MDAX Trading System Collapses: From +14% Gain to -60% Drawdown

From +14% to 60% Drawdown: The MDAX System That Drowned

A trading system presented with a 40% win rate, low drawdown, and promises of safe leverage was ultimately consumed by its own statistics. Within months of live trading on the German MDAX index, the equity curve rose from €5,000 to a peak of +14.60%, before plummeting into a 60% drawdown that its creator deemed irrecoverable. There was no hacking or false promise: just a daily published trade log, including commissions, spread, and financing costs. This is precisely what makes it an uncomfortable case study for anyone considering buying an automated system.

How the System Worked: Stop-Loss Entries and Trailing Stops

The mechanics were almost insultingly simple: act on daily MDAX candles, stay always in the market—long or short—and flip positions when the stop-loss triggered. A trailing stop trinc the trend, orders were published the night before, and executed the next day. The vehicle used was CFDs, though the author noted they could be replicated with ETFs or futures.

The detail not to overlook is leverage. With starting capital of €5,000, a 120-point loss on the index equaled 2.4% of capital. In other words, minor index movements translated into double-digit percentage losses within the account. The declared spread was 15 points, and financing costs were 3%.

The backtesting statistics sounded reasonable: low drawdown, acceptable profit factor, good recovery factor, and a hit rate around 40%. No get-rich-quick promises or impossible curves. On paper, a decent system.

Why Did the Curve Go From €5,000 to Peaks and Then Into the Hole?

The equity table tells the whole story. On February 9, 2015, the account started at €5,000. The next day, €4,884. By mid-month, €5,141. On March 3, €5,475. On March 5, €5,519. In ten days of trading, the system had added €418 to the initial capital. On March 16, new curve highs were marked with a cumulative +14.60%, which the author attributed to the strength of the German index.

Then came the hangover. On March 30, the account dropped to €4,770, a -4.60%. It rebounded to €5,480 in April. In May, the system was quarantined with a 40% drawdown; orders continued to be published but were no longer executed. The final picture, unambiguous, was a 60% drawdown and the acknowledgment that the system had stopped working.

Some argue every system has a shelf life and recognizing its death is part of the job. Counterweighing this is a colder argument: if the strategy began losing immediately after leaving the design period, perhaps it never worked for the reasons its author believed.

Overoptimization or Just Bad Luck?

Here lies the core issue. The most repeated accusation was curve-fitting: tuning the system to a specific historical segment and then launching it blindly. The author himself admitted the danger of overoptimization, though he defended that his system wasn't optimized and suggested the likely explanation was that it had exhausted its useful life. Both interpretations are compatible with the data.

The discussion shifted to something more practical: did the system fail because the market changed, or because it never had a real edge? Without a long sample of live trades, the question remains unanswered. And that uncertainty is exactly what devours those who entrust their money to a backtesting table.

Trading in Reverse: The Herd Theory That Beats Systems

The most common idea was simple: if the system loses, just do the opposite, down to the last detail. The underlying thesis is well-known—majority of accounts lose, brokers might benefit from contrarian strategies—but it has an obvious problem: it doesn't always work, because the enemy isn't direction but cost.

The clearest explanation appearing in the conversation notes that some systems lose in both directions, often due to commissions or the product itself. Options were cited as an example, where the same strategy loses with low volatility and wins with high volatility, and casino roulette, where the house's mathematical advantage ensures any system loses regardless of how you spin it.

With this in mind, the experiment of trading inverted isn't as free as it sounds: spread and financing are paid equally, whether you win or lose.

---

The episode's balance sheet isn't one of fraud, but of an honest system that ran out of edge. What remains are the figures, the complete trade log, and a cheap lesson: good backtesting says very little about what a system will do in the next streak. If the author publishes such a curve again, the useful question won't be how much it earns in the simulator, but how long it takes to stop working when the market changes mood.