Trading Your Own Capital: The Metrics That Matter When Nobody Is Enforcing a Drawdown Rule

We asked several hundred traders one question over the past two months: "what are you trading right now, prop firm or personal?" A large share of the answers were some version of "own capital", "self trading", "mine tho". Which was a useful thing to learn about our own writing, because nearly every article on this blog frames the value of a journal around a rule those traders do not have.

This one does not. Nothing below assumes a challenge, an evaluation, or a firm.

The Hidden Advantage of a Prop Rule (and Why You Should Steal It)

A prop firm's drawdown limit is usually described as a constraint. It is more accurately a subsidy. The firm pays to compute, in real time, three things most traders will not compute for themselves: how far you are from a loss you cannot come back from, how much of today's budget you have already spent, and whether the position you are about to open is compatible with both. Then it enforces the answer without asking your opinion at 2am.

That enforcement is the entire value. It is not that funded traders are more disciplined — it is that the cost of indiscipline arrives immediately and impersonally, instead of six weeks later disguised as a bad month. When you trade your own capital you keep all the upside a funded trader gives away, and you lose the one mechanism that was doing your risk management when you did not feel like doing it.

"A limit you can move is not a limit. The point of writing it down before the drawdown starts is that the version of you inside the drawdown is not the one you want negotiating with it."

So steal the mechanism, not the rule. You do not need FTMO's 10%, or Apex's trailing floor, or anybody's daily loss cap. You need a number, derived from your own data, written down somewhere that is not your head, and checked before you size a trade rather than after you close one.

Setting a Self-Imposed Drawdown Limit That Actually Binds

Most self-imposed limits are round numbers picked because they sound serious. "I stop at 10%." The problem with a round number is that it has no relationship to how your strategy actually behaves, so it fails in one of two directions: it either fires constantly during perfectly normal variance, which teaches you to override it, or it sits so far below your realistic worst case that it never fires at all.

The better source is your own trade history. Take your closed trades as a sequence of R-multiples, then resample that sequence across 1,000 or more simulated paths. Your edge stays intact — the same trades, in different orders — and what you get back is a distribution of maximum drawdowns rather than the single one you happened to live through. Read the P10: the depth that the worst 10% of plausible paths reach.

Illustrative worked example. Say 60 trades, expectancy +0.18R, risking 1% of the account per trade:

The round number is inside the strategy's ordinary variance. You will hit it doing nothing wrong, override it, and from then on the limit is decoration. The fix is not to raise the limit — it is to change the input. Drop risk to 0.7% per trade and the same P10 lands near 8%, comfortably inside a 10% line that now only trips when something is genuinely off. You size to make the limit survivable, then the limit means something. Numbers here are illustrative; run yours.

Two practical notes. Below roughly 30 to 50 trades per setup you do not have enough data for this and should use a deliberately conservative fixed percentage instead. And write the limit down as a sentence with an action attached — "at −8R from equity high I cut size in half and review the last 20 trades" beats "I stop at 10%", because it survives contact with the moment it applies.

The Four Metrics That Replace "Did I Pass?"

A challenge gives you a binary at the end of it. Self-funded, there is no end and no binary, which is why so many independent traders fall back on account balance as the only scoreboard — and balance is a terrible scoreboard, because it moves for reasons that have nothing to do with whether you traded well this month.

Four numbers do the job instead:

The first two are strategy questions and well covered elsewhere — see expectancy and SQN for the mechanics. The third is the early-warning system: comparing a rolling 30-trade window against your lifetime figures is how you notice an edge decaying while it is still a small problem.

The fourth is the one that matters most here and gets tracked least. Planned-versus-actual risk adherence measures you, not the strategy: for each trade, the risk you intended versus the risk you actually took. Funded traders get this policed for free. Self-funded, it is invisible unless you log both numbers — and the pattern it exposes is remarkably consistent. Risk creeps up on losing days and after losing days. That single correlation, visible only in your own log, explains a large share of accounts that die with a positive-expectancy strategy still attached to them.

Position Sizing Without a Rulebook

With no firm dictating maximum size, sizing becomes a decision you make several times a day, which means it degrades under exactly the conditions where it matters. Two bounds keep it stable.

The floor is fixed-R. Every trade risks the same fraction of the account, so results across trades become directly comparable and no single position can dominate the record. Fixed-R sizing is not optimal in a growth-maximising sense, and that is the point — it is robust to being wrong about your edge, which is the situation you are usually in.

The ceiling is Kelly. The Kelly criterion gives the mathematically growth-optimal fraction given your edge, and it is useful mainly as an upper bound you should never approach. Full Kelly assumes you know your win rate and payoff ratio exactly; you do not, and the penalty for overestimating is severe and asymmetric. Half Kelly or less is the practical zone, and if your fixed-R number is already well below it, that is a healthy sign rather than timidity.

One more thing that separates own-capital accounts from funded ones: size off account percentage, not fixed dollars. A funded account's balance is roughly static between payouts. Yours compounds, and it also takes withdrawals. If "1R = $200" was set when the account was $20,000 and the account is now $31,000, you are quietly trading at 0.65% while believing you are at 1%; run it the other way after a drawdown and you are at 1.4% while believing the same thing. Recomputing 1R from current equity keeps the risk you think you are taking equal to the risk you are taking, and keeps your account balance and journal in the same reality.

The Failure Mode Unique to Self-Funded Traders

There is one failure mode that a funded trader is structurally protected from and you are not: nothing stops you. At 2am, after a bad session, at three times normal size, on an instrument you do not trade, there is no daily loss limit to hit, no account to breach, no automated flatten. The only thing standing between you and that trade is you, at the exact moment you are least qualified for the job.

This is not a discipline lecture, because discipline is not a plan. It is an argument for a mechanical pre-trade gate — a small set of conditions you check before sizing, in the same order, every time:

The journal is the enforcement layer here, and it works through a slightly unglamorous mechanism: knowing the trade will be logged, tagged, and visible in a comparison you will run next week changes what you are willing to do at 2am. That is the whole trick. It is also the reason a log you keep only when things are going well is worth close to nothing.

Withdrawals, Deposits and Why They Wreck Your R Math

Funded accounts have clean capital histories. Personal accounts do not: you top up after a rough quarter, you pull money out for something unrelated, and you fund the account in stages rather than all at once. Every one of those events corrupts a specific set of numbers, and it is worth being precise about which.

R-multiples survive intact. An R-multiple is the trade's result divided by the risk taken on that trade, so it is indifferent to your balance. A +2R trade is +2R whether the account held $5,000 or $50,000 that day. This is the strongest argument for R as the primary unit of a self-funded log — it is the only common metric that a deposit cannot distort.

Everything percentage-based breaks. Specifically:

• Percentage return — a $10,000 deposit into a $20,000 account looks like a 50% gain to any naive calculation.
• Drawdown depth as a percent of balance — a withdrawal shrinks the denominator, so an unchanged dollar drawdown suddenly reads much deeper.
• The raw equity curve — vertical steps on capital-event days that are not trading results, which quietly poison any drawdown statistic computed from that curve, including a Monte Carlo run on it.

The handling is straightforward. Log every deposit and withdrawal as a dated capital event rather than letting it disappear into the balance. Keep performance analytics in R. When you do want percentage figures, compute them between capital events, not across them, and use a time-weighted return if you want one number for the year. Recompute your 1R from equity after each event, so the sizing you are using matches the account you actually have — this is the same discipline as the sizing point above, and forgetting it after a withdrawal is one of the more common ways a well-managed account starts silently over-risking.

What to Track If You Ever Plan to Attempt a Challenge Later

Plenty of self-funded traders eventually take a shot at a funded account, usually for the leverage rather than the discipline. If that is possible for you, a few extra fields cost nothing now and turn a future evaluation into a formality rather than an experiment.

• Daily P&L in R, not just per-trade R. Almost every firm's hard constraint is a daily loss limit, and a daily series is the only way to know how often yours would have breached one.
• Maximum intraday excursion, not just closed results. Trailing-drawdown firms measure your equity high including open profit. A flat day that spiked +2R and gave it back is a non-event in your log and a real cost under those rules.
• The distribution of your daily results, not just the total. Several firms apply consistency rules where a single outsized day disqualifies an otherwise passing account.
• Trades per day and time-of-day tags. Minimum-trading-day requirements are easy to satisfy and easy to fail by accident.

With that data present, the readiness question stops being a guess: run the same Monte Carlo you used to set your personal floor, overlay the firm's actual drawdown limit and profit target, and read the share of paths that pass. That is the full readiness workflow, and the traders who breeze through evaluations are almost always the ones who had a real log before they had a challenge. Rules vary by firm and change often — verify current terms directly with the firm before paying an evaluation fee.

And if you never attempt one, none of the above was wasted. It is the same log, answering the same question the firm would have been asking on your behalf: how much can this realistically lose, and am I sized for that.