Triathlon All-Terrain Strategy
Backtested Performance
In the historical backtest, the All-Terrain Strategy produced 3.9× the annualized return of the S&P 500, with slightly lower overall volatility and less than one-third of the maximum drawdown. Its Sortino ratio was 7.3× SPY’s, reflecting substantially more return per unit of historical downside risk.
| Annualized return | 1 yr | 3 yr | 5 yr | 10 yr | Since 1999 | Volatility | Worst decline | Sortino |
|---|---|---|---|---|---|---|---|---|
| All-Terrain Strategy | 20.2% | 47.6% | 32.6% | 41.4% | 33.5% | 18.4% | −16.9% | 4.74 |
| S&P 500 (SPY) | 15.8% | 22.9% | 13.8% | 15.3% | 8.5% | 19.2% | −55.2% | 0.65 |
| Nasdaq-100 (QQQ) | 23.9% | 28.2% | 16.4% | 21.0% | 10.9% | 26.9% | −83.0% | 0.58 |
Annualized total returns from March 12, 1999 through September 30, 2026, dividends reinvested. The All-Terrain Strategy is net of a 1.50% annual advisory fee; benchmarks are gross of fees, for comparison only. Worst decline is the largest peak-to-trough loss over the period, measured daily. Nasdaq-100 figures are shown for context and are not the strategy’s benchmark.
lower volatilityUptrend · higher participation →
increased volatility
How the strategy performed when conditions changed
Worst peak-to-trough decline for each side during the downturn, measured on daily closing values. Each side reached its own low on its own date. See disclosures at the bottom of the page.
Select a period to see what the market did, how the strategy performed, and what environment it was in.
The model reacts. It does not predict.
The strategy does not forecast the market. It waits for market behavior to confirm that conditions have changed. Some allocation changes follow a single condition; others require several conditions to occur together before anything moves.
Once a move is confirmed, the readings below determine which allocation is held within the new environment, and how quickly it may change again.
A rules-based process can still be wrong.
The strategy was built with the benefit of backtesting. The backtest depends on historical relationships that may not persist in the future.
Designed primarily for tax-deferred retirement accounts.
IRAs and 401(k)s, including former-employer plans.
The strategy reallocates about fourteen times in a typical year and is intended to replace part or all of an existing long-term equity allocation in a tax-deferred account. It may not be suitable for taxable accounts.
Portfolio structure
ETF-based · no short positions · no options · leveraged ETF exposure in certain allocations.
Want the full analysis?
The complete presentation includes model mechanics, leverage analysis, historical decline studies, attribution and detailed limitations.
All performance shown is hypothetical and back-tested. It was generated by applying a proprietary, rules-based model retroactively to daily closing prices from March 12, 1999 through September 30, 2026, a period during which the strategy was not in use. The strategy has not been managed in any actual client account during the period shown, and no actual performance results exist. Because the model’s rules and parameters were developed using historical data, the results benefit from hindsight and may reflect choices that improved historical results; results achieved with hindsight may differ materially from those that would have been achieved prospectively. No representation is made that any account will or is likely to achieve results similar to those shown. Strategy returns are net of a 1.50% annual advisory fee, deducted quarterly in advance, and reflect the reinvestment of dividends; the actual fee charged to a client may differ. Benchmarks are shown gross of fees, for comparison only, with dividends reinvested. Results are calculated on a closing-price basis and assume transactions at the closing price on the day a signal is registered; actual execution prices will differ. Results do not reflect brokerage commissions, transaction costs, bid-ask spreads, slippage or taxes, the deduction of which would reduce the returns shown. Results assume a single initial investment with no contributions or withdrawals. The model holds eleven positions, and because four of them were switched to equivalent Vanguard funds in 2004, fifteen distinct funds appear across the period. Several did not trade for the entire period; their earlier history was reconstructed from the underlying index or instrument each tracks, with the relevant fund-level costs applied. These substitutions are estimates and introduce model risk; actual returns over those periods may have differed. Returns are annualized (compound) unless stated otherwise. Worst decline is the largest peak-to-trough loss measured on daily closing values; a 16.9% decline is itself a substantial loss, and the strategy can decline more than it has historically. In its Growth and Transitional allocations the strategy holds leveraged exchange-traded funds, which magnify both gains and losses and carry additional cost and risk. It holds only one to five securities at a time, and a single holding can be 100% of the balance, and it has exceeded 60% on 25% of all days; that concentration is a risk in itself. Trailing trend signals can produce false signals and whipsaws, and the strategy may lag the benchmark for extended periods, as the backtest shows. The four allocations, their names and the descriptions shown here describe how the strategy is structured; they are not a recommendation and not a complete description of the holdings, which are proprietary. Descriptions of how the strategy behaves in rising and falling markets reflect its design and objectives, not a guarantee; it will not participate fully in every gain or avoid every loss. Allocation-time figures are the share of trading days spent in each allocation. Risk contribution, described here as a share of “total risk,” is an allocation’s share of total variance, computed as the sum of squared daily deviations from the mean during that allocation divided by the sum across all days; variance is additive, so the figures sum to 100%. The strategy is intended for tax-deferred retirement accounts, such as IRAs and 401(k) plans, where its allocation changes do not create current taxable events; it may not be suitable for taxable accounts. Triathlon Partners LLC is a registered investment adviser located in Connecticut and may only transact business in states in which it is registered or qualifies for an exemption; registration does not imply a certain level of skill or training. Information presented is for educational purposes only and is not an offer or solicitation to buy or sell any security or strategy. Consult a qualified financial adviser and tax professional before implementing any strategy discussed here. Past performance, whether actual or hypothetical, is not indicative of future results. All investing involves the risk of loss, including loss of principal. Additional information about the criteria and assumptions underlying these hypothetical results is available on request.