You have been told that a level tested four times is stronger than one tested once. Every level in five years of hourly candles says the opposite — and the gap is not small.
Support and resistance are prices where a move previously stopped, and they work because people remember them and leave orders there — orders that get used up a little more with every test.
Support is a price where falling stopped before. Resistance is a price where rising stopped before. That is the whole definition, and it is worth keeping that plain because almost everything else written about the subject adds mystique that the data does not support.
A level matters for one reason: people remember it. Traders who bought there and did well will buy there again. Traders who sold there and got trapped want out at break-even. Traders who missed the move are waiting for another chance. Three groups, one price, all leaving orders in the same place.
Imagine a wooden fence panel that a crowd keeps leaning against. The first time they lean on it, the panel is solid and it pushes back. The second time, some nails have loosened. By the fourth or fifth time, most of the nails are gone and the panel gives way.
Every test uses up some of the orders that made the level work. The people who wanted to buy there have now bought. That is why the sensible mental model is not a wall getting stronger, but a supply of orders being spent — which is exactly what the numbers below show.
You will read almost everywhere that a level tested three or four times is stronger than one tested once. Every confirmed swing high and swing low in five years of hourly candles was tracked forward, each test scored as held or broken.
| Instrument | 1st test holds | 2nd test | 3rd test | 4th test | Levels tracked |
|---|---|---|---|---|---|
| XAUUSD | 69% | 46.9% | 47.9% | 50.7% | 4,090 |
| GBPUSD | 61.3% | 46.6% | 43.1% | 43% | 4,056 |
| EURUSD | 60.3% | 45.4% | 45.2% | 44.8% | 4,061 |
| US30 | 67.3% | 45.6% | 47.5% | 60.7% | 2,839 |
| NAS100 | 67.2% | 45.2% | 41.9% | 44.9% | 2,898 |
What was counted: Every confirmed swing high and swing low is treated as a level and tracked forward for up to 400 hourly candles. A test is price re-entering a band of 0.1% around the level after having left it. The test held if price moved back out of the band, and broke if an hourly candle closed through the level by more than 0.1%.
Only touch numbers with at least thirty tests are shown; beyond the fourth test the samples get too small to report honestly.
Read the first two columns. On gold the first test of a level held 69% of the time across 3,239 tests. The second test held 46.9%. That is not a small decline — the level goes from genuinely useful to a coin flip in one step, and it stays a coin flip from then on.
The same shape appears on every instrument tested, on thousands of levels each. This is not a quirk of gold or a quirk of the definition. A level is at its strongest the first time it is retested, and the popular claim has it backwards.
Because of how charts are read after the fact. When you look back at a level that was tested five times and then held, it looks like proof that repeated testing builds strength. What you cannot see on that chart are all the levels that were tested twice and broke — they stopped being levels, so nobody drew them, and nobody remembers them.
This is survivorship bias, and it is the single most common way technical analysis fools people. The only cure is counting every case including the ones that disappeared, which is what the table above does.
Nadia has been watching gold hold 4,410 all week. It bounced there Monday, again Wednesday, and again this morning. She reads that as a strong floor and buys, with a stop just below.
What the data says about that decision. She is buying the third test. On gold, the third test held about 47.9% of the time — she is taking a coin flip and calling it a high-probability setup.
What she thought she saw. Three bounces looked like accumulating evidence. In reality each bounce consumed some of the buying that made the level work. By the third test, most of the people who wanted gold at 4,410 already own it.
The trade she should have taken. The first retest, on Monday, was the one with the odds behind it — and it was also the one that felt least safe, because the level had no track record yet. That discomfort is the price of the better odds.
What to do now instead. Either take the coin flip knowingly with a small position, or wait. If 4,410 breaks, the level becomes a candidate to short on a retest from below — and that setup has its own measured hit rate rather than a feeling.
This is the easiest thing in the whole series to verify yourself, and the most worth doing because it contradicts what you have been told. Replay a few months bar by bar, mark levels as they form, and score each test held or broken. Your numbers will look like the table above, and you will never read a "strong level" the same way again.
Test levels in the free backtester →Run the drill in the free backtester → Free, no sign-in to begin.
Support and resistance is the base that everything else in this series sits on. Supply and demand zones are the same idea drawn as an area rather than a line, and they measure better. Order blocks narrow it further to the candle before a break. Flip zones are what happens after a level fails.
The thread running through all of them is that a level is a place where orders were, and orders get used up. Once you hold that idea, the decline in the table above stops being surprising and starts being obvious.
What is support and resistance in simple terms?
Support is a price where falling has stopped before. Resistance is a price where rising has stopped before. They work because traders remember those prices and leave orders around them — not because of anything inherent in the number.
Do levels get stronger the more times they are tested?
No, and this is the clearest result in the series. On gold the first test of a level held 69% of the time across 3,239 tests; the second held 46.9%; the third and fourth stay near a coin flip. Every instrument tested shows the same drop. The belief survives because levels that break stop being drawn, so only the survivors are remembered.
Which test is the best one to trade?
The first retest, by a clear margin — and it is also the one that feels least safe, because the level has no track record yet. That discomfort is what you are being paid for. By the third test you are taking a coin flip that feels like a certainty.
What timeframe should I draw levels on?
Daily and four-hourly for the levels themselves, then drop to your execution timeframe without redrawing them. A level works because many people are watching it, and far more people watch the daily chart than the five-minute one.
Should a level be a line or a zone?
A zone, in practice. Marking a single tick creates false precision and pushes you into stops that are too tight. Supply and demand zones formalise this, and they measure better than lines do.
What happens when a level breaks?
Sometimes it flips and becomes a level in the opposite direction. That is worth knowing about but not worth assuming — it has its own measured hit rate, which is closer to a coin flip than most teaching suggests. See the flip zone lesson.
Replay a few months, mark levels as they form, and score every test held or broken. An afternoon of that will do more for your trading than any article about strong levels, including this one.
Open the free backtester →Beginner exploration
Open each answer for a plain-language way to read Support and Resistance, And Why Levels Get Weaker, test it carefully and decide what to explore next.
This page focuses on “Support and Resistance, And Why Levels Get Weaker”.Support and resistance explained, and the oldest claim in technical analysis tested: across 4,090 gold levels the first test held 69% and the second only 46.9%. Levels get weaker with testing, not stronger.For “Support and Resistance, And Why Levels Get Weaker”, a beginner should identify what the learning guide measures, assumes or teaches before acting on its conclusion.Treat this page's account of “Support and Resistance, And Why Levels Get Weaker” as a learning reference rather than a prediction, signal or promise of future performance.
For “Support and Resistance, And Why Levels Get Weaker”, translate the idea into a definition you could apply the same way on two different charts.While exploring “Support and Resistance, And Why Levels Get Weaker”, work through one example slowly and record which inputs or observations determined the result.Keep your “Support and Resistance, And Why Levels Get Weaker” record honest: list the limitation or counterexample before using the concept in a trading plan.Before leaving “Support and Resistance, And Why Levels Get Weaker”, practise the definition on unseen history and review consistency before judging performance.
Turn one idea from “Support and Resistance, And Why Levels Get Weaker” into a rule with explicit inputs, dates, costs and pass-or-fail conditions.Ask AI to expose missing assumptions in that “Support and Resistance, And Why Levels Get Weaker” test, not to guess the next market move.Use the FXAbsolute AI Backtesting Lab to inspect calculations connected to “Support and Resistance, And Why Levels Get Weaker” and the assumptions behind them.Reproduce any important “Support and Resistance, And Why Levels Get Weaker” result and reserve unseen data before deciding that an apparent pattern is useful.
Continue your exploration of Support and Resistance, And Why Levels Get Weaker with the beginner AI prompt guide, or inspect public calculations in the AI Backtesting Lab.