Does my TradingView indicator repaint?

An indicator repaints when a mark it already drew on a closed bar changes afterwards. Nothing in the description on the script page tells you whether yours does. Two checks will, and you can run both this week. Here is the definition, the three mechanisms behind it, and how to test. Run them on anything you are weighing up, ours included.

What does repainting actually mean?

An indicator repaints when a mark it has already drawn on a closed bar changes afterwards. The arrow that sat under Tuesday's candle has moved to Wednesday's when you come back, or it has quietly gone, or it has flipped sides. The bar is finished, its open, high, low and close are final, and the drawing still moved. That matters because it breaks the link between what you saw at the time and what the chart shows you later: the signals in history are not the signals you would have seen live.

It is not the same as a signal that simply arrives late. A mark that lands on a bar and stays where it was put, forever, is not repainting, even if you wish it had shown up two bars earlier. And a tool that never repaints can still be a poor tool. All the property buys you is that the record is real.

Is repainting the same thing as lag?

No, and two different complaints keep arriving under the same name. Lag means the signal is honest but late: it stayed put, and by the time it fired the move was half over. Repainting means the signal was never there when you needed it, and the chart now claims it was. One wastes part of a move. The other rewrites your memory of it.

"Fine, but every non-repainting indicator I've tried was useless and late." They are independent properties: a tool can be guilty of one and innocent of the other, or guilty of both. One that reads slightly ahead of itself on history looks fast and looks right, because both flaws point the same way. Worth knowing which one you are actually annoyed about, though neither label tells you whether the signal is worth taking.

Why do indicators repaint in the first place?

Three mechanisms cause almost all of it, and none of them requires bad intent.

• Higher-timeframe data pulled without an offset. A script asks for the 4-hour value while you sit on a 5-minute chart. On historical bars it gets the completed 4-hour figure and stamps it on bars that closed hours before that figure existed. On history it looks like it knew. Live, there is nothing to stamp yet.
• Calculations that revise bars already closed. Anything confirmed by what comes after it: pivots that need n later bars, zigzag legs that redraw when a new extreme prints, some centred smoothing. Often it is just the maths being what it is. The mark still lands on a bar that could not have carried it live.
• Signals that depend on the still-forming bar. A condition that is true mid-candle and false at the close. If the tool draws on that intrabar state and never re-checks at the close, the arrow you acted on can vanish while you hold the position.

So "does it repaint" is a question about mechanism. Ask how it is built. The limit of that approach is obvious the moment you try it: the source of a protected script is closed, many vendors will not answer the question in any detail, and a description that says "non-repainting" is a claim, not a check. That is why the two tests below exist.

Test one: screenshot it, then come back

This is the test that works, and the slow one.

1. Pick the symbol and timeframe you actually trade. Not the one in the vendor's screenshot.
2. When a signal prints, screenshot the chart. Write down the ticker, the timeframe and the exact bar time.
3. Come back hours or days later. Load the same symbol and timeframe, find the same bar.
4. Compare. If the mark moved, disappeared, or changed side, it repainted. If it sits where you photographed it, it did not, on that bar.

Do it three or four times, in different conditions: a clean trend, a flat chop, a session open. One clean instance is an anecdote.

Vendor screenshots prove nothing. Your own, held against the same bar days later, is the only evidence in this argument that is actually yours. Its weakness is obvious: it takes real time, it tests one bar at a time, and it tells you nothing about the bars you did not capture.

Test two: bar replay, and what it cannot show

TradingView's Bar Replay steps the chart forward one bar at a time. Add the indicator, start replay somewhere in the past, and advance while watching the marks behind the cursor. If something that already printed changes as you move forward, you have your answer in about two minutes.

Two caveats, and the first is not ours. Unless your TradingView plan includes intrabar replay, the bars arrive fully formed, with no movement inside them, and the intraday history you can step through is limited by the plan you are on. That is their limit, not ours, and it blunts this test on exactly the mechanism that produces mid-candle signals.

The second caveat is the important one. A clean replay is not proof. An indicator that reads ahead can look perfectly stable in it, because completed history is exactly where lookahead hides: the values are consistent every time you step through them, and they are still values that could not have existed live. Some honest indicators also behave oddly in replay for unrelated reasons, including how higher-timeframe data and session boundaries resolve inside it.

Treat it as a fast screen, not a verdict. A messy replay is a reason to look harder. A clean one is a reason to run test one.

What can neither test tell you?

Neither test says anything about fills, spread, slippage, commissions, or your data vendor revising bars after the fact. Neither one generalises: an indicator can behave on the daily and misbehave on the 1-minute, or hold on one symbol and not another.

Non-repainting is a property of the display, not of your results. It does not make the signals good. It does not make them timely. It does not tell you anything about what happens after you act on one.

Does it count if only the live bar changes?

"It keeps changing while the candle is open, so it repaints."

No, and this is the objection we get more than any other. Every real-time tool on every platform updates the bar that is still forming, because the price it computes from is still moving. A moving average is not final until the close. Neither is anything built on one. An indicator that froze mid-candle would just be broken.

The honest question is narrower: does anything change after the bar closes? That is the whole test, and it is the one both methods above are aimed at.

It does mean a mid-bar state is not a signal. Act before the close and you are acting on something that may not survive to the close, and no indicator can protect you from that.

Does a non-repainting indicator make more money?

No. Non-repainting settles one thing only: whether the marks sitting in your history are the marks you would have seen live. An indicator that is consistently wrong and never repaints is still consistently wrong. It is now wrong in a way you can actually measure.

The reason to care is narrower than the sales copy suggests. A repainting tool cannot be evaluated at all, because the evidence rewrites itself: every time you go back to check, the record has been tidied up in the tool's favour. Non-repainting is the precondition for judging a tool, not a property of a good one.

So treat it as a gate, not a verdict. It gets an indicator onto the list of things worth testing, and it tells you nothing at all about what you will find once it is there.

Do TradingView's own built-in indicators repaint?

Most of the standard library does not, for a dull reason: it computes from the bars already on your chart. A moving average, an RSI, a MACD read what has printed and nothing else, so once a bar closes there is nothing left to revise.

The usual sources are the ones that look forward by design.

• Pivot and fractal markers. A pivot needs n bars either side before it is confirmed, so the label appears several bars after the bar it points at, then sits there permanently.
• Anything drawn from a higher timeframe. A script requesting another timeframe's value can stamp a completed figure onto bars that closed before that figure existed, unless it offsets for it.
• Drawings that redraw on a new extreme. Zigzag legs and similar tools move the last leg whenever a further high or low prints.

None of those is a fault; it is what those tools are for. The trap is in the reading: a pivot that took five bars to confirm is easy to read on history as though it printed the moment it appeared. Check when the mark could first have existed, not where it sits now.

Whatever you end up using, test it once

You do not need to buy anything to get value out of this page. Pick the indicator already on your chart, the one you have been quietly trusting, screenshot its next signal with the bar time written down, and go back to it on Friday. Ten minutes of work, spread over a few days, settles a question you would otherwise argue about forever. What it does not settle: it clears one bar on one symbol, it says nothing about the other nine hundred bars on your chart, and it says nothing at all about whether the tool is any good. If you are still choosing, the multi-timeframe indicator guide covers the rest.

How does Trade Symbiotic behave, and how would you check?

Closed bars stay put; only the live bar updates. The honest cost: you wait one bar for that close, in exchange for a signal that's real. Wave and Tide are rebuilt from a rolling window of your own bars, so they are a live approximation, not a copy of the higher-timeframe chart.

Put us through the same checklist: the script is listed as Trade Symbiotic - Premium, the documentation is unpaywalled, and the Sandbox needs no account.

If the mechanism question is settled and the one left is price, how it compares on price and terms sets out what the alternatives charge, including where our 7-day guarantee is the shortest of the group.