Support and resistance: why price levels form and how to use them

Support is a price area where falling prices have repeatedly stopped and turned up. Resistance is a price area where rising prices have repeatedly stopped and turned down. They are the most widely used ideas in technical analysis, and the reason is simple: they are where traders' past decisions are written into the chart. A level that held three times is a level where buyers outnumbered sellers three times, and many of the same people are still watching it.

This guide explains why levels form at all, how to find and rank them, why a level is better treated as a zone than a line, how support turns into resistance and back (role reversal), why round numbers matter more than they should, and how to tell a level that is still working from one that has failed. Two real examples run through it: SPY's struggle with $700 in early 2026 and a support zone in Microsoft that held across 15 months.

What support and resistance are

Picture a stock that fell to $50 in March, bounced to $60, fell back to $50 in May and bounced again. The $50 area is support: twice, buying there was strong enough to absorb all the selling. If the stock then rallies to $70 and stalls there twice, $70 is resistance. Between them the stock is in a range; above $70 or below $50, it is in new territory where the old reference points no longer apply.

The concept works the same in an uptrend. A rising market leaves behind a staircase of old highs, and those old highs often act as support on the way back down. A falling market leaves a staircase of old lows that tend to act as resistance on the way back up. That is the link to Trends and trendlines: how to identify a trend and draw it correctly: a trend is a sequence of support and resistance levels that keep being moved.

Levels can be horizontal (a price) or dynamic (a line that moves, such as a trendline or a moving average). This guide is mostly about horizontal levels, which are the most objective to find.

Why levels form

Prices are set by orders, and orders are placed by people who remember prices. There are four reasons the same price keeps mattering.

1. Memory and regret

Suppose a stock spent weeks trading near $100, then fell to $80. Many holders who bought near $100 are sitting on losses. Research on the disposition effect (Shefrin and Statman, 1985, and many studies since) finds that investors are reluctant to sell at a loss and keen to sell once they are back to even. When the price returns to $100, some of those holders sell to "get out flat." That selling is resistance. The mirror image creates support: traders who missed buying at a low often place orders to buy if it comes back.

2. Orders cluster at visible prices

Limit orders, take-profit orders and stop-loss orders are not spread evenly across prices. They pile up at prices people can see on a chart: prior highs and lows, round numbers and the edges of recent ranges. Carol Osler at the Federal Reserve Bank of New York studied the order book of a large currency dealer and found take-profit orders clustered heavily at round numbers, with stop-loss orders clustered just beyond them (Osler, 2000 and 2003). That pattern explains two features of charts: why prices tend to stall at a round level, and why they often accelerate once it gives way, as clustered stops are triggered.

3. Volume leaves a footprint

A price at which a lot of shares changed hands is a price at which many participants hold a position at a known cost. The more volume traded in an area, the more people have a stake in what happens when the price returns. This is the idea behind volume-at-price studies (see Volume analysis: confirmation, climax volume and relative volume): the heaviest-volume areas of a range behave as the strongest levels.

4. Self-fulfilment, within limits

Because so many traders watch the same obvious levels, their orders help make the levels work. This is a real effect, not a criticism: a level matters partly because it is watched. It also has a limit. The more obvious a level, the more stops sit just beyond it, and the more attractive it becomes to push the price through it.

How to find support and resistance

Start with the highest timeframe you care about and work down. A level visible on a weekly chart outranks one visible only on a 15-minute chart.

1. Prior swing highs and lows. The peaks and troughs of the past few months are the first candidates. Mark the obvious ones, the ones you would see from across the room.
2. Areas of congestion. Where price moved sideways for weeks, the top and bottom of that range are levels, and the middle often acts as a magnet.
3. Gap edges. The prior close before a large gap and the first price after it are frequent reaction points (see Chart gaps: common, breakaway, runaway, exhaustion and island gaps).
4. Round numbers. $50, $100, $500 and $1,000 for stocks; 5,000 or 40,000 for indexes. More on these below.
5. 52-week and all-time highs and lows. Widely reported, so widely watched. An all-time high has no overhead resistance from earlier buyers, which is part of why stocks making new highs can run.
6. Retracement levels. Many traders also watch Fibonacci retracements: levels, extensions and real evidence of the last big swing.
7. Dynamic levels. The 50-day and 200-day moving averages and rising or falling trendlines.

Then rank them. A practical way is to score each level on the factors in the next section, keep the three or four strongest above and below the current price, and delete the rest. A chart with fifteen horizontal lines has no levels, only noise.

How strong is a level?

There is no formula, but technicians largely agree on what makes a level more important.

One factor cuts both ways. Each test of a level shows it is watched, but each test also fills some of the orders defending it. A level that has been hit five times in two weeks, each time with a smaller bounce, is often closer to breaking than one hit twice. Watch the size of the reactions, not just their count.

Zones, not lines

Price levels are areas where orders cluster, not single prices. Orders at "around $700" sit at $698, $699.50 and $700.25, and different traders draw the same level through wicks, bodies or closes. The result is that a well-known level usually shows up as a band, and price often reverses a little short of it or pokes a little past it.

SPY in early 2026 shows this clearly.

Between December 24, 2025 and February 26, 2026, SPY's daily high reached $690 or more on 30 sessions. Not one of them touched $700. The two highest were $697.84 on January 28 and $697.14 on February 11. A trader waiting for "$700 resistance" to be tested would have seen it fail again and again without the price ever getting there. Drawn as a band from $690 to $700, the level was obvious.

How wide should a zone be? A useful starting point is the stock's typical daily range, for example a fraction of its Average True Range (ATR): volatility, stops and position size. For a broad index ETF a zone of 1% to 1.5% is common; for a volatile single stock, several percent. The MSFT example below uses a band about 4% wide.

Role reversal: support becomes resistance

The best-known property of support and resistance is that they swap roles once broken. Edwards and Magee described it in 1948, and it is sometimes called the polarity principle. Old resistance, once cleared, tends to act as support on the next pullback. Old support, once broken, tends to act as resistance on the next rally.

The reasons are the same as above, reversed. Traders who sold at resistance and watched the price run away from them often want to buy back on a dip to the same price. Traders who bought the breakout defend their entry. Holders who bought at old support and watched it break are glad to sell at breakeven when the price comes back.

SPY's $700 area is a textbook case. After the February top, SPY fell to $629.28 on March 30, about 10% below the January high. It then rallied straight back into the zone and made its first close above $700 on April 16, 2026 ($701.66). The next two pullbacks stopped at $702.64 on April 21 and $702.28 on April 23, just above the old ceiling. By June 2 SPY traded at $760.40, about 8% above the breakout close.

Role reversal fails, too. When a broken level gives way again on the retest, the breakout was probably false. The guide to Breakouts and false breakouts: what the data says about each looks at how often that happens.

Round numbers

Round numbers matter for a reason that has nothing to do with the company: people think in them. Investors set price targets at $100, analysts write about "Dow 40,000," and order entry screens make round prices quick to type. Osler's currency data showed take-profit orders concentrated at round numbers, especially those ending in 00.

Two practical consequences follow. First, prices often stall just before a round number, as SPY did under $700, because orders placed slightly early fill first. Second, a break of a round number can accelerate, because stop orders cluster just beyond it. The Dow Jones Industrial Average first closed above 10,000 on March 29, 1999, after weeks of headlines about the milestone; round index levels draw the same attention today.

Round numbers are weaker on their own than levels backed by real swing highs and lows. They are most useful as confluence: when a prior high, a moving average and a round number fall in the same zone, that zone deserves extra weight.

A support zone that held for 15 months: Microsoft

Single levels can matter for a long time. Microsoft's chart over two years shows a support zone between roughly $345 and $360.

• April 2025. In the market sell-off of early April, MSFT made a low of $344.79 on April 7 and $355.67 on April 21. It then climbed to $555.45 by July 31, 2025.
• March 2026. After a decline of about 36% from that peak, MSFT fell back into the same area, with lows of $356.51 on March 27 and $356.28 on March 30, 2026, and bounced.
• June 2026. It returned once more, to an intraday low of $349.20 on June 25, closed at $352.83, and rose 5.7% the next day. By late October 2026 it traded near $530.

Three points matter here. The zone was tested months apart, which makes it more significant than three tests in a week. Each test reversed sharply. And the exact low differed by about $12 between tests, which is why a band, not a line, describes it. Nothing guaranteed the third test would hold; a close below $345 would have broken a level the whole market could see.

When a level fails

Levels break all the time. The question is whether a break is real. Most traders look for some combination of:

• A close beyond the zone, not an intraday poke. Intraday wicks through a level that close back inside it are common, especially around news.
• Distance. A close 1% or more beyond the zone edge, or more than a set fraction of the Average True Range (ATR): volatility, stops and position size, filters out small violations.
• Volume. A break on heavier-than-usual volume suggests real participation. A break on thin volume is more often reversed.
• Follow-through. A second close beyond the level, or a retest from the other side that holds.

A useful mental model: a level has a fixed amount of orders defending it. Each test draws them down. When they run out, the level breaks, and the clustered stops beyond it can make the first move through it unusually fast. That is also why you should expect occasional false breaks, where price pushes through, triggers the stops, and returns.

Using support and resistance in practice

Support and resistance are most useful for planning, not prediction.

• Context for entries. Buying into strong resistance or selling into strong support means working against the most crowded side of the market. Many traders wait for price to reach a level and show a reaction, such as a reversal candle (see How to read candlestick charts: a guide with real data), before acting.
• Placing exits. A stop placed just beyond a zone is placed where the trade idea is proven wrong. Placing it exactly at the obvious level puts it among everyone else's stops. The Stop-loss placement: volatility, structure and time stops guide covers the tradeoffs.
• Sizing targets. The distance to the next level in the direction of the trade is a realistic first target, which feeds directly into a Risk-reward ratio and R multiples: the math of a trade calculation.
• Reading the trend. In an uptrend, old resistance should hold as support on pullbacks. When it doesn't, the trend is weakening.

Common mistakes

• Too many lines. Every chart has dozens of minor highs and lows. Keep the few that matter most on the highest timeframe you use.
• Precision you don't have. Expecting a reversal at exactly $100.00 leads to stops that are too tight and entries that never fill. Think in zones.
• Ignoring the trend. In a strong uptrend, resistance levels break more often than they hold, and support levels hold more often than they break. A level's odds depend on the direction of the larger move.
• Hindsight. It is easy to find "perfect" levels on an old chart. Draw your levels before the test, write them down, and check how they actually performed.
• Forgetting adjustments. Splits and large dividends change historical prices on adjusted charts. A level drawn on unadjusted prices before a split will be in the wrong place afterwards.

Support and resistance compared with related tools

How Finabase helps you find and mark levels

The chart on every symbol page has a horizontal line tool for single levels and a rectangle tool that is ideal for zones: draw it from the lowest to the highest reversal in the cluster and extend it to the right. Drawings are anchored to bar dates and prices, so they stay in place as you zoom or change the range, and they are saved per symbol when you are signed in. Try marking the SPY $690–$700 zone or the MSFT $345–$360 zone yourself.

Each symbol's overview shows a 52-week range bar with the current price inside the year's high and low, two of the most watched levels for any stock. The Scanner lists the stocks closing at a new 52-week high or 52-week low each day, and the stocks whose close has just crossed one of the standard moving averages that many traders use as dynamic support and resistance.

Frequently asked questions

What is the difference between support and resistance?

Support is a price area below the current price where buying has previously stopped declines. Resistance is a price area above the current price where selling has previously stopped advances. The same price can be either, depending on which side of it the market is trading.

How do you know if support or resistance is strong?

Look for several clean reversals spread over weeks or months, heavy volume at the level, sharp reactions away from it, and visibility on weekly charts. Confluence with a round number or a major moving average adds weight. A level that keeps producing smaller bounces is weakening.

Why does broken resistance become support?

Traders who sold at the old ceiling and missed the move want to buy back at the same price, breakout buyers defend their entries, and orders placed around the level now sit below the market. Together they create demand when the price returns. It doesn't always happen; a failed retest is a warning that the breakout was false.

Are round numbers really support and resistance?

Often, yes, because orders cluster at them. Federal Reserve Bank of New York research on currency orders found take-profit orders concentrated at round numbers and stop-loss orders just beyond them. Round numbers are most reliable when they coincide with a real prior high or low.

Should I place my stop exactly at support?

Usually not. The obvious level is where many other stops sit, and prices often dip just through it before reversing. Placing the stop beyond the whole zone, at a distance the stock's normal volatility doesn't reach, is the more common approach. Size the position for that distance.

Summary

Support and resistance are the prices where the market's past decisions cluster: old highs and lows, heavy-volume areas, round numbers and watched averages. They form because people remember prices and place orders around them, they behave as zones rather than lines, and once broken they often trade places. SPY's 30 failed tests just under $700 and its quick retests just above it after April 16, 2026 show both the zone and the role reversal; MSFT's $345–$360 floor shows how long a level can stay relevant. Use levels to plan where to act and where you would be wrong, and always have a plan for the break.