The road to $3,000 — and then to six of them

The setup

One strategy per account. Six accounts, running at the same time, not one after another. Each engine trades the contract configuration it was tested with, and each account is judged on its own — an account never borrows another account’s cushion. That single rule is the whole design of the portfolio, and it is why the drawdown column below is reported per account and never blended.

The mark we walk toward is $3,000 in an account. It is a common size for a first milestone, and it is small enough that the question “how long, and how deep?” has a meaningful answer.

The journey, on a calendar

Six lanes, one per account, all opened on the same day — Sep 15, 2025, the latest date all six engines have tested history for. The hollow circle is that shared start. The solid circle is the day that account’s cumulative result first reached $3,000. The faded continuation is the rest of the window: the journey does not stop at the milestone.

Under every lane, in red, is what that road cost to travel: the deepest fall from that account’s own high-water mark on the way to the milestone. The red bar is measured against the $3,000 target, not against time — a bar half the length of its road means the drawdown was half the size of the prize. Read it as “how much of the target did the hole eat,” and judge the engine on that, not on the date it finished.

Window: Sep 15, 2025 to Aug 7, 2026 — every account measured from the same first day, so no engine gets credit for a head start. Hypothetical, backtested results.
These roads were walked straight through every major news day in the window — FOMC, NFP, CPI and PPI included. Nothing was paused, skipped or filtered out for being inconvenient. And no parameter was optimized, re-fitted or tweaked to make the journey look better than it was: these are the same settings the engines were tested with, run forward across the window as they stand.

The same six, as equity curves

The same accounts, drawn as cumulative dollars from that shared start. The dashed line is the $3,000 mark; the dot is the day that account first touched it.

Hypothetical, backtested results, all rebased to $0 on Sep 15, 2025.

“Drawdown to get there” is the deepest fall from that account’s own high-water mark on the way to the milestone, in dollars, for that account alone — never a blended figure.

Reading it honestly

All 6 accounts reached the $3,000 mark inside this window, and they took very different amounts of time to do it — Aperture in 88 calendar days, Iris in 296. That spread is not a defect. It is the entire reason for running six of them: the fast one is not the same engine every year, and nobody knows in advance which one it is this year.

The drawdowns matter more than the dates. The deepest one on the way to a milestone was −$1,634, on a single account. Six accounts, six separate holes to climb out of, none of them shared. Over the 11 months from the shared start the six accounts together produced $30,593 across 1,337 trades.

What this page cannot tell you is what the next twelve months look like. A tested history is a description of a road already walked. It is not a promise about the one ahead, and the results here carry all of the limitations set out at the foot of this page.

Every account gets $2,000. That is the whole test.

It does not matter which kind of account it is. An evaluation account, a funded account, or $2,000 of your own money in a brokerage account — the arithmetic is identical. All three give the engine the same thing: $2,000 of room, and no more.

• Evaluation account. A $50,000 evaluation with a $2,000 trailing drawdown limit is not a $50,000 account. Touch the limit and the evaluation is over, whatever the balance says.
• Funded account. Same limit, higher stakes: now it is a live account with the same $2,000 of room, and breaching it ends the account rather than a trial.
• Your own $2,000. No rulebook, no trailing threshold — but the money runs out at exactly the same place. The market does not care whose rule it was.

So the honest way to size any of this is: each account has $2,000 of room, and an engine either fits inside that or it does not belong on the account. Six accounts at $2,000 each is $12,000 of room committed, whichever of the three routes you took to get there.

Intraday margin on a handful of micro contracts is small change against that room — on the order of $50/contract on MNQ · $50/contract on MES · $100/contract on MGC — so margin is not the binding constraint here. The drawdown limit is.

Position size is not fixed on these engines: risk is sized in dollars per trade, so a wide stop buys fewer contracts and a tight one buys more. The number worth judging is therefore not the lot count — it is the drawdown that sizing actually produced, which is what the table below reports.

Does each engine actually fit?

This is the column that decides everything. It is the deepest drawdown each account took on the way to $3,000, set against the $2,000 it is allowed to lose:

Margin rates above are illustrative discount-broker intraday figures. Brokers set their own margin and can raise it at any time, including around the news days these engines trade through. Overnight margin is far higher; these are intraday engines.

All six fit, and the one that came closest — Ripple at 82% of the limit, $366 of room to spare — is the one to watch, not the one with the biggest number at the end. A drawdown deeper than what has already been seen is not a surprise; it is the normal condition of the next twelve months. That is also why each engine gets its own account: six drawdowns of this size in one account would have ended it.

This is exactly why the site never quotes a blended drawdown, and why no engine here is offered at a size that has already used most of its room.

Renting the room instead: a third-party funded-account program

The other route rents the capital instead of posting it. A funded-account program charges a monthly fee while you trade its evaluation, then a one-time activation fee once the evaluation is passed — after which the account is theirs, the rules are theirs, and the payout split is theirs. Round numbers, no firm named:

Hypothetical figures for education. Funded-account programs differ in fees, drawdown rules, profit splits and payout schedules, and a passed evaluation is not a payout. Read any program’s own rules before assuming these numbers apply to it.

Setting the cost against the road

Every figure in this comparison is hypothetical. The costs are real prices; the result they are set against is a backtest, not money anyone was paid. Over the 11 months drawn above, the six accounts together produced a hypothetical $30,593. Against that:

• Own capital: $12,000 of room committed across six accounts, plus $2,189 of subscription over the window — software running at about 7% of the tested result. Tested result against room committed: 2.55× over 11 months.
• Funded-account route: $6,389 all-in over the window — about 21% of the tested result, before that program’s own profit split, which is not modelled here.

Two honest warnings about that arithmetic. First, it divides a backtested result by a real cost: the cost is certain and the result is not. Second, and more important, $2,000 per account is a hard ceiling, not a cushion — the tightest engine here already used 82% of it once. An account run at the edge of its limit does not need a disaster to end; it needs a bad month. Every sizing decision on this site is made against that ceiling first and the return second.

The order it happens in

1. 1One account, one engine. Run a single account first. Not because the others are riskier, but because the first thing to verify is that your fills resemble the tested fills on your platform, your data feed and your connection.
2. 2Compare, do not hope. Every engine writes a per-trade record. Set it against the published history for the same days. If they diverge, that is information you want in week one, not in month six.
3. 3Add the second account — a different market. The point of the second account is not more size. It is a different instrument and a different behaviour, so that one bad week in one market is one bad week, not the whole book.
4. 4Widen to six. Six accounts, six engines, three markets, running in parallel. Scale wide, not tall.
5. 5Review on a schedule, not on a feeling. Engines are reviewed on a fixed cadence against their own baseline. An edge that fades gets trimmed or retired — on the calendar, not in a panic.

Window shown: Sep 15, 2025 – Aug 7, 2026. Figures are net of commission and slippage as configured in the tests. Trading futures involves substantial risk of loss.