From fills, not entries
The journal is built from broker executions, including the ones you would rather not type in. There is nothing to remember to record.
Tagged by strategy, account, sessionTrading bots
The bot is your strategy. This is the part that trades it.
"Trading bot" describes two very different products sold under one word. One sells you the decisions. The other executes decisions you already have. Knowing which you are shopping for saves a lot of money.
Two things called a bot
A bot that decides. Someone else's model, sold to you as a black box with a performance claim attached. You cannot see the rule, cannot test it independently, and cannot tell whether last month was skill or a market that suited it. Most of what is marketed as a trading bot is this.
A bot that executes. A rule you can read — your PineScript, your Python, your own logic — and infrastructure that turns its output into real orders on your accounts. The decision stays yours and stays inspectable.
Quanify is the second kind and does not sell the first. There is no black box here, and no strategy you cannot see the source of.
A bot that decidesYou cannot inspect the rule, so you cannot tell a good month from a lucky one.Opaque
A bot that executesYou wrote the rule or you can read it. The infrastructure is the product.Yours
Where Quanify sitsThe executor. Bring a strategy; keep the source; own the decisions.Executor
What we will not doSell you a signal with a return attached. Nobody can guarantee that, and the ones who say so are the tell.
What an executor has to get right
Turning a signal into an order sounds like a small job until you enumerate what can go wrong between the two.
- A live broker session
- Held server-side, not on your desk. If the session lives on your laptop, so does your uptime.
- Sizing per account
- Your accounts are different sizes. A signal that says "buy 2" has no idea which account it is about to trade.
- Limits enforced before the order
- Checked at dispatch, not reconciled afterwards. A limit you find out about from a fill is a report, not a control.
- Duplicate protection
- Sources retry, brokers retransmit. De-duplication is per execution inside a 60-second window per account, so a repeat lands once.
- Partial fills handled as partials
- A large order rarely fills in one piece. Every partial is mirrored on its own execution id rather than guessed at the end.
- A record from the first order
- Every fill tagged by strategy, account and session. A bot with no audit trail is a bot you cannot debug.
Nothing starts itself
The expensive failure with an automated system is not a strategy that underperforms. It is a strategy that was live when you thought it was not.
So a new strategy is created powered off, every account attached to a trader lands disabled, and a trader will not report itself armed until the engine has confirmed the routing. Configuring is not arming — they are separate actions, deliberately.
Turning it off is equally literal: new signals stop reaching your accounts, and nothing is closed on your behalf. Quanify never flattens a position for you, not on power-off, not on cancellation, not when a strategy is deleted.
How to tell if a bot is worth running
The claims are unfalsifiable. The behaviour is not, and it can be checked in a week without funding anything.
- 01Read the rule. If you cannot see what it does, you cannot tell whether a good month was the strategy or the market. That is not a small caveat, it is the whole question.
- 02Run it on paper with real costs. Built-in paper accounts fill from the strategy's own signals with commission and slippage you set. A strategy that only works at zero cost is not a strategy.
- 03Watch the losing days, not the winning ones. The number that decides whether you can actually run something is its worst stretch, not its average.
- 04Check what happens when it is wrong. A signal that fires twice, a broker that rejects, a session that drops — what does the system do, and does it tell you?
- 05Size it against the drawdown, not the return. A strategy's worst historical stretch at your contract count is the number that decides whether you can run it.
The measurements that matter
Every fill is recorded from the first order, so the analytics are computed from what happened rather than from what you remembered to log. Win rate, expectancy, maximum drawdown, Sharpe, Sortino and Calmar are all derived from real fills.
Two of them are withheld when the sample is too thin to mean anything: Sortino below three losing months, Sharpe and Sortino below twelve months. A ratio computed from six trades is a number, not a measurement, and showing it would be worse than showing nothing.
Costs included
Paper accounts apply the commission and slippage you set, per side. A strategy that is only profitable before costs is a backtest, not a plan.
Per-side commissionWithheld when thin
Ratios that need a sample are hidden until there is one. Precision that the data cannot support is a way of being wrong confidently.
Sharpe and Sortino need 12 monthsCommon questions
What is a futures trading bot?
Software that places futures orders from a rule instead of a person. In practice it is two parts: the strategy that decides, and the executor that holds the broker connection and places the order. Quanify is the executor — you bring the strategy.
Does Quanify sell trading bots or signals?
No. There is no black-box strategy for sale here and no signal with a return attached to it. You bring a rule you can read, and the platform executes it on your own accounts.
Are futures trading bots profitable?
A bot is exactly as profitable as the rule it runs, minus costs that are charged per account. Automation removes the typing and the hesitation; it does not add edge, and it makes a losing rule lose faster.
Do I need to code to run a trading bot?
No. A TradingView alert with a webhook URL is enough, and the alert body is generated for you. Python and raw HTTP requests work identically if you prefer to write your own.
What happens if my bot sends a bad signal?
It is checked before an order is placed against an off-by-default guard that refuses to grow a position you already hold. Scaling out is always allowed, and so is a flip.
Can I test a bot without risking money?
Yes. Point it at a simulated or evaluation account, or use the built-in paper accounts, which fill from the strategy's own signals with commission and slippage you configure. Nothing about the setup changes when you later point it at a funded account.
More on automation
Automated futures trading, on accounts you already have.Automate a futures strategy on your own broker accounts. A webhook fires, Quanify risk-checks and sizes it, and the order reaches your broker.Your strategy, executed on your live accounts.Automate any futures strategy on Tradovate, NinjaTrader or TopstepX. Send a webhook; Quanify risk-checks, sizes and routes it. No VPS, no glue code.Automated NinjaTrader, without writing NinjaScript.Automate NinjaTrader orders from a TradingView alert or any HTTP request. No NinjaScript, no add-on, no machine left running overnight.Automating TopstepX, without writing the layer underneath.What automating a TopstepX account actually requires, what the hard parts are, and how to route a strategy to it without writing an execution layer.Your strategy, or one you can inspect.Run your own futures strategy through a webhook, or add a published Quanify strategy with its full forward record. How strategies work and are measured.Turn a TradingView alert into a live order.Send a TradingView alert to one webhook. Quanify resolves the contract, sizes the order, applies your risk rules and routes it to your broker.The Tradovate API, and what building on it really costs.What building on the Tradovate API actually costs you: OAuth, WebSocket sessions, contract resolution, reconnects. And what you get if you skip it.
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.