag, and treat buckets under about thirty trades as unproven.
- The usual outcome of tagging is a removal: one setup, one session or one habit that has been paying for the others.
Frequently Asked Questions
How many tags should a trade have?
Two to four is enough for most traders: a setup, a market context, a session and, where relevant, an execution note. More than that fragments the sample so badly that no bucket accumulates enough trades to say anything, and the tagging itself becomes a chore that gets dropped.
How many trades before tag analysis means anything?
It depends on the bucket, not the total. A tag with ten trades behind it is anecdote, thirty is thin, and around a hundred starts to carry weight. That is why a small tag set matters, since it concentrates your history into fewer, larger buckets.
Should emotions be tagged?
Record them, but prefer mechanical proxies. Time since the last trade, whether the previous trade lost, and size relative to your average capture most emotional patterns without depending on honest self-reporting after the fact, which tends to be rewritten by the outcome.
About the Author
Alex Onta is an Executive Director at SINGUARD. He built eTrader, the terminal, the mobile apps, eTrader Broker, Copytrading, Business and Community, along with the worldwide clustered-server infrastructure it all runs on, with his brother Roman Onta helping on the design, and he leads that division today. Together with Roman he builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals, and the two of them carry worldwide compliance, payment processing and international business structuring side by side. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.