The 7% That Existed Only in a Spreadsheet, and What It Can Cost a Trader

A drawdown limit only protects you if the loss path is one you can actually follow without breaking your rules. A bot can stay within its settings while the trader behind it reaches a point of fear, pressure, or second-guessing that was never part of the plan.

In September 1998, Long-Term Capital Management faced that gap at a scale few investors had seen. The fund’s partners included John Meriwether and Nobel Prize-winning economists Myron Scholes and Robert Merton. Its models and trades were built around relationships that had worked historically. After Russia defaulted on its debt in August, markets moved in ways those relationships did not contain. By late September, the Federal Reserve Bank of New York helped organize a private-sector recapitalization to avoid a disorderly unwind.

Roger Lowenstein documents the episode in When Genius Failed. The important detail is not that the fund used sophisticated models. It is that a strategy can make sense inside a model and still demand a loss path that the people funding it cannot calmly carry through.

A limit on paper can feel different on Friday

Friday is when the account review stops being theoretical.

The strategy may show a drawdown still inside the maximum you set. Stops may have executed. Position size may have stayed within the stated percentage of capital. No single trade may have violated the rules. Yet the equity curve has fallen for several sessions, and the next queued signal arrives while you are already wondering whether the system has changed.

That is the real test. A maximum drawdown is not permission to ignore discomfort. It is a boundary that needs to reflect three things: available capital, the time you need that capital to last, and the behavior you can sustain when losses cluster.

If your plan allows a 15% drawdown but you will override the next trade after an 8% decline, then 8% is closer to your practical limit. The remaining 7% exists only in a spreadsheet.

This matters for retail traders because a drawdown can change more than an account balance. It can change position sizing, confidence, willingness to follow stops, and the temptation to recover losses with a larger trade. What Happens When a 15% Drawdown Cuts Into Your Trading Runway? examines the capital side of that problem. The behavioral side deserves the same attention.

Compliance does not equal consent

A bot follows the rules it receives. It cannot tell whether you understood the experience of following those rules through five losing trades, a gap through a stop, or a week where every setup feels late.

Before using any automated signal process, review the possible loss path, not only the expected return or win rate. Ask concrete questions:

• How many consecutive losses appeared in the backtest?
• What was the maximum drawdown, and how long did recovery take in that test?
• Did the test include periods of abnormal volatility, thin liquidity, or large gaps?
• What happens if your broker fills a stop worse than the planned exit?
• At what account decline would you pause new entries and review the system?

Backtests help describe a strategy’s past behavior. They do not guarantee future fills, liquidity, or emotional readiness. The same backtest can also produce different broker results once spreads, slippage, order handling, and market conditions enter the picture.

LTCM’s problem was larger and more complex than a retail trading account. The mechanism still applies: a model can define risk with precision while the real environment produces a loss path that changes the decision-makers’ ability to hold the position.

Use the approval gate before the pressure builds

An approval gate creates a deliberate pause between a signal and an order. That pause has value when it is used to check risk, not to chase certainty.

For each queued trade, review the planned entry, stop, position size, exposure to related positions, and the loss if the stop fills worse than expected. Then compare that risk with the account’s current drawdown and your prewritten limits. If the trade still fits, approve it. If it does not, reject it and record why.

The record matters. A trading journal can show whether you are rejecting trades because the setup changed, because portfolio risk rose, or because recent losses made you abandon a plan you previously accepted. Those are different problems and need different responses.

A quiet approval process also makes it harder to hide from the question that matters: “Would I choose this risk again, given what the account has already experienced?”

In 1998, the problem at Long-Term Capital Management was not solved by believing harder in the original assumptions. The positions had to be confronted in the market as it was. Treat a drawdown the same way. Review the actual loss path, reduce risk when your rules require it, and only approve the next order when you understand what it can cost.

Educational content, not financial advice.