How Strategies Held Up in 2008, 2020 and 2022

A trading strategy stress test replays a strategy through specific market crashes to see how far it fell and how long it took to recover, compared with simply holding the market. In historical backtests of our five strategies, most fell less than their benchmark in 2008 and 2022, but results in the 2020 Covid crash were mixed, and two strategies that use leveraged funds fell far more than the market in simulated runs through the 2000-2002 dot-com crash. These are simulated results on past prices; nobody earned them, and the next crash will not look like any of these.

What a stress test is for

A full-period backtest blends good years and bad years into one yearly figure. A stress test pulls out the bad stretches and looks at them on their own. It answers three questions:

1. How deep did it fall compared with the benchmark over the same crash?
2. How long did it take to get back to its value before the crash?
3. Did the strategy's rules do what they claim in a crisis, or did it just get lucky with timing?

The crashes chosen here are well known. The 2007-2009 financial crisis was a deep, slow bear market driven by the banking system. The 2020 Covid crash was extremely fast, with stock markets falling sharply in a few weeks and recovering within months. The 2022 rate-hike bear market was a grinding decline in both stocks and bonds as interest rates rose. The 2000-2002 dot-com crash was a long collapse concentrated in technology stocks.

The results in one table

All figures below are from historical backtests: computer tests on past prices. Nobody earned these returns, and future results will differ. "Back in" means months until the account regained its pre-crash value.

*Simulated three-fund version, because the leveraged funds these strategies use did not exist yet.

Strategy by strategy

Gold Breakout

Gold Breakout trades gold (XAUUSD) when a quiet period ends in a breakout, and can buy or sell short. Its caveat: doubling trading costs cuts its yearly result roughly in half.

Its benchmark is gold held, not the stock market. In all three crashes it fell less than gold held: 13% against 29% in 2007-2009, 7% against 12% in 2020, and 18% against 21% in 2022. Recovery was mixed. It took longer than gold to recover in the financial crisis (21 months against 14) and in 2020 (3 months against 2), and slightly less time in 2022 (9 against 11).

Importantly, its worst fall did not come during any of these crashes. Its deepest backtest fall was 41%, from May 2018 to October 2019, a quiet period for gold rather than a crisis, and it took 41 months after the bottom to recover. Over the whole backtest from 2004 to 2026 it returned +15.1% a year with a -41% worst fall, against +11.0% a year and -45% for gold held. Stress tests that only look at famous crashes can miss a strategy's real weak spot.

Leveraged Growth

Leveraged Growth rotates between leveraged growth funds, volatility funds and bonds. Its caveat: its rules may be fitted to the years after 2010, and leveraged funds can lose value quickly.

In 2020 it fell 53%, much more than the S&P 500's 34%, and recovered in 4 months, the same as the index. In 2022 it fell 29% against the index's 24%, but was back in 2 months against 23. In the simulated 2007-2009 version, it fell 48% against 55% and recovered in 19 months against 53.

The simulated dot-com result is the warning: a 91% fall against 47% for the S&P 500, with 115 months to recover against 78. A strategy built on 3x leveraged technology funds is exposed to exactly the kind of long tech decline that happened in 2000-2002. Over its main backtest from 2011 to 2026, it returned +129.1% a year with a -53% worst fall, against +13.9% and -34% for the S&P 500. See leveraged ETF decay for why long declines hurt leveraged funds so badly.

Market Switch

Market Switch holds one fund at a time: SPY normally, UVXY when markets look overheated, and leveraged funds after sharp falls. Its caveat: nearly half of its tested profit came from UVXY, and it failed outside the US. UVXY, a volatility fund, has lost most of its value over long periods, so the strategy only holds it for short spells.

Its recent crashes look reasonable: a 42% fall in 2020 against 34% for the index, recovered in 2 months against 4, and an 18% fall in 2022 against 24%, recovered in 1 month against 23. The simulated older crashes are very different: a 96% fall in 2007-2009 against 55%, taking 73 months to recover against 53; and a 100% fall in 2000-2002, meaning the simulated account was wiped out, against 47% for the S&P 500; it took 259 months to get back to its pre-crash value against 78 for the index. Over its main backtest from November 2011 to November 2026 it returned +71.9% a year with a -42% worst fall, against +15.0% and -34% for the S&P 500.

Big Tech Trend

Big Tech Trend decides each month whether the market is healthy, holds strong technology names when it is, and moves toward short-term bonds, gold and defensive sectors when it is not. Its caveat: the stock list was chosen with hindsight, and three stocks made 81% of the profit.

It fell less than the S&P 500 in every crash tested: 41% against 47% in the dot-com crash, 22% against 55% in 2007-2009, 26% against 34% in 2020, and 13% against 24% in 2022. Recovery was faster each time, for example 37 months against 78 after the dot-com crash and 21 against 53 after 2007-2009. Over 2011-2026 it returned +37.3% a year with a -29% worst fall, against +14.0% and -34% for the S&P 500. Its deepest fall in the shown period started in June 2026 and had not recovered at the end of the data, a reminder that a strategy that held up in old crashes can still struggle now.

Steady Growth

Steady Growth holds seven fixed names, NVDA, GLD, COST, LLY, MCK, FIX and AXON, and reweights them weekly. Its caveat: these seven names were picked because they already did well, which flatters the test.

It fell less than the S&P 500 in each crash shown: 39% against 55% in 2007-2009, 20% against 34% in 2020, and 10% against 24% in 2022. Its 2007-2009 recovery still took 41 months. Over 2011-2026 it returned +26.5% a year with a -20% worst fall, against +14.0% and -34% for the S&P 500. Because the names were chosen knowing how they did afterwards, its crash results deserve extra scepticism; read hindsight and survivorship bias for why.

What these stress tests actually show

A few honest conclusions:

• Market filters helped in slow bear markets. Strategies that step aside when the market looks weak tended to fall less in 2007-2009 and 2022, which were drawn-out declines.
• Fast crashes are hard for everyone. In 2020 the market fell so quickly that monthly and weekly rules had little time to react. Two strategies fell more than the index.
• Leverage turns crashes into wipe-outs. The simulated dot-com results for the two leveraged strategies show what happens when 3x funds meet a long decline.
• The worst fall may not be in a famous crash. Gold Breakout's deepest drawdown came during a calm period.
• Recovery time varies more than depth. Compare 1 month and 73 months for the same strategy in different crises.

None of this predicts the next crash. Each crisis has its own shape, and rules that handled one can fail in another. For a broader look, see what rule-based strategies do in bear markets.

How to run your own stress test on any strategy

If you are judging a strategy from anyone, ask for:

1. Its fall and recovery time in at least two or three named crises, next to the benchmark.
2. Whether those periods were inside or outside the data used to design the rules.
3. Whether any results are simulated using substitute funds, and how.
4. Its deepest fall over the full test, whenever it happened.

If a provider shows only the good crashes, or shows none, that tells you something.

Key takeaways

• A stress test isolates crashes to show depth of fall and time to recover against a benchmark.
• In these backtests, most strategies fell less than their benchmark in 2007-2009 and 2022, but results in the fast 2020 crash were mixed.
• Simulated dot-com results show leveraged strategies can lose nearly everything in a long tech decline.
• A strategy's worst drawdown can happen outside any famous crash.
• All figures are simulated; the next crash will differ.

This article is general education, not personal financial advice.

Watch them in today's market

Past crashes are only half the story. You can run all five AlgolaneBots strategies on a $10,000 paper account for a free 90-day trial, with real prices and costs, and see how they handle whatever the market does next. Start your free trial and review the risk disclosures.