A chart with every swing high and low marked as a "level" tells you nothing. Price passes through dozens of points on its way through a week. What actually matters is which of those points it came back to more than once, and what rule decided a swing counted in the first place. Below is a way to score level strength honestly, using only price and a lookback window, that works on any platform without buying anything. Define the swing before you draw the line Call a bar a swing high when its high sits at or above every high within a fixed number of bars on either side of it; the swing low is defined the mirror way, on the low. The number of bars on each side is a setting you pick once and apply everywhere, not a judgment call made by eye on each chart. That single choice matters more than it looks: two people staring at the same chart mark different highs as significant, but a fixed bar count either side removes the disagreement. A bar either passes the test or it does not. Merge before you count Raw swing points overstate how many separate levels exist. Price can chop around the same price for days, throwing off several swings a few points apart that are really one level counted five times, not five levels. Before counting touches, merge swings that sit within some distance of each other into a single level, and give that level a price equal to the mean of every swing folded into it, not the price of whichever swing was found first. Averaging in this way matters: one stray outlier bar then cannot drag the whole level to its own price. The merge distance itself should scale with volatility instead of staying fixed in points, because a gap that clearly separates two levels on a currency pair sits inside normal noise on gold. Tying the distance to a share of the instrument's own ATR does that automatically, without a separate setting for every symbol you trade. Skipping this step is the reason a level map ends up cluttered instead of useful. A tool that marks every swing it can find produces a chart with dozens of lines that answers nothing, because most of those lines are the same level counted several times under different swing points. The threshold is a filter, not a decoration Once levels are merged, count the touches behind each one and drop anything below a minimum you set. A level touched once is a swing, not a level: there is no evidence behind it beyond the fact that price was there one time. A level touched four or five times over several weeks has been tested by the market repeatedly and held or reversed on each visit. That repetition is the actual claim a support or resistance line is making. Draw every level regardless of touch count and you are asking the reader to guess which lines carry evidence and which are decoration; set an explicit minimum and drop the rest, and only the tested levels remain on the chart. Print the count, don't imply it A level's strength is only useful if it can be checked. Write the touch count and the age in days directly next to the line, for example "4 touches, 31 days", instead of leaning on line thickness or color to hint at strength. A printed number can be verified: scroll back on the same chart, find the swings near that price within the merge distance, and count them by hand. Disagree with the printed count, and you know exactly where to look for the mistake, in the swing definition, the merge distance, or the count itself. A thicker line cannot be audited the same way. Age answers a separate question from touch count. A level built from touches two years apart is not the same claim as one built in the last month from the same number of touches. Printing both numbers next to the line lets you weigh recency and repetition separately instead of folding them into one visual impression that hides which one is doing the work. Rebuild only on closed bars A swing needs the bars on its right side to already exist before it can pass the test, so nothing about a level can be known until those bars have closed. Rebuilding the map once per finished bar, rather than on every tick, keeps a level's price and touch count fixed once it appears. That stability is what makes the hand check above possible at all: a level that could still change shape while the current bar is forming has nothing settled behind it to verify against. The same scoring rule, bar-confirmed swings, an ATR-scaled merge distance, a minimum touch count, and a printed age, is what an indicator called Auto Level Map draws for you. None of it depends on that indicator specifically: a chart, a spreadsheet, and the rules above are enough to score a level by hand.
