- The mechanism that makes levels work, and when it doesn't
- Why support becomes resistance after a break
- How to rank levels by significance instead of drawing dozens
Levels are the most-drawn and least-understood objects in technical analysis. The mechanism is entirely about human positions, and once you see it, you stop drawing thirty lines on a chart.
The mechanism
A stock trades sideways around ₹400 for four months. Enormous volume changes hands there — thousands of participants build positions at roughly that price.
Then it falls to ₹300. Every one of those buyers is now down 25%, and a large fraction of them have made the same private promise: if it gets back to what I paid, I'm out.
Price returns to ₹400. A wall of supply appears — not because ₹400 is a special number, but because a large, identifiable cohort wants out at exactly that price. That is resistance. It is a queue of people, not a line on a chart.
Now invert it. If price breaks decisively above ₹400, that cohort is finally whole and stops selling. The people who sold too early want back in. The next time price falls to ₹400, buyers appear — the same level is now support.
A level is significant in proportion to how much volume traded there and how many participants are trapped by it. That is why polarity flips: the trapped cohort changes sides when price passes through.
price
│ ╭──────
│ /
│ ═══════════════╤═══════════╤═══════════╪═══════════ ← the level
│ heavy volume │ │ │ ↑
│ traded here │ RESIST │ RESIST │ SUPPORT
│ ↓ \ / \ / now holds
│ \_______/ \______/
│
└──────────────────────────────────────────────────────────► time
Phase 1: a cohort builds positions at the level
Phase 2: price falls; they are trapped; rallies back are sold into
Phase 3: price finally closes above; the trapped cohort exits or turns
buyer; the level now attracts demand instead of supply Ranking levels instead of drawing them all
- Volume transacted thereWhy it strengthens the level: More trapped participants — a bigger queue on the return
- Number of touchesWhy it strengthens the level: Repeatedly defended means the counterparty is real, not incidental
- Time spent at the levelWhy it strengthens the level: Four months of consolidation traps far more people than a two-day spike
- AgeWhy it strengthens the level: A multi-year level survives every intervening participant. Levels from last week are noise.
- Reaction size on the last touchWhy it strengthens the level: A violent rejection means a large counterparty acted, not just a few orders
- Round numbers and prior all-time highsWhy it strengthens the level: Behavioural clustering of orders — weaker, but genuinely present
Real levels are bands, usually a percentage or two wide, because the original activity happened across a range of prices rather than at one tick.
Treating a level as an exact number produces two errors: being stopped out by a wick that only clipped the band, and refusing an entry because price came within 0.4% of your line and turned. Draw the band from the actual highs and lows of the consolidation that created it.
Illustrative, but the ranking logic is exactly what the Chart DNA engine on this site applies.
- ₹412 — 6 touches over 3 years, heavy volume
- Tier 1
- ₹368 — prior all-time high, broken 8 months ago
- Tier 1
- ₹455 — 2 touches, both last quarter
- Tier 2
- ₹390 — one spike low, single bar
- Tier 3 — ignore
Everyone can see the obvious level, and a large fraction place stops just beyond it. That creates a pool of resting orders — and a temporary poke through the level triggers them, generating exactly the liquidity a large buyer needs.
This is the mechanism behind springs and stop-runs, covered in detail in the Wyckoff module. Two defences: give the level a band rather than a tick, and judge breaks on closes rather than on touches. Both cost you a little on genuine breakdowns and save you repeatedly on false ones.
Dynamic levels
- Moving averages act as levels because so many participants watch them — the 50-day and 200-day especially. Their power is behavioural, not physical, but it is measurable: our DNA engine tracks each stock's own 200-DMA respect rate, and it varies enormously between instruments.
- Trendlines are levels that move. They are more subjective than horizontals — two people will draw them differently — so treat them as secondary evidence, never as the primary structure.
- Prior day / week / month high and low matter intraday and for short-swing work, since they are the reference points everyone shares.
- Volume-by-price shows where volume actually transacted rather than where you think it did. It is the most direct expression of the mechanism this whole lesson is built on.
- ☐Mark only levels with volume, touches or age behind them — three or four per chart.
- ☐Draw them as bands taken from the actual consolidation, not as single prices.
- ☐Note which are flipped levels (old resistance now support, or the reverse).
- ☐Define, before entering, what a close beyond the level would mean.
- ☐Expect the level to be pierced before it holds — size and place stops accordingly.
In one screen
- 1Levels work because trapped cohorts create predictable supply and demand at known prices.
- 2Polarity flips when price closes through: the trapped group changes sides.
- 3Rank by volume, touches, time and age — a working chart has three or four levels, not twenty.
- 4Use bands and closes; expect piercing, because that is where stops sit.
Check yourself
Nothing is scored or stored — the explanation after each answer is the point.
A stock consolidated for five months around ₹250, then fell to ₹180. It is now back at ₹248. What should you expect and why?
Why should a break of a level be judged on closes rather than touches?