"Uptrend" is not a vibe. On a price chart it is a claim about swing points: higher highs and higher lows (HH/HL), or the inverse (LH/LL). The claim is easy on a finished daily chart of a liquid large-cap. It is unpleasant on a 5-minute NSE replay at 10:10, when the last swing high is only a few bars back and the next print might cancel it.
Most structure content skips the grey zone — the stretch where swings are not clean enough to label without cheating. That is where traders force a trend so they can take the trade they already wanted. Labelling rules exist to make the grey zone say skip, not to decorate a screenshot.
A swing definition you can repeat
Before you talk HH/HL, write how a swing is born. Example (freeze yours; do not mix with a neighbour's):
- A swing high is a bar whose high is higher than the N bars on either side (pick N, often 2 or 3 on your timeframe).
- A swing low is the mirror.
- Swings are marked only on closed bars.
- You may not relocate a swing because the story looks better after three more candles.
N is not magic. It is a resolution setting. Small N = more swings = more labels = more chances to lie. Large N = fewer swings = more "unclear." Pick N for the timeframe you are practising and keep it for the block.
This is structure as labelling, not as a forecast. Trend versus mean reversion is a later choice about what you do with the label.
Grey-Zone Label Rules (GZL)
At each pause in replay, assign exactly one market-state tag.
| Tag | Observable rule (example — write yours) |
|---|---|
up |
Last two confirmed swings: HH and HL |
down |
Last two confirmed swings: LH and LL |
range |
Swings overlapping a horizontal band you marked in advance; no HH/HL sequence |
grey |
Anything else: one-sided break without a second swing, equal highs, expanding both ways, N not yet satisfied |
Grey is not a failure to see. Grey is the honest output when the definition does not fire.
Actions tied to tags (example stub):
up— only long location setups, or skip shortsdown— mirrorrange— range practice rules, or skip if you do not have a range stubgrey— no new risk
If your stub allows trades in grey, you do not have a structure filter. You have a caption.
Checklist: labelling before the next candle
Print this beside the replay.
- N is written; I have not changed it this session
- Last swing high/low marked on closed bars only
- I can point to the two swings that justify
upordown - If I cannot, the tag is
rangeorgrey, not "probably up" - Location for any trade was marked before this segment, not on the impulse I just watched
- Invalidation is a price relative to a swing or level, not "if structure breaks" as poetry
- NSE window logged (
open/midday/last-hour) because 9:15–10:00 produces moregreyby construction
Fail any box → no entry. The checklist is the practice.
Failure modes (the ones screenshots never show)
Retcon swings. You delete a swing high after price breaks it, so the uptrend "never broke." Log retcon. Structure practice is over for that session if you do this twice.
One-swing trends. A single higher high after a low is not an uptrend under a two-swing rule. It is a break. Breaks can fail. Tag grey until the second swing confirms, or write a different stub called "break of swing" and stop calling it structure.
Timeframe smuggling. You label up on 15-minute and enter against it on 5-minute without a written lower-timeframe rule. That is two labels. Write both or drop one.
Trendline as substitute. A diagonal that you redraw is not a swing rule. Trendline practice is a separate drill. Do not let a line rescue a grey tag.
Open masquerading as trend. The first 30 minutes after 9:15 often print a directional burst that is inventory, gap fade, or both. You may tag grey by policy until a time cutoff. That is a structure rule about when labels are allowed, not cowardice.
Range denial. Equal-ish highs and lows with overlapping bodies is a range. Calling it "coiling for a trend" is a forecast. Forecasts need a different log field.
P&L-coloured labels. You flip up to down because the long stopped. The swings did not change. The ego did.
Workflow: five sessions of labels only
Hidden-date NSE replay. Liquid names. Timeframe frozen.
Sessions 1–5: No trades. Pause every N bars (or every swing candidate). Apply GZL. Advance. After the session, count tags. If up+down exceed grey+range on a choppy midday window, your N is too eager or you are retconning.
Sessions 6–10: Allow entries only on up or down (or only on range if that is the stub — not both in one block). Process-score 0 if you entered in grey.
A 30-day block can hold one N, one timeframe, and one action map. That is enough to see whether you use grey or paint over it. It is not a claim that HH/HL reading becomes a career.
Price action's Contact Card wants a one-word state. GZL is how that word stays testable. Support and resistance then gives location inside the state. State without location is a lecture. Location without state is a magnet for mid-range clicks.
Label grey on purpose for five hidden-date sessions before you allow a structure trade. Join the waitlist.
Related reading
- Trend Following vs Mean Reversion — Which to Practise First
- Range Trading Practice: When Chop Is the Setup
- Price Action Basics That Survive Contact With a Chart
- Support and Resistance: How to Practise Level Reading
Educational note
Educational practice content only. Not investment advice. No guaranteed returns. Practice ≠ live profitability.