Study · Solana
Do support and resistance zones work? We tested ours against random levels
Over 44 days of SOL, our graded zones held 46.0% of the time. Random levels of the same size, at the same distance from price, held 49.0%. Here is the method, the numbers and what we changed because of it.
The question
Every trader draws support and resistance. Wrackline is our version: a map of zones for Solana built from swing highs and lows on several timeframes, named levels and round numbers, each zone graded A+, A or B by how many of those things stack up in one place. A zone under price is a Long zone, a zone over price is a Short zone.
A map like that looks convincing on a chart, because your eye finds the bounces and skips the breaks. So the question we asked was narrow and checkable: when price taps one of these zones, does it hold more often than a random level would?
How we tested it
- Point in time. The zone ladder was rebuilt every hour, 1,048 times, using only candles that had already closed before that hour. Each tap was judged against the ladder that existed at that moment. Nothing from the future was allowed in.
- What "held" means. After the tap, price moved 0.1 of the average daily range away from the zone before it traded 0.025 of the average daily range through the far side, inside 90 minutes.
- One tap, counted once. Repeated taps of the same zone on the same side inside 90 minutes count as one.
- The control. For every real zone, every hour, we placed three random zones with the same width and the same distance from price, on a random side, and scored them the same way. If real zones are special, they should beat these.
- The trade. Entry at the zone edge, stop one 15-minute ATR away, exit at the stop or after 90 minutes, fees included.
The result
| Taps | Held | 95% range | Net R per tap | |
|---|---|---|---|---|
| Wrackline zones | 315 | 46.0% | 40.6 to 51.6% | −0.12 |
| Random levels | 412 | 49.0% | 44.2 to 53.8% | −0.11 |
The two ranges overlap almost completely. On this sample the zones are not distinguishable from random levels, and if anything they sit three points lower. Net R uses our measured trading cost of 0.063% per round trip. At a more cautious 0.12% it is −0.27 for the zones and −0.26 for random levels. Before fees both are about +0.05R: the fee decides the sign.
Did the better grades do better?
| Grade | Zones | Taps | Held | 95% range |
|---|---|---|---|---|
| A+ | 27 | 37 | 37.8% | 24.1 to 53.9% |
| A | 44 | 62 | 50.0% | 37.9 to 62.1% |
| B | 85 | 216 | 46.3% | 39.8 to 53.0% |
No. The top grade has the lowest hold rate, although with 27 zones it is under our own minimum of 30 and its range is wide enough to mean very little. What the table does not show is any sign that more stacked levels make a zone stronger.
What this does not prove
- It is 44 days of one market. A different month, or a different coin, could look different.
- The historical ladder leaves out things the live map can see, such as order-book walls and outside points of interest.
- It tests one mechanical rule: touch the zone, enter, wait 90 minutes. It says nothing about a trader who waits for a reaction before entering.
What we changed
The negative result leads the page. Wrackline opens with this finding instead of hiding it.
A zone is a location, not a signal. It tells you where a decision happens. It does not tell you which way to trade it.
No grade without a sample. Under 30 zones a grade is shown as unproven.
The test keeps running. The ladder is archived every 15 minutes so the study can be repeated on new data.
If the zone alone is not the edge, the read around it has to be: trend, volatility, who is pushing. That is exactly what our game lets you practise on the same data.