Bar replay is a better chart. A question bank is a better test.

If you practise reading charts, you have probably used bar replay. TradingView's version is the one most people mean: you pick a symbol, scrub back to a date, and the chart plays forward one bar at a time while you decide what you would have done.

It is a good tool, and this is not an argument that it is not. It is an argument that it is a different kind of thing from what we built, and that the difference is not a feature list.

What bar replay is genuinely good at#

You choose the instrument. That matters more than it sounds: rehearsing on the thing you actually trade, in the timeframe you actually trade it, is not a luxury.

You can move between timeframes, put your own indicators on, mark levels, and rehearse the whole decision — the entry, where the stop goes, where you would take it off. You can go back to a specific event you care about and study it. And it lives inside a charting platform you already have open, which is the single most underrated feature any tool can have.

None of that is something we do. We do not attempt most of it.

The one difference that matters#

Bar replay is self-administered, and that is a description rather than a complaint. You choose the chart, you choose the moment, and the tool does not require you to commit a call before the next bars appear. You can impose that discipline on yourself — plenty of people do — but it is not part of the exercise.

That freedom is exactly what makes it a good analysis tool. A question bank gives the freedom up on purpose, and gets a controlled test in exchange.

So the honest claim is not that a replay is incapable of any of this. It is that a question bank enforces it: you are dealt a chart you did not pick, at a moment you did not choose, and the call is recorded before the outcome exists. That is worse for studying a setup you care about, and better for finding out whether you can read one at all — which is what people mean when they look for a TradingView Bar Replay alternative for blind chart-reading practice.

Four controls a question bank enforces#

You do not choose the chart#

When you scrub to a date, you usually know the symbol, you often half-remember the period, and you picked it for a reason. None of that is cheating; it is just not a blind test.

⚠️ TradingView is ahead of most of this critique, and it is worth saying so: their help pages describe a Random bar option that drops you at an arbitrary point in history (as of 2026-08-11). That removes the "you chose the moment" problem, which is the biggest one.

What it does not remove is the rest of the sentence. You still know the instrument. Nothing balances what you are dealt, and nothing filters it — which are the next two.

Direction is balanced, exactly#

The competitive bank is 20,300 questions balanced 10,136 up against 10,164 down — within a tenth of a percentage point of even. That is not sampling; the selection is balanced by construction and publication aborts if the mix drifts.

It has to be enforced because US equities drift upwards. Any practice built from moments you selected yourself inherits that drift, and it quietly rewards a habit — leaning long — that will look like skill for as long as the sample is small.

Most historical bars make poor questions#

This is the one that surprised us when we measured it, and it is worth being precise about because it is easy to overstate.

We swept 140,319 decision points across 22 US equities and ETFs with no filter at all. Over the following fifteen minutes, 20% moved less than a quarter of the symbol's own ATR — barely enough to call anything — and only 31% landed in the 1–3× ATR band the question bank selects for: large enough to produce a meaningful outcome, while excluding the most extreme moves.

That is not "most bars are noise". It is that historical bars are not interchangeable as questions, and a replay hands you all of them with equal weight. Our generator keeps the band, and takes at most one question per symbol per trading day so two questions never share their outcome bars.

The call is locked before the outcome#

This is the deepest one, and it is not a criticism of any product — it is a property of practising alone.

Watching a bar print and thinking I knew that is not a record. "I knew that" is not evidence. The story assembles itself around the outcome, and afterwards you genuinely cannot tell whether you called it. Do it long enough and the skill being trained is post-hoc explanation. We wrote about the research behind that in what the research says about trader intuition.

And this is where scoring stops being a game mechanic and starts being the point — because without a committed record, you cannot tell a good session from a lucky one. A player answering at random scores 60% or better in one session out of four, and over a hundred answers hits a five-in-a-row streak 81% of the time. Establishing that a genuinely 60% player is a 60% player takes about 194 answers. None of that is knowable from a replay session, however many bars you stepped through. We took that apart in why SwipeTA is not Duolingo for trading.

The controls are worthless if you never do it twice#

Everything above is about making one answer meaningful. It leaves out the harder half.

Narrowing the task is part of what makes repetition possible at all. We are not rehearsing a trade; we are asking one short-horizon intraday question, in the same shape, over and over.

Our own arithmetic is blunt about it: establishing that a genuinely 60% player is a 60% player takes about 194 answers. That is not a session you sit down for. It has to fit into ordinary gaps in a day rather than demand a dedicated study session every time.

And this is where a replay tool and a question bank part company for a reason that has nothing to do with rigour. To use a replay you have to enter a state — open the platform, pick a symbol, scrub the timeline to a date, decide for yourself when the trade was over. That is not a flaw; it is exactly the deliberate, unhurried mode that makes it good for studying something. But it is a threshold, and thresholds are paid every single time.

We spent the design the other way. Classic's questions are bundled into the app rather than fetched, so a card deals instantly with no round trip — our own publisher's source comment calls that "the whole feel of Classic", which is an odd thing to write about a data format unless the feel was the point. One hand, a real US intraday chart, a swipe, an answer. In the beginner and survival modes the timer is eight seconds; the competitive modes give you fifteen.

That is the trade in one line: a replay is better for the session you plan, a question bank is better for the two hundred you did not.

Low friction gets one session started. It does not get the next one started, and that is why the rest of the product looks more like a game than a charting tool: daily goals, streaks, a level progression and competitive modes with a rank attached. None of that makes an individual answer more valid — the controls above do that. Their job is narrower and less glamorous: to make the next hundred answers more likely to happen.

It costs us something real, and the section that admits it is below. Eight seconds is not long enough to think properly — that is the point of the number, not an oversight. What is being trained is fast directional recognition, not analysis. A swipe is also not a place to write down why. If you want to sit with a chart, we are the wrong tool, and no amount of instant dealing fixes that.

The implementation question underneath all of this#

Every tool in this category, ours included, has to answer one more thing: are the indicators on the screen computed only from data that existed at the decision point? A line calculated over the whole series and then hidden on the right is a different artefact from one that was never drawn past the cut-off, even though they look identical.

We are not claiming TradingView gets this wrong — we have not audited anyone's implementation, and this piece is not the place to. What we can document is how our own generator handles it, which is written out in four places look-ahead bias leaks into a chart replay.

Where bar replay is better, plainly#

Several places, and some of them are not small.

It practises the whole trade. Entry, size, stop, target, when to get out. Our competitive modes have a stake and a multiplier, but there is no exit at all — once a call is committed it settles at its fixed horizon whatever happens in between. That is a deliberate narrowing, and it means the thing paper trading and replay teach about managing an open position, we do not teach.

It is multi-timeframe. We are deliberately not. SwipeTA narrows the exercise to short-horizon intraday reading: one intraday chart, one decision point, one fixed outcome horizon. That narrowing is what makes the questions comparable to each other and the loop short enough to run hundreds of times — but it removes context a real trader may care about. If the daily chart is in a downtrend, our question does not tell you. Bar replay can.

It is your instrument. Ours are 50 US large caps and ETFs in the competitive modes and 22 in Classic, dealt to you. If you trade something else, a replay on your own symbol is closer to the thing you are actually doing.

It is a tool, not a game. You can stop, draw on it, take notes, go back. We are a timed swipe.

So which one#

One row says SwipeTA rather than "a question bank", and the distinction is real: the controls above are a property of any question bank, but whether you come back tomorrow is a property of a product.

If you want to rehearse a strategy — entries, stops, position management, on the instruments you trade — use bar replay. We do not do that, and a comparison that pretended otherwise would be useless to you.

If you want to find out whether you can read a chart at all, on a sample that is balanced, filtered and scored, and where the answer is not knowable in advance, that is what a question bank is for.

Most people who take chart reading seriously should use both, and the order matters less than the honesty: one of them tells you what a setup did, the other tells you what you do.

What neither of them settles#

Both are practice on historical data, and neither has shown that practice transfers to real markets. We make no such claim about ours, no study we know of supports it for any tool in this category, and we would be sceptical of anyone who told you otherwise. SwipeTA is a training game and a simulation — no real money, no broker, and it does not provide investment advice.

Everything we do claim, with the parameters behind it, is on our methodology page.