Sim Trading vs Live Trading: What Changes When the Order Leaves

Sim trading and live trading differ in one structural way, and everything else is downstream of it: a simulated order is filled by software that never sends anything to a market, while a live futures order in a US market is executed on an exchange the CFTC has designated a contract market, and the transaction that results has to be cleared through a clearing organization the Commission has registered.

Fill quality, slippage, and what a regulated firm may call the resulting track record all trace back to that.

What is the difference between sim trading and live trading?

Sim trading — also called paper trading — means your orders are priced and filled by the platform instead of being sent anywhere. Live trading means the order leaves.

A designated contract market is an exchange operating under CFTC oversight, and in a US market a venue wanting to let people who are not eligible contract participants trade futures generally has to apply for that designation, absent an exemption or exclusion. Retail customers are among the traders the CFTC says a designated market may allow in. The designation runs off Section 5 of the Commodity Exchange Act, 7 USC 7, and what a market has to satisfy to obtain one is set out in that section and in Part 38 of the CFTC's regulations.

Clearing is a second step, and not optional. Under 17 CFR 38.601, a transaction executed on or through a designated contract market has to be cleared by a derivatives clearing organization registered with the Commission, with a separate route for security futures products, which may instead be cleared by an SEC-registered clearing agency. Such an organization substitutes credit: each side ends up facing it rather than the party that took the other side, by novation or an equivalent arrangement, and one that wants to clear futures has to register with the Commission before it starts.

One broker's published account of its paper trading, beside what the rules require on the live side in a US market:

Where does a simulated futures order go if it never reaches an exchange?

A simulated futures order stops at the platform that took it. The software prices it and reports a fill; nothing is sent onward. The number looks convincing because the input is genuine: in one broker's published paper-trading documentation, the price a simulated execution is given is worked out from the prices and sizes quoted by the market.

What the order misses is the live path's second half. It is not executed at an exchange and never reaches settlement at a clearing house, so no clearing organization ends up in the middle of it, and the credit substitution the CFTC describes never happens. None of that absence is visible on screen: same chart, same ladder off the same feed, same arithmetic on the position once you click.

Why do simulated fills differ from live fills?

Because a simulated fill is an estimate, built from a narrow slice of the market. Two limits stated in one broker's paper-trading documentation show the shape of it. A stop there is handled by the simulator, as the other complex order types are, without exception, so the moment it triggers and the price it gets are both the software's doing. Fills are drawn from the top of the book only, the depth behind the best bid and offer being out of reach.

It is worth resisting the assumption that simulation always flatters. The disclaimer the CFTC prescribes for hypothetical results allows error either way — results may have compensated too little or too much for market factors, a lack of liquidity among them — and NFA's Interpretive Notice 9025 takes the same two-way position, adding price slippage to the factors it names. The thinner your edge per trade, the more of your result the simulator is deciding.

What do US regulators say a simulated track record cannot capture?

US regulators have written the list down, which makes it useful even to a trader who will never advertise anything. Under 17 CFR 4.41(b), simulated or hypothetical performance of a commodity pool operator, commodity trading advisor or principal may not be shown unless a prescribed statement goes with it — the CFTC's own, or one prescribed by a registered futures association. Where the presentation is anything but oral, that statement has to be prominent and sit in immediate proximity to the figures.

Three things the prescribed statement puts on the record. Programs of this kind are, as a rule, put together with hindsight available. The trades were never executed, so what is shown is not a record of actual trading. And in the CFTC's prescribed wording at 17 CFR 4.41, results "may have under-or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity".

Notice 9025 adds a limitation with nothing to do with execution. Nothing is at stake in producing a hypothetical record, so it can only partly reflect the risk-related factors that shape a real one — among them a customer's or an advisor's capacity to absorb losses, and to stay with a trading program while they arrive.

Note who these rules reach. Section 4.41 addresses commodity pool operators, commodity trading advisors and their principals; NFA Compliance Rule 2-29 governs FCM, IB, CPO and CTA Members and Associates. Neither reaches your own practice account. What they list is still worth borrowing as a description of what a simulated record leaves out.

Why does NFA treat calling simulated results "live" as an abuse?

Because the label is most of the claim. Dressing a hypothetical record up as "live" or "real-time" presents it as actual performance, and Interpretive Notice 9025 lists that relabelling among the abuses NFA wrote the notice to stop, beside disclaimers shrunk towards illegibility or parked far from the figures they qualify.

What the notice requires instead reads as a checklist, whether or not any rule applies to you. Where a Member or Associate — FCM, IB, CPO or CTA — has under a year of actual results behind it, the disclaimer is pinned in front of the hypothetical figures rather than allowed to follow them; either way it has to carry as much prominence as the results it qualifies, which 9025 says will generally mean type at least as large. Every material assumption behind the figures has to be described as well, and the notice sets a floor for that: how the performance was calculated at all — settlement prices, say, as against real-time pricing — what commissions were charged and what management or incentive fees were taken, what became of profits as they accumulated, and the size of the opening balance. Those are the same dials a simulator hands you.

One rule closes the door rather than qualifying it. Compliance Rule 2-29(c)(4) stops a Member or Associate using hypothetical results for a trading system once three months of actual results for it exist, with a single exception at 2-29(c)(6) for material directed exclusively at qualified eligible persons.

Can a simulator be configured to behave more like a live account?

Partly, and it takes about ten minutes. Quantower — one of the platforms we support — puts its simulator's assumptions in front of you as settings rather than burying them: the commission charged, a delay on execution, the balance the test account opens with, the symbols covered, and the netting method applied to each. It will also run on a connection that cannot place a real order at all — the simulator supplies the execution instead.

Two of those settings do more work than the rest. An execution delay of zero assumes your order arrives the instant you send it; a commission of zero deletes a cost a real account pays. Commission charges are one of the assumptions 9025 obliges an FCM, IB, CPO or CTA Member to spell out, a fair signal of how much leaving them out can change the answer. Setting both honestly does not make a simulator predictive; it removes two ways it can be kinder to you than a market would be.

What does sim trading vs live trading mean for a funded account?

Accounts at Lumen Futures, evaluation and funded alike, are simulated accounts trading live market data, and the payouts on the profits are real money. Our own comparison of the simulated and live sides is blunt about the part that belongs here: fills on the live side are real and they are worse, and the strategies that lose most in the switch are the ones scalping a tick or two.

So the honest way to read a passed evaluation is as evidence that you stayed inside the rules — the drawdown first among them — rather than as a measured execution edge. The risk disclosure says plainly that most people who attempt this will fail, and if you are choosing what to trade in either environment, the Micro E-mini and E-mini comparison covers what contract size does to your risk per tick.

FAQ

Are simulated results always better than live results?

Not necessarily, and neither the CFTC's prescribed wording nor NFA's says they are. The statement 17 CFR 4.41 prescribes puts the error in both directions at once: a simulated result may have made too little of certain market factors, or too much of them, lack of liquidity being the example it gives. Interpretive Notice 9025 agrees and adds price slippage to the list. Treat a simulated result as a different measurement from a live one, not a reliably generous version of it.

Do CFTC and NFA rules on simulated performance apply to prop firms?

Those rules name their own audiences. 17 CFR 4.41 is addressed to commodity pool operators, commodity trading advisors and their principals. NFA Compliance Rule 2-29 applies to FCM, IB, CPO and CTA Members and Associates. Neither text names proprietary trading firms, and nothing here should be read as saying that it does. What they offer any reader is an authoritative list of a simulated record's limitations: no financial risk behind it, hindsight in its design, and market factors it may have allowed too little or too much for.

Why do stop orders behave differently in a simulator?

Because the platform, not a market, decides what happens to them. Stops are among the order types one broker's paper-trading documentation describes as simulated in a paper account without exception. Orders there are never executed at an exchange, so the moment a stop triggers and the price it is given both come from the software's model of the market.