Where Price Turned: Levels and Role Reversal

Why price responds where it turned before

A chart is a record of completed transactions, and every completed transaction leaves something unfinished nearby. Someone who bought at $40 with a plan to cut the position below $38 left a sell order at $38 that may still be sitting there. Someone who sold at $60 on the way down and watched the stock rebound to $80 left a buy order at $60 they are still waiting to fill. A market maker who was run over at a level is more careful there the next time. Round numbers attract orders because people place them in fives and tens, and options strikes cluster on them too, so a strike that expires near a price can create genuine hedging flow at that price. None of this is mystical: it is a small set of resting orders, and resting orders are why price behaves a little differently at prices where a lot happened than at prices where nothing did. The consequence is that a level is better understood as a zone than a price. The buyers who want $48 do not all want it at exactly $48.20; they want it anywhere from $47.60 to $48.40, and the reason a chart shows several touches at slightly different prices is that the level is a band of interest rather than a line. That distinction matters practically the moment a stop is involved: a stop placed exactly at the level is inside the zone where price is most likely to wander, so structural stops belong just beyond it. Treating the level as an exact number is what produces the experience of being stopped out "right before" the move resumes. There is a second source of level behaviour that has nothing to do with orders: the level is the price at which a disagreement was settled. Anchor points are how people make comparisons — a high, a low, the price they paid — and the behavioural literature finds that adjustments away from those anchors are smaller than they should be. So a prior high becomes the reference everyone quotes, the level is discussed, and the discussion itself concentrates orders. The mechanism is circular, and that is precisely why it works and why it decays: attention to a level creates the orders that make the level matter, and attention moves on. One level, four kinds of order — Stops from earlier buyers: Sell orders resting just below a prior low — which is why a brief undercut often comes first · Unfilled bids from people who missed the rally: Buy orders at the old ceiling once it becomes support ← · Options hedging: Strike clusters create flow at and around round numbers and listed strikes · Anchored comparisons: The prior high and the entry price are the numbers people quote, so orders concentrate there Every one of these decays. Stops get taken, the people who missed the rally buy elsewhere or give up, options expire and attention moves — which is why recency is part of what makes a level worth drawing.

Role reversal, and drawing fewer lines than you want to

When price closes decisively above a level, the population at that price changes sides. The traders who sold into it have been proven wrong and will want to exit near break-even on the next test, which is buying; the traders who were waiting for a breakout and missed the initial surge want the pullback, which is also buying; and the market makers who sold the level are short from a price they now want to buy back. That is why the first retest of a broken ceiling is one of the more reliable patterns in a chart, and why the level is worth marking *before* the break rather than after. The mirror version happens below: a broken floor becomes the ceiling that caps the next rally, and the rally that fails at the old support is the classic shape of a downtrend continuing. The reversal is not automatic, and the tell is the behaviour on the return. If price comes back to the level and closes through it — especially on expanding volume, and especially if it cannot reclaim within a session or two — the memory has been spent and the correct action is to stop treating the level as support. This is where a beginner loses money with levels: having marked the level as support, they keep buying every touch as it descends, converting a level into an argument. The level is evidence about where orders were, and once price has cleared the zone the evidence has been used up. The other discipline is scarcity. Beginners mark every swing, every round number and every gap, which guarantees that some line will appear to explain any move after the fact — a form of overfitting done with a pencil. The useful test is to ask whether a stranger would notice the same level: it should show repeated rejections, it should be recent enough that the participants who created it are still active, and the move away from it should have been decisive rather than a drift. Three or four levels on a chart is usually the whole picture. Twenty levels is a description of the past with no predictive claim at all, because with enough lines drawn, every price is near one. • A broken ceiling becomes support; a broken floor becomes resistance — the traders who were wrong become the buyers or sellers on the retest. • The retest is the only real test of a break: holding is confirmation, closing back through is refutation. • Draw levels that are repeated, recent and sharply rejected — observable before the trade, unlike your cost basis. • Stops go just beyond the zone rather than inside it, because the zone is where price is most likely to wander. • Scarcity is a quality: a few levels that a stranger would notice, not every swing on the chart. The trap that costs real money is not a wrong level but a level treated as a promise. A chart marked with support before a decline makes it easy to add at every touch — each purchase feels disciplined because a line was drawn in advance. The line is a probability, and the close through it is the update; the level is refuted by evidence, and treating refutation as a discount is how a losing position becomes a large one.

Volume at price: where the trading actually happened

A line drawn across a chart says that price turned at this level. It does not say how much trading happened there, and the amount of trading is what makes a level behave the way it does. The tool that answers the question is a **volume-at-price** profile — sometimes called market profile or a volume profile — which sorts a period’s volume by the price it traded at rather than by the time it traded. The picture is quite different from a bar chart. Where price spent a long time at a narrow range, the profile shows a bulge: a **high-volume node**, an area where a great deal of stock changed hands and where buyers and sellers were broadly in agreement. Where price moved quickly through a range on light volume, the profile shows a thin zone — and that thinness is the important part, because a market that traded little at a price has had little chance to build a consensus there, so price tends to move through it quickly the next time. That gives three things a drawn line cannot. First, the level’s *quality*: a support line sitting on top of a high-volume node has real turnover behind it, while one drawn over a gap in the profile is a line through a vacuum. Second, the **unfinished business** above and below: the low-volume pockets are where the market moved without deciding, and they are the places price revisits fastest — which is a plausible mechanical explanation for why gaps are often filled. Third, the **value area**, the band containing most of a period’s volume, which acts as a range that price tends to revert to when there is no news, and to break away from when there is. The mechanism behind all of it is holding cost and attention. Prices where a lot of stock changed hands are prices where a lot of participants have a cost basis, and cost basis drives behaviour: holders defend it, sellers emerge near it, and stops accumulate around it. Prices where almost nothing traded have no such population, which is why a rally through them can be fast and unattractive to fade — there is nobody there to sell. Two cautions keep the tool honest. It is created from the same trade data as everything else, so it cannot see liquidity that never traded — a quiet price can be quiet because it was never offered. And it is a description of the past distribution of activity rather than a forecast, which means it grades a level rather than predicting a reaction. Used that way — as a way to rank candidate levels by how much business sits behind them — it is a substantial improvement on drawing lines by eye, and it is the same information the earlier reads called the level’s supply and demand story. • Volume sorted by price, not time, shows where the market reached agreement. • High-volume nodes carry turnover and cost basis; low-volume pockets carry neither. • Thin zones are traversed quickly, which is a mechanism behind gap fills. • The profile grades a level; it does not forecast a reaction. A useful pairing: draw the level by hand first, then look at the profile. Where a hand-drawn line lands on top of a high-volume node, the level has business behind it; where it lands in a gap, what you drew was a coincidence of two endpoints.

Sources

• Support, resistance and the role reversal of broken levelsMurphy, “Technical Analysis of the Financial Markets”
• Why price clusters: resting orders and round numbersHarris, “Trading and Exchanges”
• Anchoring on prior extremesBehavioural finance literature on reference points