Loading market quotes…

Subscribe free
Market Fieldnotes11 minFreeGBC Research

Support and resistance: zones, breaks and false breaks

In short

Support and resistance are price zones where, in the past, declines stopped because buyers stepped in (support) or rallies met selling (resistance). They work better as bands than as lines: three lows at $48.80, $49.10 and $48.95 draw a zone $0.30 wide, 0.6% of the price. Whether the zone has broken is decided by a rule written before you look at the session, for example a close below the lower edge.

On January 18, 1966, the Dow Jones Industrial Average rose above 1,000 for the first time during the trading day, then closed below it. The first close above 1,000 took almost seven years: November 14, 1972, at 1,003.16. On January 11, 1973, the index reached 1,051.70, and then the 1973-74 bear market began. That high wasn’t topped until November 3, 1982.

For sixteen years the roundest number on Wall Street acted as a ceiling, and the ceiling had thickness. The index pierced it intraday, closed above it, fell back below it, and the cycle’s high came 5% above 1,000. Anyone who drew an exact line at 1,000 would have counted two breaks, both reversed. Anyone thinking in zones saw a single band of resistance between roughly 1,000 and 1,050.

The same thing happens on a single stock’s chart on a smaller scale, and it changes where you put your stop, how you size the position and how you judge whether a level held.

What are support and resistance?

Support is a price zone where declines have stopped in the past because enough buying arrived to absorb the selling. Resistance is a price zone where selling has been sufficient to halt previous rallies. Both are observations about trades that already happened, and nobody guarantees the same reaction will repeat.

There is, though, a mechanical explanation for why some levels matter. Carol Osler, an economist who also studied the subject for the Federal Reserve Bank of New York, published an analysis of actual stop-loss and take-profit orders in the currency market in the Journal of Finance in 2003. Take-profit orders clustered especially at round numbers, and stop-loss orders just beyond them. The first cluster helps explain why prices tend to stall at certain levels; the second, why they tend to run once those levels give way.

In earlier work, published by the New York Fed in 2000, Osler found that support and resistance levels issued by six currency-market firms anticipated intraday trend interruptions, with results that varied widely across currencies and firms. This is currency data: for stocks it describes a plausible mechanism, which you still have to check on the stock you follow.

Why does support work better as a zone?

Suppose a stock stops falling three times, at $48.80, $49.10 and $48.95, on different sessions. If you draw an exact line at $49.00, two of the three lows sit below it: with that line you would have logged two breaks, both reversed within the same session. If you draw a zone from the lowest low to the highest, from $48.80 to $49.10, the three episodes become three touches of the same area.

Three lows, two readingsThree lows, two readings. The same sessions read with an exact line and with a zone as wide as the observed lowsFIG. 1 · LINE OR ZONEGB.CAPITALThree lows, two readingsThe same sessions read with an exact line and with a zone as wide as the observedlowsLine at $49.0048.8049.1048.952 “breaks” in 3 touches, both of them falseZone $48.80–49.1048.8049.1048.953 touches, no breaks. Width $0.30, or 0.6%Illustrative chart with hypothetical prices, not a real stock.
Figure 1 With an exact line two of three touches look like breaks; with a zone as wide as the lows they are three touches of the same area.

The zone keeps the information the data actually contains: buyers interested across an area 30 cents wide, 0.6% of the price. The $49.00 line adds precision nobody gave you. The 49 is still a handy reference, because it is round and therefore probably crowded with orders; the exact edge to the cent is something the person drawing invents.

How do you draw a support or resistance zone?

Pick the timeframe before the level

A zone on a weekly chart and one on an hourly chart are different objects, built on different observations. Work on the timeframe you make decisions on and check the one above it for context, as we explain in the guide to daily and weekly trends.

Use the candles’ lows and highs

For support, take the lows of the sessions where the decline stopped; for resistance, the highs. If you need a refresher on wicks and bodies, see our guide on how to read candlestick charts. The lower edge is the lowest low and the upper edge the highest low. Then compare the zone’s width with the stock’s typical daily range: if the zone is wider, you are probably joining episodes that have nothing to do with each other.

Count the touches and note the dates

Write down how many touches you have and when they happened. Three lows weeks apart describe steadier interest than three lows on three consecutive days, which may be a single pause in the price.

Freeze the drawing

Save the chart with its date. From then on the zone stays where it is: if price goes through it, you log that as it is, without widening the edges to make it fit.

When is support considered broken?

Back to the example stock. After the three touches, a session drops to $48.70, 10 cents below the edge of the zone, and closes at $49.30. Before you look at the table, write your answer down: is support broken? Yes or no, and why.

Did support hold? It depends on the ruleDid support hold? It depends on the rule. Three hypothetical months of daily sessions, with both zones drawn before the test sessionFIG. 2 · SUPPORT AND RESISTANCE AS ZONESGB.CAPITALDid support hold? It depends on the ruleThree hypothetical months of daily sessions, with both zones drawn before the testsession$46$48$50$52$54$56Resistance $53.60–54.20 · two touchesSupport $48.80–49.10 · three touchesTest session: low $48.70,close $49.30Illustrative chart with hypothetical prices, not a real stock.
Figure 2 The test session dips below the zone and closes above it: broken or held depends on the rule written beforehand.

The answer depends on the rule you wrote before the session. Here is the same day judged by five reasonable rules.

The session with a $48.70 low and a $49.30 close, judged by five rules
Rule written beforehandVerdictNote
Line at $49.00, any intraday cross countsBrokenUnder this rule the $48.80 and $48.95 lows were breaks too
Line at $49.00, the close countsHeldClosed 30 cents above the line
Zone $48.80-49.10, any cross below $48.80 countsBrokenBy 10 cents, or 0.2%
Zone $48.80-49.10, a close below $48.80 is neededHeldThe session becomes a false break
Zone, two consecutive closes below $48.80 are neededHeldFilters more, and flags real breaks late too

None of these rules is right in absolute terms. Intraday rules react sooner and get fooled more often; closing rules filter out noise and, when the break is real, get you out lower. What makes the work useless is changing the rule after seeing the session. If support “holds” when it suits you and “breaks” when it suits you, you aren’t measuring anything.

What is a false breakout, and why does it happen?

A false breakout is a move beyond the edge of a zone that is quickly reversed, often within the same session. The order clustering Osler described helps explain it. Just below a visible support sit the sell stops of everyone who bought at that support. If price reaches them, the stops turn into market orders and push price lower for a few minutes; once they are used up, the buy orders resting in the zone remain, and price can climb back above.

Orders pile up inside the zone and just outside itOrders pile up inside the zone and just outside it. Schematic of a support near $49: order sizes are illustrativeFIG. 3 · WHY FALSE BREAKS HAPPENGB.CAPITALOrders pile up inside the zone and just outside itSchematic of a support near $49: order sizes are illustrativeRESTING ORDERS (SCHEMATIC)48.4048.5048.6048.7048.8048.9049.0049.1049.2049.3049.4049.5049.60Limit buys andprofit-taking: they slowthe declineSell stops just belowthe zone: oncetriggered, they speed itupTEST SESSION$48.70: stops triggerclose $49.30Hypothetical schematic. Clustering of take-profit orders at round numbers and of stops just beyond themis documented in currency markets by C. Osler, Journal of Finance, 2003.
Figure 3 Stops sitting just below the zone explain why price pokes through it and then comes back.

The practical problem is where to put your stop. Say you buy at $49.50 and want to lose no more than $100 if the idea is wrong. The table compares two stops using hypothetical numbers.

Hypothetical buy at $49.50 with a planned maximum loss of $100
Stop A, just below the zoneStop B, about 1% below the edge
Stop level$48.75$48.30
Risk per share$0.75$1.20
Shares for $100 of risk13383
Session with a $48.70 low and $49.30 closeStopped out: −$99.75, plus any slippageStill open: −$16.60 at the close
Gain if the stock reaches $53.60$545.30, had it not already been stopped out$340.30

Stop A sits immediately below the visible support zone, and the test session takes it out before turning back up. Stop B survives, but with a smaller position: 83 shares instead of 133, and a potential gain 38% lower if the idea works. Widening the stop cuts the share count and leaves the dollar risk you chose unchanged; the full calculation is in our guide to position size.

Neither stop protects you from a price jump. If after an earnings report the first executable price is $46, the old touches at $49 don’t create a buyer at $49: the stop becomes a market order and fills wherever it finds a counterparty, as we describe in the guide to earnings gaps.

How do you test support without fooling yourself with hindsight?

On today’s chart every level that worked looks obvious, and with enough lines you can explain any historical chart. To know whether your zones are worth anything, you have to draw them using only the data available at a given date, then reveal the following sessions one at a time and note what happens.

Log every case, including the ones where price goes through the zone and keeps going. A folder of clean bounces can’t tell you how often your definition failed, or whether the result would have been tradable after commissions and spreads. On thinly traded stocks those costs weigh heavily, which is the subject of the guide to volume, liquidity and spreads.

A tidy workspace makes the exercise sustainable: dated screenshots, the same interval, the same price-adjustment setting. We work on a TC2000 setup we refined over about two years, but the log below works on any platform.

Frequently asked questions

How do you draw support and resistance?

Pick the timeframe, then mark the lows of the sessions where the decline stopped (or the highs, for resistance). The lowest and the highest of those lows become the edges of the zone. Save the chart with its date and write down in advance what you will count as a break.

How many times does support need to be tested to be valid?

No number of touches makes a support valid. Two touches are enough to draw a zone, and each extra touch, especially one far in time from the others, adds an observation. Logging the cases where the zone gave way matters more, so you know how often it happens.

What is a false breakout?

It is a move beyond the edge of a zone that is quickly reversed, often within the same session: for example, a stock that drops to $48.70 below support at $48.80 and closes at $49.30. One possible explanation is sell stops clustered just below the level, which push price down briefly and then run out.

Should you count a break intraday or on the close?

An intraday rule reacts sooner and gets fooled more often by false breakouts; a closing rule filters noise but, when the break is real, gets you out lower. Choose one before you look at the chart and keep it for every zone you test.

Does broken support become resistance?

It is one of the most described behaviors in the textbooks, and it has a plausible explanation: people who bought at support and are sitting on a loss tend to sell when price returns to their break-even point. Treat it as a hypothesis to test: after a confirmed break, mark the old zone as possible resistance and log what happens on the first return.

Do support and resistance actually work?

Carol Osler’s research on the currency market found that levels published by professional firms anticipated intraday trend interruptions, with results that varied widely across currencies and firms. For stocks, the answer has to come from your own data: a fixed rule and a log of every case, false ones included.

Sources and further reading

Related research
Daily vs weekly chart: multiple timeframe analysisMarket Fieldnotes · 05.10.2026Bid-ask spread, volume and liquidity: how to read the costMarket Fieldnotes · 05.10.2026How to review a losing trading week before changing your strategyMarket Fieldnotes · 28.10.2026How to read candlestick charts: an OHLC exampleMarket Fieldnotes · 05.10.2026Stop-limit order vs stop order: what each one guaranteesMarket Fieldnotes · 05.10.2026 Full archive →
FREE · EVERY SUNDAY

Get the next Radar
before Monday’s open.

The first name, the reasoning, and our public research.

FREE · EVERY SUNDAYGet the watchlist