A rules-based, data-driven investment model designed to help everyday people navigate volatile markets with less anxiety and greater confidence.
Most long-term investing advice is straightforward: buy broadly diversified assets and hold forever.
In theory, that works. In practice, it is harder than it sounds.
Markets are erratic. Periods of steady growth are routinely interrupted by sharp pullbacks, prolonged downturns, and sudden shocks. Even disciplined investors with long time horizons can find themselves second-guessing their strategy when their savings are on the line.
Since 2000, the S&P 500 has returned about 6.1% per year on average, excluding dividends. Those are real, meaningful returns — and they should not be dismissed.
But they came alongside several severe drawdowns. This is exactly why most professional investors and institutional funds do not rely exclusively on a buy-and-hold strategy. Instead, they shift capital between equities and more defensive assets as conditions change.
The math behind this is simple. A portfolio that falls 50% must double just to break even. Reducing exposure during the worst downturns — even imperfectly — can have an outsized impact on long-term results.
For most individual investors, the hardest part of investing is not knowing what to own. It is knowing when to stay in and when to step back.
The AnyMarket Algorithm was built to answer that question systematically. Rather than reacting to headlines or gut instinct, the model evaluates a set of objective market signals each day and determines whether conditions favor equity exposure or a more defensive posture. The goal is not to predict the future — it is to apply a consistent, rules-based framework that takes emotion out of the equation.
The AnyMarket Algorithm is built on a straightforward premise: capital should be deployed differently depending on what the market is doing.
Each day, the model evaluates a combination of momentum indicators, trend signals, volatility filters, and statistical thresholds. Based on those signals, it rotates between two asset categories:
When conditions look favorable, the model stays in equities. When warning signs emerge, it rotates toward safety. See the Methodology page for a full breakdown of how each signal works.
Following these signals since 2000, the AnyMarket Algorithm has produced an annualized return of 18.39%, more than tripling the S&P 500 over the same period.
A $1,000 investment in the S&P 500 at the start of 2000 would be worth $4,652.83 today. The same $1,000 following the AnyMarket Algorithm would have grown to $81,411.00.
The model’s allocation changes have historically coincided with meaningful turning points in the market.
Many sell signals have occurred when equity markets were extended or beginning to weaken. Many buy signals have occurred just as markets were stabilizing after major declines — frequently within a day or two of the S&P 500’s lowest price in a given pullback.
The chart below shows every trade the model has made since 2000. The pattern speaks for itself.
A buy-and-hold strategy captures the full upside of equity markets — but it also absorbs every downturn in full.
The AnyMarket Algorithm takes a more adaptive approach. Since January 1, 2000, the model has been in equities 74.18% of the time and in safe-haven assets 25.82% of the time.
The timing of those shifts has mattered enormously:
Across the full period, the S&P 500 gained an average of 0.031% per day.
On days when the model was in equities, the S&P 500 rose an average of 0.090% per day.
On days when the model was in safe havens, the S&P 500 fell an average of 0.139% per day.
The model has spent its time in equities when the market was rising and in safety when it was falling — not perfectly, but consistently enough to make a substantial difference over time.
Small differences in annual returns compound into very large differences over time.
Outperforming the market by 12.29% per year may not feel dramatic in any single year. But over decades, the gap widens significantly — especially because the model’s edge is not symmetrical. When it underperforms, it tends to trail by a modest amount. When it outperforms, the margin is often substantial.
Preserving capital during major downturns is the engine behind this. Money that is not lost during a crash does not need to recover — it continues compounding during the next advance.
Major market corrections can permanently impair long-term returns for investors who stay fully invested through them.
The AnyMarket Algorithm has historically done its best work during the market’s worst periods. Across 26 full calendar years, the model finished in the red just once — in 2000, when it still beat the S&P 500 by 6.41%. Every subsequent year has been positive, giving the model 25 consecutive years of gains.
No model can guarantee this continues. But the track record illustrates what consistent downside protection looks like when compounded over time.
Since January 1, 2000, the AnyMarket Algorithm has:
The asymmetry matters: when the model outperforms, it tends to do so by a wide margin. When it trails, the gap is typically small.
The chart below lets you explore any individual year in detail. The model’s performance during major downturns — the Dot-Com Crash (2000–2002), the Great Recession (2008), and the COVID-19 shock (2020) — is especially worth examining.
The AnyMarket Algorithm started as a personal tool.
Like a lot of investors, I wanted to participate in long-term market growth without spending every day anxious about my portfolio. I also noticed something that bothered me: the same financial professionals who advised ordinary people to buy and hold through everything were quietly managing their own money with far more sophistication.
I wanted a model that could do what those professionals do — evaluate market conditions daily, apply a consistent process, and reach decisions without emotion or headlines getting in the way. Something that could run in the background and let me get on with my life.
Early versions were modest experiments. My bar was low: beat the market by a percentage point or two per year, reduce volatility, and make investing less stressful. That alone would have been a success.
After many months of refinement — adjusting parameters, adding filters, improving signals — the results improved well beyond what I expected. The gap between the algorithm and a simple buy-and-hold strategy kept widening with each iteration.
Eventually the results were too compelling not to share. If a rules-based model could do this for me, it could do it for others.
This site exists to share that work openly and without pretense — not as financial advice, not as a product, and not as a promise. Simply as a data-driven framework that any investor can observe, evaluate, and use to inform their own thinking.
Data is delayed by 30 days. All performance figures on this site reflect data through 30 days ago. This allows time to verify accuracy before publishing. Subscribers receive real-time allocation alerts the day a signal fires. Visit the Support page to learn more.
Dividends are excluded. For simplicity, performance calculations do not include dividends. This means reported returns slightly understate the true historical performance of both the S&P 500 and the AnyMarket Algorithm.
Safe-haven returns use a 4% assumption. When the model is in defensive assets, returns are modeled at a flat 4% annual rate. In practice, short-term Treasuries and similar instruments have varied meaningfully around this figure, but 4% serves as a reasonable long-run approximation.
Past performance is not indicative of future results. See the full disclaimer for more detail.
