Bar Replay vs. a Real Replay Desk: What Each One Can Actually Teach You
Chart replay tools and execution replay desks solve different problems. Knowing which problem you have saves you months of practicing the wrong thing.
- tools
- practice
- execution
There are three ways to practice day trading without live money, and traders pick between them almost at random.
Chart replay, where you scrub historical candles forward and mark your entries. Broker paper trading, where you trade the live market with fake money. Execution replay, where a historical session plays back through an order book and a fill engine.
They’re not competing products. They teach different skills, and picking the wrong one means grinding for months on something you weren’t trying to improve.
What bar replay is genuinely good at
TradingView’s bar replay does one thing extremely well: it removes hindsight from chart reading.
Scroll back on any chart and every setup is obvious. The breakout that worked is the only one you can see, because your eye skips the eleven that failed. Bar replay fixes this by hiding the right edge. You step forward candle by candle, and you have to decide now, with only what’s on screen.
For pattern recognition that’s close to ideal. Cheap, quick, works on any symbol and timeframe on the platform you already have charts on, and you can burn through fifty setups in an afternoon. If your problem is “I can’t tell a real breakout from a fake one,” bar replay is the correct tool and it’s not close.
Where it stops is the boundary of what a chart is.
Bars aren’t a market
A candle is a summary. Open, high, low, close, volume, for an interval. Everything that made those four numbers has been discarded: the sequence of trades, the quotes standing behind them, the sizes, which venue printed what.
You can’t rebuild an order book from a candle. The information isn’t compressed, it’s gone.
So when a chart replay tool lets you “buy here,” it has exactly one price available: a number you picked off the bar. It has no way to know whether 400 shares or 40,000 were offered at that price, whether the quote was firm, or what the book looked like 180 milliseconds after you decided. It fills you at your number because your number is the only input it has.
That’s not a flaw in the tool. It’s the honest consequence of operating on bars. But it means every execution lesson is unavailable, because there is no execution being modeled.
What that costs you
Here’s the concrete list of things you cannot practice on a chart, however many hours you put in.
You can’t learn sizing against liquidity, because there’s no displayed size to size against. You can’t learn queue discipline, because there’s no queue. You can’t develop a feel for when a marketable order is worth the spread versus when to rest, because both fill identically. You can’t read a montage. You can’t build hotkey muscle memory for an order ticket that doesn’t exist. And you can’t discover that your favorite setup only works at prices you’d never actually get.
That last one is the expensive one. A strategy validated purely on chart replay has been validated at ideal prices. Whether it survives real fills is an open question your practice hasn’t touched.
Broker paper trading: the opposite trade-off
Paper accounts flip the problem. You get the real platform, the real montage, real hotkeys, a live market with genuine liquidity on the screen.
What you lose is history and honesty about fills.
History, because you can only trade what’s happening right now. You cannot practice the open on a specific day, cannot run the same setup twenty times, cannot go back to the day you blew up and take it again. Your practice is rate-limited to the market’s schedule, and small-cap momentum concentrates into a window that’s over by 10:30.
Honesty, because most paper engines fill at the touch. Some add a fixed penny penalty. Either way there’s no latency, no queue, and no walking the book, so the fills are systematically better than yours would have been. Paper-trading fills lie goes through the specific mechanics.
Execution replay
The third mode replays a real historical session through an actual matching engine.
The tape is tick-level: NBBO with bid and ask sizes and exchange routing, trade prints with their condition codes. Quotes get filtered for eligibility before anything trades against them, so manual, stale, crossed, and condition-flagged quotes never become fills. Orders draw a latency before they’re actionable. Marketable orders walk the book when they exceed the inside. Resting orders take a queue position and can sit there unfilled while price moves away. Every fill comes back with an attribution: what the quote was when you submitted, when the engine evaluated, and when you filled.
That’s a different experiment. You’re no longer asking “would I have spotted this,” you’re asking “would I have gotten this, at what price, at my size.”
The cost is that it’s heavier. There’s more to configure, more to read afterward, and a fill log takes real work to interpret compared to a chart with arrows on it.
Since this is the category we build in, here’s what ours actually does, so you can hold it against whatever else you’re looking at. DAS Replay runs US equities across NASDAQ, NYSE, ARCA, AMEX and BATS, about 12,000 symbols, from 04:00 to 20:00 ET on one continuous timeline, so the pre-market setup and the open are the same session rather than two disconnected charts. Speed runs 0.1x to 10x and you can pause, scrub, or seek to a candle index mid-session.
Order entry is a Lightspeed-style ticket priced to $0.0001 with a two-column Level 2 montage beside it. You pick a latency profile before the session: 180ms give or take 40 for a consumer round trip, 120 plus or minus 25 for a direct-access desk, 30 plus or minus 10 colocated, or set your own mean and deviation. Every submission draws fresh from a seeded generator, cancels draw their own, and fill-versus-cancel races resolve in tick order. Same seed, same session, same fills, which means a result you don’t believe is one you can re-run.
Two scanners replay alongside it, Top Leading Gainers and High-of-Day Momentum, and their alerts fire at the timestamps they really fired at. Click one and the desk jumps to that ticker ten seconds before the trigger, so you see the setup rather than the print.
Afterward you get the fill log, the slippage attribution split three ways, and a journal that keeps the sessions.
Picking by the problem
| Your problem | The tool |
|---|---|
| Can’t distinguish real setups from fakeouts | Bar replay |
| Don’t know the platform / fumble hotkeys | Broker paper account, or a replay desk |
| Strategy looks great on charts, loses live | Execution replay |
| Need to practice a specific day or the open, repeatedly | Execution replay |
| Want to know if your size is realistic for the name | Execution replay |
Most traders should use more than one. Bar replay for volume of chart reps, execution replay when a strategy graduates from “I like this pattern” to “I’m going to risk money on this.”
What replay can’t teach you either
I’d rather say this than have you find out the hard way.
Sizing pressure is not simulated. Losing $800 of simulator money and losing $800 of rent are different experiences, and the second one changes your behavior in ways the first cannot rehearse. Anyone claiming their simulator solves trading psychology is selling something.
Borrow availability is real and we don’t model it. On a hard-to-borrow small cap, the short you take instantly in replay might not have been available live at any price, or might have carried a fee that changed the trade.
Your own market impact isn’t there. The tape is a recording. It doesn’t react to you. Small size, this barely matters. Get large in a thin name and reality would have pushed back in a way the recording won’t.
Knowing the day is a leak. Replay a session you’ve already traded and you’re not testing your read anymore. Fresh days only, or the experiment is contaminated.
None of that makes replay useless. It makes it a tool with a boundary, and knowing the boundary is what separates practice from theater. Everything inside the boundary is real: the tape, the quotes, the queue, the fills. Outside it, you’re still going to have to find out live.
Try it against whatever you use now
DAS Replay is in free beta. No card, no course, nothing upsold at the end.
The useful way to evaluate it isn’t to read more about it. Pick a day you’ve already traded or already studied, run it here, and compare the fills you get to the ones you believe you’d have gotten. If ours look generous, that’s the interesting outcome and I’d want to hear about it. If they look punishing, that’s the number your strategy has to clear before it’s worth real money.
Related: why your limit order didn’t fill on the queue mechanics chart tools can’t show you, and slippage has three sources on reading a fill log once you have one.