Bar Replay Tools Compared: TradingView, KTrainer, ColdRead
All three replay bar by bar. The differences: hidden ticker and date, recorded decisions, and a rule-based post-round review. Compared from public vendor pages.
All three tools can play a chart forward one bar at a time, so "hiding the ending" is not what separates them. Three things do: whether the ticker and date are hidden, whether your decisions are recorded before the reveal, and whether there is a fixed-rule review after the round. TradingView's Bar Replay has none of the three; it is a playback feature inside a charting app. KTrainer hides the ticker and adds AI review and error statistics. ColdRead hides both ticker and date, records every trade as you make it, and grades the round with 13 behaviour tags on ATR-based rules without looking at profit. Reviewing a chart whose ending you know measures storytelling, not judgment1, so the first question about any tool is what it hides.
How was this compared?
Three caveats up front. First, the source is each vendor's public pages and help docs as of 7 September 2026; we did not open paid accounts to test every item, and anything a page does not state is marked "not stated" rather than guessed. Second, the comparison looks at four things only: hidden future, hidden source, recorded decisions, post-round review. Those four are the skeleton of a blind review and decide whether hindsight gets in. Third, ColdRead is our own product; its limits are listed as plainly as the others' strengths.
What is each tool?
TradingView Bar Replay is a playback feature inside the charting app: pick a start date on the chart, the bars after it are hidden, and you step or auto-play forward with indicators and drawings intact, optionally synced across charts. The pricing page lists Bar Replay under Essential and higher paid plans; what the free plan can do is whatever the site says at the time. Its premise is that you already chart on TradingView.
KTrainer is a Chinese-language web platform that describes itself as a "double-blind candlestick training system": real historical candles stepped forward one at a time, ticker and future hidden, four timeframes (30-minute, daily, weekly, monthly), plus AI-assisted review, error attribution with history statistics, leaderboards and themed drills. The page notes that its simulated points cannot be cashed out and do not represent returns.
ColdRead is a web trainer: a random slice of history with both ticker and date hidden, stepped forward bar by bar, revealed only at the end. Every entry and exit is recorded as it happens; after the reveal a rule engine assigns 13 behaviour tags using ATR thresholds, judging behaviour rather than profit. The free tier is 15 rounds a day with no sign-in; local features are a one-time purchase. Market data never passes through the server: a curated set of Chinese A-share history is built in, and anything else is imported by the user.
Side by side
| Dimension | TradingView Bar Replay | KTrainer | ColdRead |
|---|---|---|---|
| Hidden future (bar-by-bar) | Yes | Yes | Yes |
| Hidden source | No; ticker and time axis stay visible | Ticker hidden; whether the date is hidden is not stated | Ticker and date both hidden |
| How a segment is chosen | You pick the symbol and start date | Not stated | Random |
| Recorded decisions | None; keep your own notes | Not stated | Every entry and exit logged, reviewed against the record after the reveal |
| Post-round review | None | AI review, error attribution, history statistics | 13 tags on ATR rules, profit ignored; AI commentary explains, never judges |
| Leaderboard / points | No | Yes; points cannot be cashed out | No |
| Timeframes | Whatever history your plan includes | 30m / D / W / M | Daily on free; all timeframes after purchase |
| Instruments and data | Platform data, widest coverage | Not stated | A-shares built in; everything else imported locally, never via server |
| Price | Listed under Essential and above | Not stated | Free 15 rounds a day; one-time $29 |
| Platforms | Web, desktop, mobile | Web | Web; iOS and Android will share the account when they ship |
When is TradingView enough?
If you already chart on TradingView and want to rehearse entries on instruments you know, with your own indicators, Bar Replay is the least friction: no new app, no data import, drawings and studies stay put. Its strengths are coverage, timeframes and tooling, and neither of the other two comes close on those.
Its problem is the first column of the table: you chose the symbol and the start date, so you know the year and the name. From the moment you recognise the setting, every call carries memory. Hindsight is not something willpower suppresses: people explicitly asked to pretend they do not know the outcome still cannot reconstruct their earlier judgment.1
The fixes exist, but all sit outside the software: have someone else pick the start, cover the symbol and the date axis, write your calls in a notebook away from the chart. All of that is in the seven steps of a blind review. Do them and Bar Replay trains; skip them and it is a video player.
Where KTrainer is strong
Error attribution with history statistics is its most substantial part: mistakes accumulated round after round and grouped by type is exactly what "judge the sample, not the round" needs. Leaderboards and themed drills are a steady motivator for people who like competitive feedback, and a 30-minute timeframe is one more dimension than daily-only free tools offer.
Two things to watch by our criteria. First, it hides the ticker; the page does not say whether the date is hidden, and if the date is visible, recognising the year still leaks the ending, so the source is only half covered. Second, points and rankings pull attention back to how much you won, while training needs to watch behaviour: the same trade can be a chase when it profits and a clean stop when it loses. Both are judgments from the page, not from testing the paid product.
Where ColdRead differs, and its limits
The difference is the three rows of the table: source and date hidden together, decisions logged trade by trade, a fixed-rule review that ignores profit. It takes all three to keep hindsight out. Hide only the future and a recognised setting still gives the ending away; skip the log and after the reveal you remember the result, not the reasoning; look at profit and one round's luck drowns out the behaviour.
The limits, plainly. Web only for now; iOS and Android have not shipped. The free tier is daily timeframe, three basic indicators, 15 rounds a day; all timeframes and indicators, the classic-crash sessions and cross-device sync require the one-time purchase. The built-in dataset is A-shares; crypto, US stocks and FX have to be exported from your own provider and imported. There is no leaderboard and none is planned: it conflicts with the rule of ignoring profit.
How to train this
Try it first without spending anything: any charting tool with history, a random slice, the right side and the symbol bar covered, the seven steps, the record kept. After a dozen segments you will know whether what you need is more indicators or stricter blinding and review; the former points to TradingView, the latter to a dedicated trainer.
If you would rather skip the manual masking and bookkeeping, ColdRead makes that flow the default: 15 free rounds a day, no sign-up, starting from the free practice page.
FAQ
Does TradingView's free plan include Bar Replay?
The pricing page lists Bar Replay under Essential and higher paid plans. Whatever the free plan allows is defined by the site at the time; we did not test it.
What separates blind training from ordinary replay?
Ordinary replay hides the future but not the source: you know the ticker and the year, and the ending is one search away, so memory contaminates the call. Blind training hides the source too, leaving only the chart to judge from.
What does "not stated" mean in the table?
We record only what each vendor's public pages say and do not infer the rest. If something is wrong or out of date, email us; we will verify and update.
References
- Fischhoff, B. (1975). Hindsight ≠ foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299.