About
What Quantradin is
A trading lab for you and your AI. You build a strategy in plain rules, our engine grades it on real prices and real spreads, and you trade it on a paper desk. Not one real dollar moves anywhere in the product.
The short version
Most people learning to trade never get to find out whether the strategy they just read about actually works. The tools that promise to tell them are usually selling something, and a backtest is extremely easy to flatter. So the number comes back beautiful, the real account doesn’t, and nobody learns anything except distrust.
Quantradin is the opposite arrangement. The engine is built to take results away from you when they aren’t real — and to show you exactly which rule cost you the flattering version. Then you put what survived on a paper desk marked at real quotes and watch it forward, with nothing at stake but your own attention.
Your AI can do all of it too. A desk comes with its own key pair, so an agent can research, backtest, deploy and report back on the same numbers you see. There is no better answer behind an API.
Why it exists
Most backtests are advertising. That isn’t a swipe at anyone in particular — it’s a structural fact. If a tool’s job is to make you feel good about a strategy, every ambiguous modelling choice drifts the same direction: fill at the midpoint, tune and score on the same stretch of history, quietly drop the trades that couldn’t have happened. None of it is fraud. All of it inflates.
We built the version that can’t do that, because the rules are enforced in the engine rather than left to a setting. Every one of them makes our own numbers look worse. That is the whole point.
The backtester that doesn’t lie — including when the truth is that your idea, or ours, didn’t hold up.
What “honest” concretely means
Not a value we aspire to. Four mechanical rules, enforced in code and pinned by tests, that we cannot switch off for a nicer screenshot:
You buy at the ask and sell at the bid
Never the midpoint. The midpoint is an average of two prices, not an offer anybody made you, and pricing every fill there is the single fastest way to invent an edge. The spread lands on the ticket as a cost. Where we have a recorded two-sided quote we use the real one; where we don't, we model the spread and label it as modeled, on every result it touches.
A signal never sees tomorrow
A rule reading a given day sees that day and every day before it, and nothing after. The trade is entered at the NEXT bar's open, never the bar it fired on. An option's price path starts strictly after the signal date.
It's graded on history the rules never saw
The history is split. Choices like the stop are settled on the first stretch, then the result you're shown is scored on the second. When we shipped that split, every number in the product got worse and truer on the same day.
Take-profit is left out of the score
A profit cap makes a curve look calm without earning anything, so it never enters the grade. You can still run one — it works on your live paper desk. We just won't count it as evidence.
There are more — thin contracts get skipped instead of imagined into the results, delisted companies stay in the history so a winner isn’t just a survivor, and a bot can’t fill outside real market hours. They all follow from the same instinct: if it couldn’t have happened, it doesn’t count.
Where the engine came from
It wasn’t written for this website. The backtest engine was lifted out of a working trading rig that had already been run in anger for years, then re-validated so it grades a reference strategy identically before and after the move. That check runs on every change we make to the maths.
That history is why the rules read the way they do. Each one is a lesson that rig paid for: profit targets flatter a curve without earning it, a strategy that only works on one stretch of history isn’t a strategy, and a deep in-the-money option quietly eats the very thing you bought it for. We didn’t reason our way to those from first principles. We got them wrong first.
The product is a smaller, safer thing than that rig: US markets, daily bars by default with minute bars where the data supports it, and paper money only.
Our engine advises. It never decides.
People come here to test their own ideas, so the engine is not a gatekeeper. Every result carries our read — LIVE, SHADOW or SHELF, with an expected value and how much evidence it rests on — and that read sits beside your judgement instead of replacing it.
- You choose whether a bot runs live on your paper desk, shadows, or sits on the shelf. A read never blocks a deployment.
- You keep your own parameters. Set a 60% stop and we grade the 60% stop — and tell you beside it that our engine reckons 50% would have done better. A suggestion, never a substitution.
- Deploying a bot doesn’t secretly run a backtest. Not every strategy can be honestly graded, and grading uninvited is the engine deciding instead of advising. Ungraded is a perfectly normal state.
We will never tell you your strategy is worthless. We’ll tell you what our engine found, show you the workings, and leave the call where it belongs.
What Quantradin is not
Not a broker
We hold no money and route no orders. There is no brokerage account behind your desk, and nothing to fund.
Not an adviser
Nothing here is investment advice or a recommendation to buy or sell anything. A verdict is our engine's read on a set of rules, not guidance for your money.
Not real money
Every desk is a simulation funded with $100,000 of pretend cash. The prices marking it are real; the money never is. Simulated results carry inherent limitations — a paper fill has no market impact and never gets a partial.
Not a signal service
We don't sell picks, alerts to follow, or a strategy that's supposed to make you money. You bring the idea; we grade it and let you paper-trade it.
Who builds it
A very small team. Small enough that the person who wrote the fill model is the person who reads support@quantradin.com — which is deliberate, because a support reply that has to guess how the engine prices an exit is no use to you.
We’re a trader and an engineer building the tool we wanted while learning this, so the roadmap is short, the copy is written by the people who wrote the code, and the unflattering changes get published too. You can read every one of them on what’s new, including the features we built and then switched off because the data behind them wasn’t good enough to be honest with.
It is free while we’re early — no card, no trial clock. When that changes, it’ll change on the pricing page in daylight, not in an email you have to catch.
Where to go next
- Learn the basics— what a backtest can and can’t tell you, and how to read one of ours.
- Join the Discord — Ask anything, show your desk, read what other people's bots did today.
- Support — or email support@quantradin.com. A person reads it.
- Try the bench on the front page — a real graded run, no signup, nothing withheld before the result.
Simulated results only. Hypothetical performance has inherent limitations. Not investment advice. Paper trading uses zero real money. The longer versions live in our disclaimer and terms.