Repainting occurs when the calculation uses data that did not yet exist at the moment the signal is displayed. The most common case is when an indicator marks an extremum, but you can only be sure that it is really an extremum after a few candles. Until there is confirmation, the mark is placed and removed as the price moves.
The danger is that on historical data, the repainting is invisible: the chart already shows the final, settled picture. The arrows are exactly where they need to be, and the tool gives the impression of being almost flawless. In real time, the same indicator places a mark, then removes it, then places it elsewhere — and it is impossible to trade using it.
How to check a tool for this property. First: see if the signals change on an unclosed candle, and whether they remain in place after closing. Second: run it through history in step-by-step playback mode, when future candles are hidden. Third: just observe for a few days in real time and compare with what is shown in the history.
It is worth separately distinguishing between repainting and ordinary lagging. A lagging indicator shows the signal late, but honestly and only once. A repainting one shows the signal on time, but then may cancel it — and these are fundamentally different things when evaluating any method based on past results.
