Index EA vs. Corona Shock: Did It Survive Intraday Crashes? The Ultimate Test!

A beginner-friendly summary of the verification: “Index EA vs. Corona Shock: Did It Survive Intraday Crashes? The Ultimate Test!”.

What’s the idea?

We’re always looking for robust algorithmic trading strategies (EAs) that can handle whatever the market throws at them. This time, we focused on a specific “sleeve” of our portfolio: trading major stock indices like the S&P 500 (USA500), Nasdaq 100 (USATECH), and Dow Jones Industrial Average (USA30). The core idea is simple: it’s a long-only strategy. This means we only ever buy, betting that these indices will generally trend upwards over time. We use an exponential trend indicator, like a 200-period Simple Moving Average (SMA200), on daily charts to identify when to be in the market. If the price is above the SMA200, we’re looking to buy; if it falls below, we exit. The big question, though, is how such a strategy would perform during sudden, violent market crashes. While daily charts might look fine, what happens within those crash days? Could we get wiped out by a massive intra-day drop, even if the daily signal eventually gets us out? That’s what this research aimed to find out.

How I tested it

Testing intra-day risk for a strategy that primarily operates on daily signals is tricky. Why? Because 1-minute (M1) data for many years and many instruments is absolutely massive! Downloading and processing it all is a huge undertaking. So, we came up with a smart, efficient way to stress-test:

Daily Strategy Run: First, we ran our index strategy on standard daily (D1) data, using historical cash prices (like from Yahoo Finance). This gave us all the trades the strategy would have taken and, crucially, identified the days when we were holding a position.
Identify Stress Days: From those holding days, we pinpointed the top 25 days where the market experienced the biggest intra-day downward movements. Think of these as our “stress test” days – the moments when things got really hairy.
Targeted High-Resolution Data: For only those 25 specific days, we then downloaded ultra-high-resolution 1-minute (M1) tick data from Dukascopy. We even built a custom tool to do this efficiently, converting their raw tick data into standard 1-minute OHLC (Open, High, Low, Close) bars. This included data from 2019 right up to the present, crucially encompassing the entire COVID-19 market turmoil.
Intra-Day Reconstruction: With this M1 data, we could then precisely reconstruct what would have happened to our positions within those stressful days. Why this method works: Our strategy is a “swing” strategy, meaning we hold positions for several days or weeks. The intra-day risk profile for such a trade doesn’t depend on the exact entry price you got on that specific day, but rather that you were in a trade. So, using daily data for strategy signals and M1 data for stress-testing intra-day drops is a perfectly valid and efficient approach. We applied this rigorous testing to the USA500, USATECH, and USA30 indices, covering a period from 2019 to 2025 (projected, based on available data).

What happened?

First, let’s look at the strategy’s overall performance on the daily timeframe, which we re-confirmed:

Return: A solid +17.7%.
Maximum Drawdown (DD): -6.5%. Drawdown is the largest peak-to-trough decline in your equity curve. A lower DD means smoother performance.
Profit Factor (PF): An excellent 3.92. Profit Factor is your gross profit divided by your gross loss. A PF greater than 1 means you’re profitable; a PF of 3.92 is outstanding, meaning for every $1 lost, the strategy made $3.92 in profit! This daily performance alone is fantastic, but the real test was the intra-day stress analysis. This is where the magic happened:
Maximum Daily Loss: Across all 25 top stress days, including the intense COVID-19 period, the worst single-day loss for our strategy was a mere 1.82%. This occurred on February 25, 2020, right during the first wave of COVID-related selling, when we were holding positions in all three indices.
No Extreme Losses: Crucially, not a single day saw a daily loss of -5% or even -10%!
The Safety Net: How did it manage to avoid the worst of the crashes, especially during events like the “limit down” days in March 2020 (March 12/16)? The key lies in the “long-only + SMA200” rule. The strategy’s logic dictated that if the price fell below the 200-period Simple Moving Average, we would exit our positions. This acted as an early warning system, causing the strategy to be flat (out of trades) before the most catastrophic drops occurred. It was effectively out of the market during the worst of the COVID-19 crash!
Even other significant market events, like the hypothetical “Yen carry trade unwind” in August 2024, showed contained losses of less than 1.3%.

What I learned

This research is a huge win for verifying our strategies! By combining these findings (2019-2025, covering COVID-19) with our previous real M1 data verification (Research 68, covering 2015-2018, which included events like August 2015, Brexit, and February 2018), we can now confidently conclude: The intra-day risk of this index sleeve strategy is safe across all major market crashes. This is a monumental finding because it eliminates a significant “blocker” for another promising strategy we’ve been working on (Research 59). Research 59, which combines a “core” strategy with this index sleeve, showed great potential: monthly returns increased from 0.47% to 0.61%, maximum drawdown was -10.2%, and Profit Factor was 1.41. However, the unknown intra-day risk of the index component was the only thing preventing its full “promotion.” Now, with this verification, we’ve found the first confirmed path to consistently exceed the typical ~0.5% per month ceiling we’ve observed in some strategies, opening up exciting new possibilities for portfolio construction! What’s next for live implementation? While the core verification is complete, there are a few important steps before we roll this out live:

Data Alignment: For live trading, we’ll ensure consistency by using Dukascopy CFD daily data (which is resampled from M1 data) for the index sleeve. Our current validation using Yahoo cash daily data for entries and Dukascopy M1 for intra-day risk is sound because the intra-day risk is independent of the exact entry price.
Broker Specifics: We need to confirm the exact specifications and costs of index CFDs with our target brokers (e.g., Fintokei) to ensure real-world performance matches our backtests.
Risk Management Refinement: While the -10.2% maximum drawdown in Research 59 is acceptable, it’s a bit on the edge. We’ll implement a slightly more conservative risk setting for the index component (adjusting its risk from 0.003 to 0.004) to provide an even greater safety margin. (Good news: lower-risk versions of Research 59 already exist, making this adjustment straightforward!)

How this connects

This verification builds on earlier ones (what failed before and what I tried this time, comparisons between approaches).