A level that has already rejected price twice is not stronger on the third attempt — it is thinner. Each failed test consumes the resting orders that caused the rejection and stacks protective stops just beyond the line. Trade the third test by waiting for the candle close, then fading the hold or following the break.
This is one of the places where the chart tells you the opposite of what the order book is doing. Two clean rejections look like confirmation. Draw the line, note that it has held twice, and the natural conclusion is that it is a strong level worth leaning on. Meanwhile the thing that actually produced those two rejections has been spending itself, and the rest of the market has had two chances to notice the line and put a stop behind it.
What a twice-failed level actually tells you
A level holds because there are orders resting at it. Someone is willing to sell every time price reaches 5,000, or buy every time it reaches 4,940, and while that willingness lasts, price turns away. Nothing mystical is happening. There is supply, and it is enough.
So the honest reading of two rejections is: there was enough supply at this price, twice, and some of it has now been used. That is genuinely useful information — it tells you the level is real rather than a line you drew because it looked tidy. What it does not tell you is that the level will hold a third time, and the chart gives no direct view of how much is left.
There is real evidence that published levels carry information. A Federal Reserve Bank of New York study tested the support and resistance levels that six firms active in the foreign exchange market supplied to their customers, and found strong evidence that the levels help predict intraday trend interruptions — while noting that their predictive power varied across the exchange rates and firms examined (Osler, “Support for Resistance: Technical Analysis and Intraday Exchange Rates”, FRBNY Economic Policy Review vol. 6 no. 2, July 2000, pp. 53–68). Read that carefully, because both halves matter: levels work, and they do not work uniformly. A level is a probability, not a wall.
Why the third test is different from the first two
Two things change between the first visit and the third, and they pull in the same direction.
- The absorbing side is smaller. Every rejection fills some of the orders that caused it. Unless that seller keeps replenishing, the queue in front of price is shorter on each visit.
- The stop cluster behind it is bigger. Two rejections advertise the line. Anyone who bought the bounce now has a stop below it; anyone who sold the rejection has a stop above it. By the third visit there is a dense band of resting stop orders just beyond the level.
- Time has passed. If the two rejections happened three days apart, the participant who was defending the level may simply be gone.
That second point is why third-test breaks tend to move fast rather than drift. Stop-loss orders are not neutral: they buy into strength and sell into weakness, which is why the same researcher devoted a separate Fed paper to the way stop-loss orders generate positive-feedback trading and produce price cascades in currency markets (Osler, “Stop-Loss Orders and Price Cascades in Currency Markets”, FRBNY Staff Report no. 150, July 2002). A level that has failed twice is, in plain terms, a level with less in front of it and more fuel behind it. The mechanics of that cascade, and how to tell one from a genuine breakdown, are unpacked in liquidity grabs and stop hunts.
The two ways a third test resolves
There are only two outcomes worth planning for, and the candle close separates them.
| What you see on the third test | What it means | Action |
|---|---|---|
| Price reaches the level and closes back inside the range | Supply is still there and still winning | Fade, stop beyond the wick cluster |
| Price closes through the level, then the next candle holds | The absorbing side is finished; stops are being taken | Follow, on the hold — not on the break |
| Price closes through, then immediately closes back inside | Stops harvested, no follow-through | Stand aside. This is the trap |
| Price stalls into the level on falling volume | No conviction from either side | No trade. Wait for participation |
| “It has held twice, it will hold again” | An assumption, not a read | No action |
How to trade the hold
If the level holds a third time, the trade is the rejection, not the touch. You want a candle that goes to the level and closes away from it, which is the same confirmation logic as break and hold applied in reverse. The entry is on that close. The stop goes beyond the level and beyond the wicks of the first two rejections, because that is where the trade is genuinely wrong — if price trades cleanly past three visits' worth of wicks, the level is not doing what you thought.
Two honest caveats. First, your reward is capped by the range you are fading back into, so the risk-to-reward on a third fade is usually worse than it was on the first. Second, this is the attempt with the highest chance of being the one that breaks. Fading a twice-failed level is a legitimate trade; pretending it is the safest trade on the chart is not.
How to trade the break
The break is the side of a third test most traders get wrong — by entering on the break itself. A break is a price touching a line. A hold is a candle closing beyond it and the next candle refusing to come back. Only the second one is evidence.
The clean version is to wait for the close through, then take either the immediate hold or the pullback that comes back to the broken level and fails to re-enter the old range. That second entry is the classic flip, and it is covered in detail in what a retest is. It gives you a tighter stop and a defined invalidation, at the cost of sometimes missing the move entirely when there is no pullback. That trade-off is the whole argument, and it never fully resolves.
When to skip the level entirely
Not every twice-failed level is worth a trade, and the discipline to pass is most of the edge here.
- The two rejections were not comparable. One on heavy volume at the open and one on a thin lunchtime drift are not two tests of the same thing. They are two different markets touching the same number.
- The level is not clean. If you had to squint, or if the wicks are twenty points apart, you have a zone rather than a level — and the stop that respects the whole zone is usually too wide to size properly. Supply and demand zones are traded differently for exactly this reason.
- You are already at your daily risk limit. Third tests resolve fast in both directions. That is not the trade to take when the account is already down.
- You cannot say what would make you wrong. If there is no price at which you would admit the read failed, there is no trade — only a position. That is what invalidation means in practice.
Frequently Asked Questions
Does a level get stronger every time it holds?
No. That is the most common misreading of a chart. A level holds because there are resting orders at it, and every test that holds uses some of those orders up. The line looks more impressive each time it is drawn, because the chart now shows two clean rejections, while the thing that produced those rejections is quietly getting smaller.
Why does the third test of a level often break?
Two things have changed by the third visit. The resting orders that caused the first two rejections have been partially filled, so there is less left to absorb the next push. And the two rejections have advertised the level to everyone watching, so protective stops now sit in a cluster just beyond it. Less supply in front, more fuel behind.
Should you fade or follow a third test of a level?
Let the candle close decide rather than deciding in advance. If price reaches the level and closes back inside the range on a rejection, that is a fade with a stop just beyond the level. If price closes through the level and then holds on the next candle, that is a follow. Trading before the close is guessing, and a twice-failed level is the worst place to guess.
Where does the stop go on a third-test trade?
Beyond the level, and beyond the stop cluster that the first two failures created, not inside it. If the level is 5,000 and both prior rejections wicked to 5,003, a stop at 5,002 sits in the middle of everyone else's. Place it where the trade is genuinely wrong, then size the position from that distance rather than shrinking the stop to fit the size you wanted.
Bottom line
Two rejections prove a level is real. They do not prove it will hold again, and the order flow argues the other way: less resting supply in front of price, a denser band of stops behind it. So treat the third test as the open question it is and let the market resolve it for you. Wait for the close, take the rejection or take the hold after the break, put the stop where the read is genuinely wrong, and skip the ones where the two prior tests were not really the same test. The level is information, not a promise — and how a single read like this becomes a repeatable rule is the subject of trading strategies explained.