Repainting (indicators)

A repainting indicator shows you one thing on the chart today and something else on the same bar tomorrow, with no code changed in between. The arrow that flagged a perfect bottom on Tuesday's close can slide or vanish by Thursday, because the calculation behind it was never finished when it first appeared. What you saw in real time and what the history shows now are two different records of the same candle.

Most repainting comes from an indicator leaning on information a bar doesn't have until it's over: a pivot high that needs several later bars to confirm itself, a zig-zag line that only knows where a swing turned once price has moved well past it, a signal tied to the closing price that keeps recalculating candle by candle while the candle is still open. None of that is malicious. It's just how the math works. The problem is the chart presenting the final, settled version of history as if it had been visible the whole time.

That makes it the live cousin of look-ahead bias. Look-ahead bias is a backtest quietly using tomorrow's data today; repainting is an indicator doing the same thing on your actual screen, redrawing its own past in real time to hide it. Both hand the tool a record it didn't earn. The difference is you can watch repainting happen, if you leave the chart open long enough and remember what the arrow used to say.

Our audit covers 1,700+ indicators alongside the strategies, and repainting is one of the most common ways an indicator's on-chart history oversells itself. A script can look uncannily accurate in a screenshot and still be useless the moment you'd need to act on a signal before it finishes forming. Every indicator that clears our first gates goes through the same adversarial pass we run on strategies, where a second, stronger model tries specifically to catch calculations that only make sense in hindsight, the same discipline we apply to checking intrabar fill assumptions on the strategy side.

Why it matters is simple. A repainting signal can't be traded on the bar it appears to trade on. By the time it settles, the move it flagged is already obvious on the raw price chart, no indicator required. The backtest and the screenshot lied in the same direction, and they lied for free.