You Are Down Two Percent With a Three Percent Daily Limit

[Fairman]

You Are Down Two Percent With a Three Percent Daily Limit

Situation: a 10,000-dollar account, a daily loss limit of 3 percent, which is 300 dollars, and two stopped trades so far today for a combined loss of 200 dollars. A valid setup appears. Your standard risk is 1 percent, 100 dollars.

Work through the options. Option one: take the trade at full size. The worst case, a stop at the planned level, brings the day to minus 300 dollars, exactly at the limit. Any slippage pushes you past it. The limit becomes a target to touch.

Option two: take it at half size. The worst case loses 50 dollars, bringing the day to minus 250, leaving a buffer of 50.

Option three: skip it and end the day.

What does your plan say? If it includes a rule such as "stop after two consecutive losses," the decision is already made. Rules written in advance exist precisely for this moment, when judgment is under pressure.

Consider the psychology. After two losses, the pull to recover is strong, and a trade taken to recover has a different quality from a trade taken because it meets the rules. Even a valid setup may be approached with tension, hurried management, and an urge to move the stop.

The mathematical cost of skipping is the expected value of one trade, perhaps 0.2R or 20 dollars. The cost of a breached limit may be far larger, in money and in discipline.

Verdict: skip, or at most take half size if the plan explicitly allows it.

Exercise: write the rule for your own account's behavior at two-thirds of the daily limit.

[Shadow Trader]

MT4's strategy tester reports a modelling quality figure, and 99 percent looks like a seal of approval. It means little.

The number reflects how much of the tested period was covered by minute data and interpolated ticks. It doesn't measure how realistic those ticks are. The ticks are generated by an algorithm from one-minute bars, with a fixed pattern.

The result is a backtest that looks precise and isn't. A scalping strategy with tight stops and targets depends on exactly the information that the interpolation invents.

A strategy that shows a profit factor of 2.0 at 99 percent quality can show 0.9 on real tick data. The difference comes from the order of highs and lows within each bar and from spread behavior.

There's a related problem with fixed spreads. The tester usually applies a constant spread, either the one at test time or a user-set value. Real spreads vary, widen at rollover, and spike during news.

For any strategy with targets below 5 pips, tests built on real tick data, with variable spreads and added slippage, are the minimum standard.

Did a strategy ever lose its edge when moved from 99 percent modelled data to real ticks?