Your fills, not a survey
Built from broker executions rather than a journal you write afterwards, so the analysis includes the sessions you would have skipped writing up.
No survivorship in your own logQuanify Replay · In development
A backtester that reviews you, not just the strategy.
Every backtester answers the same question: would this rule have made money. Almost none answer the one that matters more for a discretionary trader — what are you repeatedly doing that costs you. Quanify Replay is being built for the second one. It is not released yet, and this page is what it is being built to do rather than a claim that it already does it.
Where it is
In development. Not released, no date announced, and nothing on this page describes a feature you can use today.
It is here because people search for a futures backtester and reasonably want to know whether the platform they are already using is going to have one. The answer is yes, and this is what it is aimed at. When there is a date, this page will carry it.
Bar-by-bar replayStep through historical sessions at your own pace and trade them as if live.Building
Strategy backtestRun a rule over history with your own commission and slippage.Building
AI trade reviewPattern analysis over your own recorded fills — the mentor layer.Building
Available todayPaper accounts fill from your strategy's live signals with your costs applied.Live
If you want to test a strategy this week rather than wait, paper accounts already do the forward half of this.
What a backtest is actually worth
Worth saying plainly, since this is a page about a backtester we are building rather than one we are selling.
A backtest is an argument, not evidence. It is the best case for a rule under conditions that already happened, produced by someone who knew what happened next. Every honest backtester is a tool for rejecting bad ideas cheaply, not for confirming good ones.
The failure modes are well known and mostly unavoidable: overfitting, where enough parameter tuning fits the noise instead of the signal; survivorship, where the instruments that stopped existing are missing from the data; look-ahead, where a calculation quietly uses a bar it could not have seen; and cost omission, where a strategy that averages $30 a trade is tested without the $4 of commission and slippage that takes an eighth of the edge.
Which is why the part being built with the most care is not the simulation engine. It is what is done with the result afterwards.
The mentor layer
This is the actual point of the product, and it operates on something a normal backtester does not have: your real fills.
Because Quanify places the orders, every execution is already recorded — tagged by strategy, account and session, including the trades you would rather not have logged. That is a dataset about you, not about a rule, and it is where the answerable questions live.
Do you cut winners early after a losing session? Is your worst hour of the day the same hour every week? Does your average loss grow after a win? Are your best trades the ones where you waited for the retest, or the ones where you chased? Does your size creep on Fridays?
None of those are strategy questions and none of them show up in a backtest. All of them are visible in a few hundred recorded fills, and all of them are more fixable than the strategy is.
Patterns, not verdicts
The aim is to show you a repeated behaviour with the evidence attached, not to grade you. A tool that says "you are undisciplined" has told you nothing.
Specific and checkableCosts are in the numbers
Commission and slippage applied per side, so what you see is after the costs rather than the version that looks better.
Same as paper accountsWhat you can use in the meantime
Two things already do the forward half of what a backtester does, and both are available now.
Paper accounts fill from your strategy's own live signals with commission and slippage you set per side. They answer the questions a backtest answers badly — how often does this actually fire, what does a bad session look like, does it survive costs — using live market conditions rather than remembered ones. They are also incapable of reaching a live broker by construction.
Forward records on published strategies are kept separate from the backtest each was published with, so you can see how a rule has behaved since anyone could actually trade it. That gap is the most honest number on any strategy card.
Between them you get the part of a backtest that is hardest to fake: out-of-sample behaviour, at real cost, in a market nobody has seen the end of.
Common questions
Is Quanify Replay available yet?
No. It is in development with no announced date, and nothing on this page describes a feature you can use today. Paper accounts and forward-test records are available now and cover the forward-testing half of it.
What is Quanify Replay?
A backtesting and market-replay tool being built for the Quanify platform, with an AI review layer that analyses your own recorded fills for repeated behaviours rather than only scoring a strategy.
How is it different from other backtesters?
Most backtesters answer whether a rule would have made money. Because Quanify places your orders, it already holds your real executions — so the review layer can look at what you repeatedly do, which is usually more fixable than the strategy is.
Can I backtest a strategy on Quanify today?
Not historically. You can forward-test one: attach it to a paper account and it fills from its own live signals with your commission and slippage applied, or run it on a simulated broker account through real sessions.
Are backtest results reliable?
Treat a backtest as an argument rather than evidence. Overfitting, survivorship bias, look-ahead and omitted costs all flatter a result, and every one of them is easy to introduce by accident. It is a good tool for rejecting bad ideas cheaply.
Will it cost extra?
Pricing has not been announced for any part of the platform yet. When it is, it will be on the pricing page.
More on free tools
Losing 50% needs 100% to get back.A 50% loss needs a 100% gain to recover. Work out your drawdown, the return required to reach the old peak, and how many trades that is.How many contracts should you actually trade?Work out how many contracts to trade from your account size, your risk per trade and your stop distance. Free, instant, and it rounds down.What is a tick actually worth?Tick size and dollar value for every major futures contract, and what a move is worth at your size. ES, NQ, MES, MNQ, YM, RTY, CL, GC and more.Is your best day going to cost you the payout?Check whether your best day breaks your firm's consistency rule, and work out exactly how much more profit you need before it stops being a problem.Build it, test it, run it. Without the export step.Design a strategy, test it against history, and route it to your live accounts without exporting anything. In development — here is what it is being built to do.A 3:1 setup is worthless if you win 20% of the time.Enter your entry, stop and target. Get the risk-to-reward ratio and the win rate that makes it break even — the number that decides whether it is any good.Your win rate is not the number that matters.Win rate on its own decides nothing. Work out expectancy, profit factor, reward-to-risk and the break-even win rate your strategy actually needs.
Start on a simulated account.
Connect a sim or evaluation account and run the whole product against it before you point anything at live money. Nothing about the setup changes when you do.