Trade Tagging: Finding Which Setups Drag Down the Account
Looking at the account as one number can end with a conclusion such as “−$120 this month.” But that total usually combines several types of trading. One type may erase the gains of another. Tagging trades makes that structure visible.
One total does not identify what to change
A monthly review often produces trade count, win rate and net PnL. Those alone do not identify what needs changing. A 46% win rate offers little direction beyond trying to do better.
Forty trades; eighteen wins; twenty-two losses; win rate 45%
Net PnL +$120
Possible conclusion:
“Near breakeven; I should improve win rate.”
→ What to change remains unclear.
Split the same forty trades by setup and the picture changes. This example uses only two setup tags.
A: Trend-pullback entries, twenty-two trades
Twelve wins; ten losses; win rate 55%
PnL +$430
B: Chasing sharp rises, eighteen trades
Six wins; twelve losses; win rate 33%
PnL −$310
Total = +$120, unchanged.
The guide's proposed action becomes specific: reduce B.
+$120 and “A +$430 / B −$310” describe the same data, but the latter suggests a concrete review target. Removing B retrospectively would leave +$430 for that month. The original guide presents this as improvement through stopping an unhelpful activity, rather than becoming more skilled at every trade.
Five basic tags
More tags are not always better. Keep fields that can be completed within ten seconds after a trade. These five reveal many common patterns.
② Direction: long or short.
③ Session: Asia, Europe or the US.
④ Holding time: under ten minutes, under one hour or longer.
⑤ Rule compliance: followed the plan or changed it mid-trade.
Record immediately after closing.
Filling records days later encourages hindsight rewriting,
such as remembering that a losing trade felt wrong from the start.
Rule compliance can be uncomfortable but informative. Separating planned trades from trades with moved stops or increased size can reveal concentrated losses. The trading-journal guide provides the broader template; this article focuses on breaking those records into groups.
What to calculate for each tag
Win rate alone can mislead. A low-win-rate setup may still contribute strongly if its winners are large. Review at least three measures for every tag.
① Count: sample size.
② Net PnL: actual result including fees.
③ Average PnL: net PnL divided by count.
Example
Breakout entries: twelve trades, 33% wins, +$240 net, +$20 per trade.
Rebound entries: twenty-six trades, 62% wins, −$130 net, −$5 per trade.
The rebound win rate is nearly twice as high,
but breakouts added to the account.
See win rate and reward-to-risk for this relationship, and expectancy and R-multiples to normalize contribution per trade. Calculating profit factor separately for each tag also supports comparison.
When is there enough data? The over-segmentation trap
More detailed categories mean fewer trades per cell. Dividing forty trades among five groups leaves eight per group on average, far too little for a stable win-rate estimate.
Two tags → average twenty per cell, treated as reviewable in the guide.
Five tags → average eight, highly unstable.
Twelve tags → average 3.3, nearly uninformative.
Reference: eight fair coin tosses
Probability of at least six heads ≈ 14%.
Even without an edge, six wins from eight occurs
roughly one time in seven.
Start with broad categories and consider subdividing after a cell reaches around thirty trades. Trade sample size discusses uncertainty, while losing-streak probability helps interpret adverse runs. Keeping only the best-looking tiny group repeats the mistake of overfitting within a trading journal.
Two rules against hindsight and look-ahead contamination
The guide highlights two routes to distorted tags.
Winners become “planned”; losers become “impulsive.”
→ The compliance tag can appear to win 100% by construction.
② Using information unavailable at entry
Tags such as “near the eventual top” or “before the crash”
describe facts that were not yet known.
The guide's rule: tag within thirty seconds after closing,
and use entry information that was knowable at the time.
The second issue is the journal equivalent of look-ahead bias. Later labels can make performance look artificially strong. Measure the eventual maximum favorable and adverse excursion separately through MFE and MAE, rather than treating future movement as an entry tag.
A monthly fifteen-minute breakdown
A spreadsheet is sufficient; elaborate tooling is unnecessary.
Step 1: Sort this month's trades by a tag field.
Step 2: Calculate count, net PnL and average PnL for each group.
Step 3: Isolate tags with negative average PnL.
Step 4: Its proposed rule pauses a tag next month if it has at least twenty trades.
With fewer than twenty, defer judgment and keep observing.
Step 5: Record this month's loss from the paused tag
as a historical reference for the cost avoided by omission.
The emphasis is on stopping a weak activity, rather than necessarily increasing another one. The guide argues that one such omission can alter the account curve. See the skill of waiting and average holding time for related perspectives.
Key points
② The same forty trades split into +$430 and −$310 groups.
③ The guide emphasizes stopping weak activity as a source of improvement.
④ Five tags: setup, direction, session, holding time and compliance.
⑤ Review count, net PnL and average PnL, not win rate alone.
⑥ A low-win-rate tag may still contribute profits.
⑦ Eight-trade cells are unstable; six wins can occur with about 14% probability without an edge.
⑧ Start broadly and subdivide after roughly thirty trades per cell.
⑨ Record promptly and distinguish entry-time information from later outcomes.
⑩ The guide's monthly rule pauses negative-average tags with twenty or more observations.
Separating activity makes specific review possible. An undivided account total can lead to the same vague reflection every month; tags reveal which behavior contributed what.
NOONOO TRADING invites you to follow live trading in our free chat.
Start in the bot📈 OKX trading fee discount for new registrations
Register for the OKX Fee Discount →