My first charts looked like sheet music. A horizontal line under every wick, another above every high, a different color for each timeframe, until the candles were the least visible thing on the screen. I thought the density proved I was doing the work. What it actually did was hand me a justification for any trade I already wanted to take, because when everything on the chart is a level, some level always agrees with you.
The fix I eventually landed on is a much stricter definition of what earns a line, plus a habit of deleting. Three sources of levels, one weekly cleanup, about five minutes per asset. Here is the whole thing.
Where real levels come from
The first source is swing points, and only the obvious ones. Zoom out to the daily or weekly, lean back, and mark the highs and lows a stranger would circle within a couple of seconds of seeing the chart. If you have to hunt for a swing, it does not count. Levels work because a lot of participants share a memory of being right or wrong at a price, and nobody remembers a minor pivot from a random Tuesday on the 15 minute chart. The swings you can see from across the room are the ones everyone else can see too, which is the entire reason they get defended.
The second source is high-volume closes. Scan for sessions where volume ran well above normal and note where price closed rather than where it wicked. A heavy-volume close marks a price where an unusual number of people actually took positions and went home holding them. Those people now care about that price. They add on dips back into it, or they bail at breakeven when a losing position finally comes back to it, and both behaviors show up later as reactions on the chart. A volume profile tool will show you the same thing as fat nodes, but eyeballing the volume bars underneath the candles gets you most of the way there.
The third source is the edges of old consolidations. When price goes sideways in a box for weeks, the top and bottom of the box become reference points for everyone who traded inside it. After price finally leaves, the edge it broke through tends to get retested, and the reaction at that retest is some of the cleanest price action you will find anywhere. I mark the edges of the box and skip the middle. The middle is where price churned, and churn is the opposite of a level.
Zones beat exact prices
Everything I keep gets drawn as a rectangle. A single line is a bet that thousands of strangers will all act at one exact price, which is a strange thing to expect. Some of them anchor to the old wick low, some to the close, some to the round number sitting nearby, so price rarely reverses at one exact number. The reversal happens somewhere inside a band.
The practical way to draw the band is to cover both the wick extreme and the candle body of whatever move defined the level. On a daily crypto chart that typically gives you a zone roughly one or two percent tall, usually tighter on a large-cap stock. The exact width matters less than the habit, because the single-line version fails in two directions at once. Either price pierces your line by a fraction, takes out your stop, and then reverses exactly the way you expected, or price turns slightly in front of your line and you never get filled and watch the move from the sidelines. A zone turns both of those from regular events into rare ones.
Touch counts, and when a level dies
A single touch tells you price turned there once, and once can be an accident. I want to see a second respected test before I treat a level as real, and I want the two tests separated in time, ideally by days or weeks on a daily chart rather than two candles in the same afternoon. Two clean tests and the level moves from candidate to keeper.
Past two, more touches make me more confident, but only up to a point, and here I will admit the evidence is murkier than the textbooks make it sound. The classic view says every additional touch strengthens support. Plenty of practitioners argue the opposite, that each test consumes the resting orders sitting at the level, so the fifth touch is more fragile than the second because most of the willing buyers have already been filled. I lean toward the second camp when the retests come fast and tight, and toward the classic view when they are separated by months, but I hold both loosely. What I will say with confidence is that a level getting hammered repeatedly in a short window usually breaks, and treating touch number six as a high-conviction bounce gets expensive.
Deleting levels matters as much as drawing them. When price closes through a zone decisively, meaning a full candle body beyond it on the timeframe you drew it on, do not remove the level right away. Flip the label instead, because broken support has a long history of acting as resistance on the retest, and the reverse holds too. What does get deleted is anything price has chopped through in both directions more than once, anything drawn on a lower timeframe that price left behind long ago, and anything so old you cannot remember why you drew it. That last rule sounds like a joke but it is the one I use most. If a level does not come with a story you can retell, it is clutter.
The five minute routine
Here is the full pass, per asset, once a week or after any large move.
- Open the weekly chart and mark the two or three swing highs and lows that jump out instantly. Skip anything you had to search for.
- Drop to the daily and add heavy-volume closes and consolidation edges, but only the ones within reach of current price. A level sitting half the chart away is trivia for now, so leave it off.
- Convert every line into a zone that spans wick to body.
- Delete anything with fewer than two touches, unless it is a fresh consolidation edge that has not had time to be tested yet.
- Flip anything that broke cleanly. Delete anything that churned.
- Count what is left. More than six or seven zones and I cut the weakest until I am back under, because the discipline of cutting is what keeps the survivors meaningful.
The finished chart should look almost empty, a handful of rectangles and a lot of open space, and that used to bother me because it felt lazy. It stopped bothering me when I noticed that trades taken at marked zones aged noticeably better than the ones I improvised in the space between them. Most sessions price never even enters one of your zones, and on those days the nearly empty chart is telling you the most useful thing a chart can tell you, which is that there is nothing to do yet.