Futures or forex prop firm? Compare the structure, not the fee

Most traders approach the futures vs forex prop firm decision the way they compare two phone plans. They line up the entry fee, the account size and the profit target, pick the cheaper column and pay. That comparison hides the variable that decides the outcome. These are two different evaluation markets with different instruments, different risk granularity, different session structure and completely different politics around automation. The fee is the smallest item on that list, and it is the only one most people look at.

Nothing below is a quotation of any firm's current rules. Every parameter here is commonly structured a certain way, varies by account type and platform, and changes over time without anyone emailing you about it. Read the live rulebook on the firm's own site before you buy, and if automation matters to you, get that answer in writing.

Disclosure: this blog is published by PraxAI. We sell trading software, not evaluations, so we have no commercial reason to push you toward either side.

A futures contract is a standardized agreement listed on an exchange. It defines the underlying, the size, the minimum price increment and the value of that increment, and every participant trades the identical thing. Your order meets a central order book, the exchange publishes real volume, and the depth you see is the depth everyone else sees.

A currency pair at a retail broker is an over the counter product. There is no single central book. Your price comes from your broker's aggregation of liquidity providers, the spread can widen when conditions change, and position size is effectively continuous rather than stepped. Two forex accounts at two different firms can print different fills on the same second of the same news release.

That single difference is the root of almost everything else here. Standardized contract means fixed increments and a fixed value per tick. Over the counter pair means flexible increments and a variable spread.

On the forex side you can normally size down to one micro lot, stepping in hundredths of a lot. You shape the position around the stop distance instead of shaping the stop around the position, so a fixed risk budget resolves to almost any number you want.

Futures do not work that way. The smallest unit is one contract, and even with micro sized contracts the step is discrete. When a setup calls for a wider stop than usual, you cannot trade 1.4 contracts. You trade one, you trade two, or you skip. That quietly compresses the useful range of an account: below a certain buffer, the smallest position you can open already represents a meaningful slice of your daily loss allowance.

Here is an invented worked example, not a quotation of anyone's numbers. Say an evaluation gives you a $2,000 loss buffer and you decide no single trade may cost more than one percent of it, so $20. On a pair where one micro lot risks about ten cents per pip, a twenty pip stop costs roughly $2 and you can scale cleanly toward your limit. On a futures contract where one tick is worth $5 and your stop sits eight ticks away, the smallest possible position already risks $40, double the rule you just set. The numbers are invented, the shape of the problem is not. Forex lets you fit the risk to the trade. Futures often lets the contract size decide whether the trade exists. The same buffer logic decides how large an account you should buy in the first place, which is [why the size you can afford is not the size you should buy](/blog/ftmo-account-sizes-which-to-buy).

Both sides use drawdown limits, but the models differ. On the forex side the most common shape is a static maximum loss measured from the starting balance, plus a separate daily loss limit. Once you are in profit, the hard floor usually stays where it started.

On the futures side, a trailing threshold is much more common. The floor follows the account upward, in many programs tracking unrealized profit during the session, and stops trailing only at a defined point. That punishes a very ordinary behavior: letting a winner run and giving part of it back can breach a limit that did not exist when the trade was opened. If you have not internalized how that mechanic moves, read [how a trailing drawdown floor actually behaves](/blog/trailing-drawdown-explained) before comparing anything else, because it decides which strategies are viable at all.

Firm level detail matters more than market level detail here. Two futures programs can price their evaluations almost identically and treat the trailing floor very differently. A structural comparison of two well known ones sits in our [breakdown of two futures evaluation models](/blog/apex-vs-topstep). Treat that post and this one as a map, not a rulebook: the binding numbers live on each firm's site and change without asking your permission.

The currency market never really closes between the Sunday open and the Friday close, and its liquidity concentrates in the London and New York overlap. An automated system there mostly needs [a session filter and an honest plan for the thin hours](/blog/forex-session-times-ea-trading), when the spread widens and the stop you calculated on paper is not the stop you get.

Futures sessions are shaped by the exchange instead. Contracts commonly have a daily maintenance break and a daily settlement, and evaluation programs frequently add a flatten by time of their own, after which open positions may be closed for you or counted as a breach. Holding through settlement is treated differently too, since the margin on an overnight position is usually not the intraday one.

For a discretionary trader that is a scheduling detail. For an automated system it is a design constraint. A system that assumes it can hold through a quiet stretch and manage the position later has to be rebuilt around a hard flatten time. Treat the exchange calendar and the program's own flatten rule as two separate constraints, and read both before you write a single line of logic, because only one of them is published by the exchange.

This section decides the choice for a lot of readers, and it is also where the worst information circulates.

On the forex side, retail prop firms overwhelmingly run on the MetaTrader and cTrader family of platforms, which shipped with a scripting layer for automated strategies. Automation there is commonly permitted under written conditions: no latency or arbitrage exploitation, no strategy built to game the simulated fill engine, limits on copying one signal across unrelated accounts, and usually a requirement that you alone trade the account. Those conditions are checked. We covered what firms look for in [what makes a bot allowed or banned](/blog/are-trading-bots-allowed-prop-firms), and how the written policies differ in [the current state of EA permissions](/blog/prop-firms-that-allow-eas-2026).

On the futures side, assume nothing. Automation policy varies by firm, by platform and by stage. A program can treat the evaluation phase differently from the funded phase, require a human present at the machine, allow semi automated tools while prohibiting unattended execution, or prohibit automation outright. We deliberately do not publish a list of futures firms that allow bots, because the only answer worth anything is the one the firm gives you in writing, for your exact account type, on the day you ask.

So ask. Open a ticket, name the program you are about to buy, ask whether fully automated execution is permitted on it, and keep the reply. If the answer is vague, treat it as a no. A rule you cannot quote back is a rule you cannot defend when a payout review starts.

Fee comparison is where most people start, so count it correctly. The number that matters is not the sticker price of one attempt, it is the expected cost of reaching a funded account and holding it, including the attempts you fail on the way.

Forex evaluations are usually a one time fee per attempt, discounted often enough that paying full price is a choice. Futures programs commonly use a monthly subscription during the evaluation phase, sometimes with an activation cost when you convert, which makes the calendar itself billable. A trader who takes four months to pass a subscription priced evaluation has paid four times, and will still call it the cheap option.

Both models punish the same behavior. Repeat challenge fees can add up to $2,400 or more a year for someone who keeps buying attempts without changing how the account is actually executed. And passing is not the finish line: only 1 to 3 percent of funded traders keep the account long term, a separate failure with separate causes, catalogued in [why funded traders lose the account](/blog/why-funded-traders-lose-the-account).

Neither market is the easy one. They fail different people, and the honest way to choose is to match the structure to how you actually trade.

If you decide to automate on either side, the platform question comes before the strategy question. PraxAI runs on MetaTrader 5, on cTrader through a cBot that is new and currently in validation with our first clients, and on NinjaTrader 8 for the futures side. Running on a platform is not the same as being permitted on a program. Whether any futures firm allows it is that firm's call, in writing, for your exact account type, and that answer can change after you get it. For the wider view of what is being used in evaluations right now, we went through the field in [the current landscape of automated tools](/blog/best-ai-trading-bot-prop-firms-2026).