Home › Learn › Lesson 11

Support and resistance — why levels hold and why they break

By TradingPrimer · Published 29 Aug 2026

Ask ten traders to mark support on the same chart and you will get ten different lines, each drawn with total confidence. That is not a failure of skill. It is what happens when a tool that describes a fuzzy area gets used as though it described an exact price. This lesson does the thing most support-and-resistance guides skip: it puts numbers on the fuzziness. You will finish with a rule for how wide a zone should be, a calculation for what widening it actually costs, and a fixed order for drawing one — so the level does not quietly get redrawn to fit the trade you already wanted.

What are support and resistance actually made of?

Support is a price band where demand previously became strong enough to slow or stop a fall. Resistance is the same thing for supply, on the way up. Neither is a property the asset owns. Each is a record of behaviour at one moment, written in candles after the fact.

It is worth being precise about what creates that behaviour, because the mechanism explains everything else in this lesson. When price arrives back at an area where it previously turned, four different groups of orders tend to sit close together. People who bought there before and want to add. People who missed the first move and left resting limit orders in the hope of a second chance. People who sold there and are waiting to buy back. And — the group that causes the most trouble — stop orders belonging to everyone in the first three groups, parked just beyond the edge of the area.

Those four populations are why an area can absorb a decline. They are also why the area often gets pierced by a few cents before it works: the fourth group is stacked exactly where a hairline drawing would put your own stop. This is the same order-book plumbing covered in Lesson 8, seen from the chart side rather than from the depth ladder.

One honest caveat before any of it is useful. None of these orders are visible to you. You are inferring the presence of resting interest from the shape it left behind on a chart built from opens, highs, lows and closes — the four numbers from Lesson 9. That inference is often good enough to organise a decision. It is never proof, and every rule below is written to survive being wrong.

What does one full cycle look like on a single chart?

Five things happen to almost every level in its life: it gets tested, tested again, swept, broken, and retested. Here is all five in twelve candles, with the exact prices used throughout the rest of this lesson.

Read the sequence rather than the shapes. The first test makes a low at $98.40 and closes back up at $99.10. The second makes a low at $98.75 and closes at $99.25 — a different price, the same behaviour, which is precisely why the honest mark is the band $98.40–$98.80 and not the number $98.40. That band is $0.40 wide, or 0.41% of its $98.60 midpoint.

Then a candle wicks to $98.20, below the band, and closes at $98.95, above it. Nothing was accepted below the zone; price went there, found no takers, and came back. Two candles later a close prints at $97.90 and stays there. That is a different event. Finally price rallies to $98.65 — back inside the old support band — and closes at $98.05, having been turned away once by the area that used to hold it up.

Why is a zone more useful than a single line?

Because the orders that create a level are spread across a range of prices, and because the outer edge of a line is exactly where the stops sit. A line does not just describe the level imprecisely; it puts your stop in the worst available place.

Take two traders with identical analysis and identical risk. Both enter long at $98.85 as price comes back into the band. Both risk exactly $50 on a $5,000 account. The only difference is what they drew.

The trader who marked the line at $98.40 puts the stop ten cents under it, at $98.30. The $98.20 sweep takes them out and price closes the candle back at $98.95 without them. The trader who marked the band puts the stop below the whole area at $98.15, survives the same wick, and is still holding when the setup either works or is properly invalidated.

Now the part that catches people out. The line trader's stop was tighter, so their position had to be larger to risk the same $50: $8,986 against $7,061, about 27% more exposure and 27% more in fees. The tighter drawing felt safer and was worse on both counts at once — the same trap Lesson 10 found in fast timeframes, arriving here through a completely different door.

How wide should a support or resistance zone be?

Judge the width against the instrument's average daily range, never in dollars and not even in percent of price alone. The same width means completely different things on different charts, and the difference is not small.

That gives a working rule, and we will state plainly that it is this site's threshold rather than an industry standard: a usable zone is roughly 10% to 50% of one average daily range. Both edges of that band come from arithmetic rather than taste.

Below about 10%, a routine intraday swing crosses the whole zone several times in a session. "Price is in the zone" then carries almost no information, because price is in the zone most of the time. Worse, your stop sits inside the ordinary noise of the instrument, so you are paying to be flat during moves that mean nothing.

Above about 50%, the stop that must sit beyond the far edge is more than half a daily range away. The trade now needs more than a full day of typical travel simply to reach 1R. That is not automatically a bad trade, but it is a swing trade being run on a level you probably marked with a day trade in mind. The next section puts a number on exactly that.

What does a wider zone actually cost you?

Not a bigger loss. The loss is fixed by your risk rule and does not move. What a wider zone costs is distance: every dollar you add to the band is a dollar the trade has to travel further before it reaches the same reward multiple. Almost nobody prices this, and it is the single most useful thing in this lesson.

Here is the model, stated openly so you can rerun it with your own instrument. Price $98.60. Average daily range 2.5%, so $2.47. Risk $50 on a $5,000 account. The stop goes one quarter of a daily range ($0.62) beyond the far edge of the zone. Round trip costs 0.10%.

Read the last two columns together. Widening the zone from $0.40 to $2.50 leaves your loss untouched at $50 but changes the trade from one that needs 0.82 average daily ranges of travel to make 2R into one that needs 2.53 — roughly three times longer to work, on an instrument that will not move faster to accommodate you. That is the real bill for a generous zone, and it arrives as waiting rather than as a loss, which is why it goes unnoticed.

Now look in the other direction, because the narrow zone is not free either. The $0.10 band has the shortest journey but pays $6.88 in fees against $1.58 — more than four times as much, because a tight stop forces a big position and fees are charged on position, not on risk. And its stop sits inside the instrument's ordinary noise, so it will be hit by movement that means nothing.

Costs rise at one end. Required travel rises at the other. The two middle rows are the only place where both are tolerable at once — and those two rows are 16% and 41% of a daily range. That is where the 10%–50% rule in the previous section comes from. It is not a tradition; it is the shape of this table.

How do you mark a zone without fooling yourself?

Fix the order of operations and follow it every time. Support and resistance is unusually easy to draw dishonestly, because you are choosing the evidence after you already have an opinion.

Three of those steps do more work than the others. Step 1 matters because scrolling through timeframes until a level appears will always succeed — there is a chart out there on which anything is a level. Pick the timeframe from your routine, as Lesson 10 sets out, then look once.

Step 4 is the one almost nobody does, and it converts drawing from taste into arithmetic. Measure the band, divide it by one average daily range, and if the answer is under 10% or over 50%, you have found something other than a tradable level. Do this before you get attached.

Step 6 is the habit that quietly costs the most. Widening a zone while a trade is open is moving the stop while pretending you did not, and it converts a $50 loss into an open-ended one. If the band was wrong, the trade is over; redraw it after you are flat, with the new information, and note in your journal that the first version was wrong. That record is worth more than the trade was.

Why does broken support so often become resistance?

Because a break does not remove the people who were positioned at that area. It only moves them into a worse position, and gives them a new reason to act at the same price.

Work it through with the chart above. Price closes at $97.90, below the $98.40–$98.80 band. Everyone who bought in that band is now underwater. When price rallies back to $98.65 — inside the old zone — those buyers get an unexpected chance to exit near breakeven, and many take it. At the same time, traders who saw the break read the retest as confirmation that sellers still control the area and sell into it. Two different motivations, one price, and the candle closes at $98.05.

What that is not is a rule. Role reversal is a behaviour you observe after the fact, and it fails often. Price may reclaim the band and close back above $98.80, which would invalidate the whole reading. It may cut through the old zone without pausing. It may never come back to test it at all. Treat one rejection as a piece of evidence with a defined expiry, not as a licence to keep shorting the level until it works.

How do you separate a false break from real acceptance?

By closes and by time spent beyond the edge, never by the wick. A wick tells you price visited a level; a close tells you it was still there when the bell rang.

The two events in the chart look almost identical for a moment and are opposites. The sweep went $0.20 below the $98.40 edge intrabar and then closed at $98.95, $0.15 above the $98.80 top of the zone — price went outside, found nothing, and finished the candle further inside the band than it started. The break printed a close at $97.90 and then failed to reclaim the zone on the retest. Same direction, different evidence.

Our own working definition, offered as a starting point you can test rather than a law: require two consecutive closes beyond the far edge of the zone on your chosen timeframe before treating a break as accepted. The trade-off is explicit and worth stating. On a daily chart two closes is two days of movement you have given away; that lost distance is what you pay for not being swept. If you cannot accept that cost, the honest fix is to move to a lower timeframe with its own zone, not to lower the standard of evidence on this one.

One more gap to keep in mind: the price you see at the edge of a zone is not the price you will get. In fast conditions the fill can be several ticks away, which is enough to matter when your whole edge is fifteen cents of stop placement — see slippage.

When does this whole framework stop working?

Support and resistance is a memory system. It fails whenever the thing that happens next has nothing to do with what the chart remembers, and it fails quietly, which makes knowing the failure modes part of using the tool.

When the order book has been replaced. A liquidation cascade, a listing, a delisting or a protocol failure clears out the resting interest that made the level a level. Afterwards the drawing is a record of an order book that no longer exists — see how cascades work. Zones drawn before a violent flush should be re-derived, not reused.

When the market is too thin for the reaction to mean anything. Two "tests" produced by a handful of orders on an illiquid pair is not evidence of demand; it is evidence that nobody was there. Check depth before you trust the shape, as in Lesson 8.

When the trend is strong enough that pullbacks never reach the zone. Waiting for your level here means never entering while the move runs without you, and the usual response — moving the zone up to meet price — is the goalpost problem in a new hat.

When you have more than about five zones on the screen. With enough bands drawn, something always "held", and the chart stops being able to tell you that you were wrong. A chart that cannot disagree with you is decoration.

Common mistakes

Marking every minor turn. Ten zones make any outcome look predicted afterwards, which is the opposite of what the tool is for. Calling one reaction a level. One turn is a data point; two separated reactions are the minimum before the word "zone" is earned. Measuring the zone in dollars. A $0.40 band is 160% of a daily range on a $10 coin and 0.03% of one on Bitcoin — always divide by the range. Putting the stop just under a hairline. That is exactly where the stops cluster, which is exactly where sweeps go. Widening the band mid-trade. A moved stop with better manners. Reading a wick as a break. Closes decide; wicks visit. Treating role reversal as automatic. It needs an observed rejection on the retest, and a close back inside the band cancels it. Trading a touch on its own. A level tells you where to look, never whether to buy — size the trade with the position size calculator and a stop you decided in advance, as in Lesson 6. Reusing zones drawn before a violent move. The participants changed; the drawing did not.

FAQ

How many touches confirm a support or resistance level? There is no universal minimum, but two separated reactions are the point at which the word "zone" is earned. Separated matters more than the count: two adjacent candles from the same bounce are one reaction seen twice, not two. More reactions make the area more visible to everyone, which cuts both ways — a level everybody can see is also a level where everybody's stops are parked. Treat two clean reactions on your chosen timeframe as the floor, and treat a fifth or sixth test as a sign the area is being worn down rather than reinforced.

Should I draw support and resistance from wicks or candle bodies? Use the bodies to set the core of the zone and repeated wicks to set the outer edge. A body shows where trade was actually accepted for a period of time; a wick shows where price went and was rejected. In the worked example here, closes at $99.10 and $99.25 above lows of $98.40 and $98.75 give the band $98.40–$98.80, while the one-off wick to $98.20 is annotated rather than absorbed. If a later reaction confirms that outlier, widen the band then, while you are flat — never while a trade is open.

Does broken support always become resistance? No. Role reversal is a behaviour you observe after a break, not a rule you assume before one. It happens because the break leaves earlier buyers underwater, so a return to the old area gives them a chance to exit near breakeven while break traders sell into the same price. But price may also reclaim the zone with a close back above it, cut straight through without pausing, or never retest at all. One rejection on the retest is evidence with an expiry date, not a licence to keep selling the level until it eventually works.

Can support and resistance predict where price is going? No, and treating them that way is the most expensive mistake in this topic. They identify locations where behaviour previously changed, and therefore where behaviour is worth watching again. They cannot tell you direction, cannot guarantee a reaction, and do not replace a stop or a position size. The useful test is whether your zone can tell you that you were wrong: if you have drawn so many bands that something always holds, the chart can no longer disagree with you and it has stopped being analysis.