Support and resistance: turning a vague idea into a testable rule

Support and resistance is the first concept every trader learns and the last one they manage to automate. The idea is intuitive: support is a price where buying tends to appear and halt a fall; resistance is a price where selling tends to appear and halt a rise. The trouble starts the moment you try to turn that intuition into a rule a computer can trade — because 'a level the market respects' is not, as stated, a definition at all.

What the levels actually are

Support and resistance are not magic prices. They are zones where the recent balance of supply and demand shifted, and they matter largely because they are remembered. Traders who bought at a level want to defend it; traders who missed a move want a second chance at it; traders who are trapped want to exit at break-even. A level becomes meaningful because enough participants act around it — which means support and resistance are partly self-fulfilling, and partly just the visible footprint of where real volume changed hands.

Why hand-drawn lines don't backtest

Open any chart and you can draw a support line that looks perfect. That is exactly the problem. Drawn in hindsight, with the freedom to choose which wicks to touch and which to ignore, a support line always fits — it is the most overfit object in technical analysis. A computer cannot trade "the line I would have drawn." And a strategy built on hand-picked levels has no honest backtest, because the levels were chosen knowing what price did next.

Definitions a computer can use

To trade support and resistance systematically, replace the hand-drawn line with a rule that computes the same idea from data, identically every time:

• Swing highs and lows — a swing high is a bar whose high exceeds the N bars on each side. These pivot points are objective, repeatable support/resistance candidates, and the same N always produces the same levels.
• Prior period extremes — the previous day's, week's, or session's high and low are unambiguous, widely watched levels that need no judgement to identify.
• Round numbers — psychologically significant prices are trivially objective and genuinely attract activity.
• Volume-based levels — VWAP and high-volume price nodes mark where real participation occurred, computed the same way every time.
• Moving averages — a long MA acts as dynamic support/resistance and is fully objective by construction.

Each of these is a precise, computable stand-in for the intuition. Once a level is defined by a rule, you can build a strategy on it — a bounce, a breakout, a retest — and, crucially, backtest it honestly, because the levels were generated by the rule, not chosen by hindsight.

Trading the levels — two opposite plays

An objective level supports two opposite strategies, and which one works is a question of regime. The reversion play fades the level — buy at support, sell at resistance — and it works in ranging markets. The breakout play does the reverse — it trades the level giving way, betting a break of resistance starts a new move — and it works in trending markets. Same levels, opposite trades. As always, a regime filter decides which one the current market rewards.

Noon Barbari's strategy designer lets you build support/resistance rules from objective inputs — swing pivots, prior-period extremes, moving averages — and backtest them honestly, because the levels are computed by rule rather than drawn by eye. Swing-based levels in particular are first-class in the engine.

Support and resistance is a genuinely useful idea. It just has to survive the move from a line on a chart to a rule in a backtest. Define the level objectively, and an intuition becomes a strategy you can actually test.