Clara Core: MNQ (2019–2026)

The edge comes from consistency over a large sample of trades, not from individual outcomes.

Short-term behavior is unpredictable. You could start during a strong week or a losing week. The market does not adjust to when you begin. What matters is that the system continues executing the same rules regardless of recent results.

Starting during a stronger period may produce faster early gains. Starting during a drawdown period means temporary losses before recovery. Both are part of normal system behavior. The May 2019 – June 2026 view shows how these periods balance out over time.

On a prop-firm account with a trailing drawdown, the account-loss level follows your equity up until you lock in a cushion. Early on, before you have banked profit above your starting threshold, you are closest to losing the account. The chart below shows the $300 model's equity over the full period, with the drawdown-from-peak underneath. The lower panel is what a trailing drawdown actually watches.

Across the full period, the deepest drawdown from a peak on the $300 model was −$1,406 (measured intra-trade). That single number drives the two decisions below.

When to increase risk or add an account: size your buffer to the Monte Carlo worst case, not the historical average. Historical maximum drawdown on this model was about −$1,406, but that is a single sequence of history. The Monte Carlo analysis reshuffles the same trades 10,000 times and shows drawdowns can reach roughly −$3,800 in an adverse order (the 5th-percentile, worst-case figure). So for full safety, wait until you have built about $3,900 in realized profit above your trailing floor before increasing risk or adding a parallel account. At that level, even a worst-case Monte Carlo drawdown would not reach the account-loss level.

How much buffer you require is a risk choice. The most conservative approach uses the Monte Carlo worst case (~$3,825). A more moderate approach uses the Monte Carlo median (~$2,545), but note that half of simulated sequences drew down worse than the median, so this leaves less protection. The most aggressive uses only the historical actual drawdown (~$1,406), which assumes the future resembles the one path history took. We use the Monte Carlo worst case ourselves; the others are shown so you can size to your own tolerance with the trade-off stated plainly.

We would not scale on anything thinner than the historical maximum. And if the strategy is ever drawing down worse than the Monte Carlo worst case, that is a signal to stop and re-evaluate, not to add more buffer.

When to take payouts (earlier): payouts are a separate decision. Because a payout usually reduces your balance and moves you back toward the trailing floor, taking a modest payout once you are clearly above breakeven, with a small cushion of a few hundred dollars, banks real money without re-exposing the account. Waiting for the full worst-case buffer before taking anything can mean a long stretch with nothing in hand.

The realistic part: this is slow and sequential. The table below shows how long a single drawdown took to recover to a new peak, per year. Some drawdowns lasted one to two months. Building a clean $3,825 buffer on a model that averages around $7,000 a year can realistically take several months, and in a difficult stretch, a large part of a year. Scaling is something you earn over time, not something you do in the first month.

Drawdown depth is measured intra-trade (including the open-position dip). Recovery is calendar days from a peak until a new peak is made. Figures are for the $300 reference model and scale by dollar risk across the other models. These are historical backtest results and do not guarantee future behavior.

Clara Core uses several entry and exit models. Not all trades behave the same; this is by design, allowing the strategy to adapt to different market conditions.

Trades can close through several mechanisms:

Because of this, some trades will close early and protect capital, some will run and capture extended moves, and some will give back unrealized profit before exiting. This variation is expected and intentional. The system is designed to balance win rate, drawdown control, and capital preservation over time.

The largest win, up close. January 4, 2022. The opening range was 36 points, tight enough to size three contracts on the $300 model. The system went short at 10:00 AM New York time, the market sold off, the trailing exit followed the move down, and the trade closed at 11:34 for $1,461 net: the largest single win in the seven-year record, and also the best single day. Worth saying plainly: that trade is an outlier. The average win is $133. Outliers like this happen a few times a year and the trailing exit exists to catch them, but the system is carried by the ordinary trades, not the memorable ones.

This report is generated directly from NinjaTrader strategy data, including all entries, exits, and execution costs. Every trade carries a $1.90 round-turn commission per contract, $0.95 per side. That is NinjaTrader's free-account rate, all-in, and it is the highest commission anyone running this strategy is likely to pay: a funded prop account through a firm like Apex pays closer to half that. We publish at the highest rate on purpose, so that almost everyone's real commission comes in lower than the figures you see here.

All published results are net of that commission. The $1.90 round-turn is deducted from every trade in the data, so the figures you see are after-cost, not gross. Your own commission will most likely be lower, which means your own backtest will most likely look better than ours. That is the direction we want the error to run.

A full raw trade export is available for download. You can upload this data into any analysis tool to review the system behavior yourself.

Live Execution Comparison
Clara Core acts only on bar close, never intrabar, so live fills track the backtest closely in principle. A full side-by-side execution record across account types will be published here as it builds, with the underlying account statements available for verification. Even running identical settings on the same day, results differ between accounts: fill speed, slippage, latency, broker and data-feed differences, account type, firm rules, and hardware or VPS environment all move outcomes by dollars in either direction. Your results will not match ours, or anyone else's, exactly; small differences are a property of live markets, not a defect. The record below started on July 23, 2026, tracking the same trades across a cash account, a Roth IRA, and two prop accounts against the backtest at the published $1.90 cost. It grows as trading days accumulate, and every day goes in, including the ones that run against us. A record this small is not evidence of anything yet. It is simply more than most in this space publish at all.

We do not post these daily. We also trade manually, swing trade, and test other strategies alongside Clara Core, so a daily feed would mix things that should not be mixed. What we do plan to document in full is running new prop-account challenges with Clara Core, journey included.

A note on manual intervention. Ours runs unattended almost all of the time. On the rare occasion we are at the screen and a trade has run well past its usual range, we sometimes step in, and one rule governs it: on the $300 model, never touch a trade that is under $150 in profit. Below that it runs to its stop, its target, or the session close, whatever comes first. Most days nobody is watching and none of this applies.

We are telling you this because it is true, not because we recommend it. We cannot measure whether it helps or hurts, and every figure published on this site comes from a fully automated backtest with no intervention in it at all. The moment you step in, your results stop being comparable to the record you are reading.

If you think you will intervene, do this instead: run a fully automated sim account alongside your live one, touch nothing on it, and compare after a week, a month, a quarter, using the free trade journal. In our experience the two end up in nearly the same place. The market is going to do what the market is going to do.

A second platform, as a check. Everything above is computed in NinjaTrader, because that is the platform you run the strategy on and the one whose figures you can reproduce. As an independent check, we run the identical logic in TradingView and compare the two. TradingView's 2-minute chart only holds about five to six years of history, so it cannot reach back to 2019; the cross-check therefore runs over the window both platforms cover, and the full side-by-side is in the next section.

All performance data on this website reflects Clara Core. Internal execution parameters may be refined over time to improve consistency without changing the strategy's core logic, risk structure, or behavior.

Updates will only be reflected in published statistics when changes produce a meaningful and verifiable improvement, such as a measurable reduction in drawdown or a demonstrable increase in profit factor. Minor internal refinements that fall within existing platform variance are noted but do not require a full data update.

Any refinement must prove itself on data it was not built on: it has to hold up across the full 7-year record and the live period that follows it, not just on the stretch that inspired it. A change that merely makes the historical backtest look better is rejected; optimizing to the past is the failure mode this industry runs on. The focus is always on long-term structural improvement, not short-term optimization.

Any meaningful update to published statistics will be announced on our social channels and reflected in the Performance section.

The published figures include commission but do not model slippage, and we would rather say that plainly than pretend otherwise. Two things work in your favor here. First, MNQ is one of the most liquid instruments in the world, so in normal conditions slippage is minimal; it is during high-volatility sessions that entry and exit prices can differ from the bar-close price the backtest assumes. Second, we publish at the highest commission rate in the audience, so there is already room on the cost side, and we track real live fills against the backtest in the open, so the actual gap is something you can watch rather than a number you have to trust.

One mechanism that reduces this gap: all exits are calculated from the actual entry price, not the signal price. This means live execution logic is identical to backtest logic, and real fills closely mirror backtest fills.

For a full explanation of how this works, see the Execution Model and Entry-Based Take Profit sections on the Strategy page.