Trade Tags & Setup Analytics: Find Which Trading Strategy Has an Edge

An account can finish profitable while most of its decisions have no edge. One strong setup may carry several weak setups, random entries, and avoidable mistakes. If every trade lives in one undifferentiated list, the overall win rate and P&L hide that structure.

Trade tags turn a journal into a testable dataset. A tag can identify the setup, session, market condition, or execution mistake behind a trade. Once tags are applied consistently, you can answer practical questions: Which setup has positive expectancy? Where do I lose after fees? Does London outperform New York? Are breakout losses caused by the setup or by chasing late entries?

What is a trade tag?

A trade tag is a reusable label attached to a trade so similar observations can be grouped and compared. Unlike a free-form note, the same tag can be filtered across dozens or hundreds of trades. Notes preserve detail; tags create categories.

• Setup tag: breakout, pullback, VWAP fade, opening-range break.
• Session tag: London, New York open, Asia, after-hours.
• Market-condition tag: trend, range, high volatility, news.
• Execution tag: late entry, moved stop, early exit, size creep.
• Process tag: on-plan, revenge, FOMO, boredom, missed checklist.

The purpose is not to label everything. It is to make repeated decisions measurable. Every tag should answer a question you expect to revisit.

The most important rule: separate setup from execution

Suppose ten trades are tagged “breakout” and the result is −3R. That does not prove the breakout setup is weak. Four losses may have come from entering late, moving stops, or taking the pattern outside its planned session. If setup and mistake are collapsed into one label, your analysis blames the strategy for your execution.

This separation also improves losing-trade review. You can compare clean breakout trades with all breakout trades and calculate exactly how much late entries or stop movement cost in R.

Build a tag taxonomy before analyzing results

A taxonomy is simply the controlled list of labels you use. Without one, spelling variations and overlapping meanings fragment the sample: “ORB,” “opening range,” and “opening breakout” become three tiny datasets instead of one useful group.

Layer 1: primary setup

Give every trade exactly one primary setup tag. This is the hypothesis you intended to trade. Keep the active setup list small enough that you can collect meaningful samples—usually three to five setups at a time.

Layer 2: market context

Context tags describe where the setup occurred: session, trend/range, volatility, instrument family, or planned event. Add only context that changes a decision. If you will never compare Tuesday with Thursday, weekday tags add maintenance without insight.

Layer 3: execution and mistakes

Mistake tags describe controllable deviations. Use objective names such as late-entry, moved-stop, oversize, revenge, or outside-session. Avoid vague labels like bad-trade unless your written plan defines exactly what that means.

• Good: breakout, London, trend, late-entry.
• Weak: setup-1, morning-ish, felt-bad, unlucky.
• Good names remain understandable six months later.
• Weak names depend on memory and cannot support a rule.

How many tags should a trader use?

Start with fewer than you think you need. A practical initial library is three to five setup tags, two to four session or context tags, and five to eight mistake tags. You do not apply all of them to every trade; they are the approved vocabulary.

Too few tags mix different decisions. Too many create samples of one. If you have 40 tags after 50 trades, the system is describing rather than measuring. Merge synonyms and archive labels that do not lead to a review question.

A clean naming system for tags

• Use short nouns or kebab-style phrases: breakout, late-entry, high-volatility.
• Choose one spelling and one level of detail.
• Do not combine dimensions inside one tag such as london-breakout-winner.
• Do not encode the outcome; the journal already stores win, loss, P&L, and R.
• Do not rename a losing setup after the fact to protect its statistics.

If your journal supports colors, use them as visual groups rather than meaning by themselves—for example, setups in blue, mistakes in red, and sessions in neutral colors. The tag name must remain understandable without the color.

Tagging manually logged and broker-imported trades

Manual trade entry can capture the setup at the moment you log. Broker import is different: the exchange knows fills, size, prices, fees, and timestamps, but it does not know why you entered. That is why tagging after import is the critical second half of the workflow.

1. Import the complete session so winners, losers, scratches, and fees are present.
2. Open the newest unreviewed trades and assign one primary setup tag.
3. Add session or market-context tags only when they are part of your hypothesis.
4. Add mistake tags where execution deviated from the plan.
5. Write a short note when the tag alone cannot explain the decision.

Traderizz broker import supports Delta Exchange India and Shark Exchange. Re-importing a date range does not require rebuilding your journal; after fills land, add the decision context and analyze them with the same R-multiple workflow as manual trades.

The four numbers to compare for every setup

1. Trade count

Count determines confidence. A setup that earned +4R across four trades is interesting, not proven. Always display sample size beside performance. Avoid ranking a three-trade tag above a 60-trade tag without acknowledging uncertainty.

2. Total R

Total R shows contribution to the book. It answers which setup added or removed the most risk-adjusted return. A low-frequency setup can have excellent expectancy but contribute little because it rarely appears.

3. Expectancy in R per trade

Trading expectancy estimates the average R produced per trade: (win rate × average win) − (loss rate × average loss). Compare expectancy by setup to find quality independent of frequency.

4. Win rate and payoff ratio

Win rate alone is incomplete. A 38% win-rate trend setup can be profitable if winners are much larger than losers, while a 75% scalp setup can fail when occasional losses and fees overwhelm small wins. Read win rate beside average win, average loss, and expectancy.

• Count tells you how much evidence exists.
• Total R tells you contribution.
• Expectancy tells you average quality.
• Win rate and payoff explain how the edge behaves.

Worked example: one account, three very different setups

Imagine a 90-trade sample with +9R overall. The account looks profitable, but setup tags reveal the engine underneath:

• Opening-range break: 35 trades, +14R total, +0.40R expectancy, 46% win rate.
• VWAP fade: 30 trades, +1.5R total, +0.05R expectancy, 60% win rate.
• Impulse breakout: 25 trades, −6.5R total, −0.26R expectancy, 52% win rate.

The overall +9R hides a negative setup. It also shows why win rate is misleading: impulse breakout wins more often than opening-range break but loses money because its winners are too small or its losers too large. The next step is not immediately deleting the tag; split clean executions from mistake-tagged executions.

If clean impulse breakouts remain negative across a sufficient sample, reduce or retire the setup. If clean trades are positive and late-entry trades create the loss, preserve the setup and change the execution rule.

Compare clean expectancy with all-in expectancy

For each primary setup, calculate two views. “All-in” includes every trade carrying the setup tag. “Clean” excludes trades with mistake tags. The gap estimates the cost of execution.

• All breakout trades: 40 trades at +0.08R expectancy.
• Clean breakout trades: 29 trades at +0.31R expectancy.
• Late-entry breakout trades: 11 trades at −0.53R expectancy.

This result does not call for a new breakout strategy. It calls for a rule that blocks late entries. Tag analysis prevents the common mistake of changing entries, stops, and targets when the actual leak is process.

Segment setups by session and market condition

After a setup has enough trades, add one context dimension at a time. Compare breakout + London with breakout + New York, or trend + high-volatility with trend + range. Do not stack five filters on a small sample; almost any story can look convincing when only two trades remain.

For a visual setup such as a liquidity sweep, tag the level type and confirmation model separately. That prevents prior-day sweeps, equal-high sweeps, and random wick reversals from being treated as one strategy.

1. Begin with the primary setup across all conditions.
2. Require a useful sample before splitting it.
3. Apply one context filter and compare count, total R, and expectancy.
4. Write the hypothesis before inspecting more combinations.
5. Test the rule prospectively on future trades instead of trusting only the historical slice.

How much sample size is enough?

There is no universal number because strategy variance differs, but fewer than 10 trades is usually anecdotal. Around 30 on-plan trades per setup is an initial checkpoint, while 50–100 often provides a more stable view. High-variance, correlated, or low-win-rate systems may require much more; see how many trades a strategy test needs.

Sample quality matters as much as count. Thirty trades collected across inconsistent rules are not one sample. If stop placement, market, timeframe, or entry criteria changed materially, mark the version or begin a new sample rather than averaging incompatible strategies.

Weekly tag review workflow

1. Make sure every trade from the week has one primary setup tag.
2. Filter each active setup and note count, total R, and expectancy.
3. Filter mistake tags and calculate their total R cost.
4. Open the highest-impact winner, loser, and mistake—not every chart.
5. Choose one setup to continue observing and one execution behavior to reduce.

Combine this with the 30-minute weekly trade review and the trading diary calendar. Tags explain which category drove the week; the calendar explains when the behavior clustered.

When to keep, pause, or retire a setup

• Keep: positive clean expectancy, adequate sample, and rules executed consistently.
• Observe: promising or weak result with too few trades to trust.
• Pause: rules are changing, execution is inconsistent, or market conditions no longer match the hypothesis.
• Retire: sufficiently large on-plan sample remains negative after fees and realistic execution.

Do not increase size merely because a tag leads this month. Require stable execution, enough trades, and performance that is not dominated by one outlier. Tag analytics guide attention; they do not remove uncertainty.

Common trade-tagging mistakes

• Using a different tag name for the same setup.
• Applying several primary setup tags to one trade.
• Tagging winners carefully while leaving losers unclassified.
• Using outcome labels such as good-trade or bad-trade instead of observable criteria.
• Combining setup, session, and result into one tag.
• Creating so many tags that every sample stays tiny.
• Judging a tag only by win rate or only by total P&L.
• Changing tags after seeing results to protect a favored strategy.

How Traderizz helps with setup analytics

[Traderizz](/) keeps tags connected to the same trade history used for overview analytics, R-multiples, and trader diary review. Filter the journal by a setup or mistake tag, compare the resulting performance, and move between individual trades and aggregate patterns without rebuilding spreadsheet pivots.

Start with one primary setup tag on every new trade. At the end of the week, compare that tag’s count, total R, and expectancy with the full book. Add context only when you have a specific question and enough trades to answer it. A small, consistent tag system will teach you more than a large taxonomy you stop maintaining.