Setup Tagging 101: From 'Breakout' to a System You Can Backtest

A journal with one tag per trade — breakout, pullback, fade — does not actually tag trades. It labels them. Labels are useful for sorting your own memory of what happened, but they're useless for answering the questions a journal is supposed to answer: which of my setups have positive expectancy, which time-of-day buckets bleed, which patterns work in which regime.

The fix is a three-layer tag taxonomy. Every trade gets three tags, one from each layer. The combinations let you filter analytics in ways a single label never can.

Layer 1: Pattern

The pattern layer is what the chart looks like. This is what most traders mean when they say "what setup is this." Keep this layer small — six to ten tags total, no more.

Common pattern tags:

• breakout — price clears a defined resistance level
• pullback — price retraces to a moving average or prior support in an established trend
• vwap-reclaim — price crosses back above VWAP after spending time below it
• fade — counter-trend entry against an extended move
• range-extreme — entry at the boundary of a defined trading range
• gap-fill — entry sized into a measured gap close
• consolidation-break — entry on a break of a tightening range
• flag — bull or bear flag continuation pattern

Ten tags is a lot. Most traders find they actually use four or five over the long run. The unused tags are diagnostic — if you tagged 200 trades and three pattern tags account for 90% of the trades, your strategy is more concentrated than you thought, which is useful to know.

Layer 2: Trigger

The trigger layer is what made you push the button at the specific moment you entered. Pattern is what the chart was doing; trigger is the specific signal that produced the entry. This layer prevents the common journal failure where every trade gets called a "breakout" when half of them were actually breakouts that you entered for different reasons.

Common trigger tags:

• volume-spike — entry on an above-average-volume bar at the pattern level
• l2-imbalance — entry on a visible bid/ask imbalance in the Level 2 book
• news-catalyst — entry on a specific named news event
• time-trigger — entry at a specific time (e.g., 9:35am open-range break)
• confirmation-bar — entry on the bar after the pattern bar closes
• retest — entry on the retest of a broken level

Pattern × trigger gives you a much finer view than pattern alone. A breakout × volume-spike is a different trade than a breakout × confirmation-bar. Both might be tagged breakout in a single-label system; the dual-tag system separates them and lets you see whether one of the two has positive R while the other is a coin flip.

Layer 3: Context

The context layer is what the broader market was doing when you took the trade. This is the layer most traders skip, and it is the layer that explains the most variance in setup expectancy.

Common context tags:

• trend-day-up — broad market trending up cleanly
• trend-day-down — broad market trending down cleanly
• range-day — broad market chopping in a defined range
• vix-elevated — VIX above a threshold you've set (e.g., 25)
• vix-suppressed — VIX below a threshold (e.g., 14)
• earnings-week — major earnings season backdrop
• fomc-day — FOMC announcement day
• sector-strength — the traded ticker's sector outperforming
• sector-weakness — the traded ticker's sector underperforming

Context is the layer that answers questions like "do my breakouts work better in trend days or range days." If you've ever felt like your strategy works great for two weeks and then stops working for two weeks, the answer is almost certainly in the context layer.

Three tag hygiene rules

These three rules make or break the taxonomy. Without them, the tag system rots into the same single-label mess you started with.

Rule 1: Max four tags per trade. Pattern + trigger + context = three. The fourth slot is for one optional discretionary tag — oversized, late-entry, revenge, whatever you want to flag for personal review. Beyond four, tags lose meaning because every analytics filter becomes either too narrow (only five trades match) or too broad (everything matches).

Rule 2: Never invent a tag on the fly. The single fastest way to ruin a tag system is to add a new tag in the moment because the trade "doesn't quite fit any of the existing ones." Two things happen: you create one-trade tags that show up nowhere else, and you contaminate the analytics with an inconsistent taxonomy. If a trade genuinely doesn't fit, tag it other and review at week's end whether the taxonomy needs to grow.

Rule 3: Tag at entry, not at exit. The point of tagging is to record what you thought the setup was at the moment of the trade. Tagging after the exit lets the outcome contaminate the label — winners get tagged generously, losers get retconned into different setups. The discipline is to tag before you know how the trade ends.

How to filter analytics with the taxonomy

Once you have 50-100 trades tagged across all three layers, the analytics tab in the Tapeboard trade journal can answer specific questions you couldn't ask with a single label.

Examples:

• Average R for breakout × volume-spike × trend-day-up. (Probably positive.)
• Average R for breakout × confirmation-bar × range-day. (Probably scratchy.)
• Average R for fade × l2-imbalance × vix-elevated. (Could go either way; this is the bet.)

The point of these queries isn't to optimize against past data — that's curve-fitting. The point is to identify which combinations have showed consistent positive R over a long enough sample to suggest the edge is real, and which combinations are basically noise. The first set gets traded with conviction; the second set gets a hard pass.

What this system is not

This is not a complete strategy. The tags tell you what worked historically. They do not tell you what will work tomorrow. Markets shift, edges decay, and a tag combination with positive R over your last 100 trades may go cold for the next 50. The tag system makes the cold periods visible — that's all it does.

This is also not a substitute for the qualitative notes. The markdown notes field on every journal entry is where you record the things tags can't capture: how the chart felt, what you were thinking, whether anything outside the trade was distracting you. Tags are the quantitative layer. Notes are the qualitative layer. Both matter. Most traders skip the notes; most traders also can't tell you why their summer drawdowns happen.

If you build the three-layer taxonomy and stick to it for a calendar quarter — pattern × trigger × context, max four tags, never invent on the fly, tag at entry — you will end the quarter with the first journal you've ever owned that can answer a useful question. That's the threshold most traders never cross.