Backtest Performance

System performance across real market cycles.

Tested across 41 years of expansions, crashes, and recoveries. All results are hypothetical, derived from historical price data.

See all 18 tests → How the System Works
Worst leveraged-equity window · 2000–2013
UOPIX / USPIX · 2× Nasdaq-100 mutual fund
+31.5%
CAGR
1.71
Sharpe
−17%
Max DD
1.85
Profit Factor
−13.9%
B&H CAGR
−99%
B&H Max DD

While leveraged buy-and-hold experienced drawdowns near −99% during the same period.

Positive in every worst year: 2000 (+27%), 2001 (+11%), 2002 (+9%), 2008 (+1%). $100k starting capital ended at $4.59M on the JEDI path; the same $100k in UOPIX buy-and-hold ended at $12.4k.

Full-cycle baseline · 1985–2025
NDX 3× · 40 years · synthetic pre-2010
Long-term compounding
58.2% CAGR
Strong risk-adjusted return
2.00 Sharpe
Contained downside
−31.8% Max DD
Return per unit of drawdown
1.83 Calmar
More winners than losers
54.7% Win Rate
Gains exceed losses by 2.3×
2.35 Profit Factor
✅
Out-of-sample validation

Results after 2025 were not used in model design — ensuring unbiased evaluation.

$10K → millions across 41 years of market cycles

Compared against a 3× leveraged Nasdaq-100 benchmark

Event windows shaded in red. Log scale recommended over 41 years.

JEDI Portfolio
Nasdaq-100 Buy & Hold
Major market event
Events highlighted on the chart

Drawdown profile (underwater plot)

Max DD: -31.8% · Recovery typically under 12 months

Drawdowns remained contained relative to leveraged benchmarks.

Crisis performance

Outperformance during major market crises

The system remained adaptive across extreme market stress events.

JEDI return vs Nasdaq-100 buy-and-hold across the four largest drawdowns of the past 25 years.

Dot-com Bust
2000–2002 (cumulative)
JEDI +104% Nasdaq-100 −72%
+176pp
Outperformance
Global Financial Crisis
2008 (single year)
JEDI +1.3% Nasdaq-100 −40.9%
+42pp
Outperformance
COVID Crash
2020 (full year)
JEDI +267% Nasdaq-100 +45%
+222pp
Outperformance
Inflation Bear
2022 (single year)
JEDI −11.1% Nasdaq-100 −33.7%
+22.6pp
Outperformance

Across the four largest drawdowns of the past 25 years, the system delivered positive outperformance every time. Across 41 years total, the system's regime classifier identified 800+ state transitions, enabling the adaptive positioning above.

Profitable in 34 of 41 years.
Beat the Nasdaq-100 in 32 of 41 years · View full year-by-year table

Annual Returns · 1985–2025

JEDI system vs Nasdaq-100 index. Pre-2010 uses synthetic 3× daily-rebalanced ETF model.

Alpha measures annual outperformance of JEDI versus the Nasdaq-100.

Download Annual Returns (CSV)
Year JEDI Return NDX Return Alpha (JEDI − NDX) Data Source
2025+107.2%+20.4%+86.8 pp📈 Real
2024+129.6%+27.0%+102.6 pp📈 Real
2023+100.2%+54.9%+45.3 pp📈 Real
2022−11.1%−33.7%+22.6 pp📈 Real
2021+82.2%+28.6%+53.6 pp📈 Real
2020+267.2%+45.3%+221.9 pp📈 Real
2019+130.7%+37.3%+93.4 pp📈 Real
2018+38.2%−2.8%+41.0 pp📈 Real
2017+110.8%+30.2%+80.6 pp📈 Real
2016+6.7%+8.1%−1.4 pp📈 Real
2015−0.9%+8.6%−9.5 pp📈 Real
2014+145.1%+18.9%+126.2 pp📈 Real
2013+131.6%+30.8%+100.8 pp📈 Real
2012+78.7%+14.6%+64.1 pp📈 Real
2011−2.1%+1.0%−3.1 pp📈 Real
2010+152.8%+17.6%+135.2 pp📈 Real
🔬 Pre-2010 Results — Modeled Using Synthetic 3× Nasdaq-100 ETF Data (TQQQ/SQQQ launched in 2010)
2009+64.0%+47.2%+16.8 pp🔬 Synthetic
2008+1.3%−40.9%+42.2 pp🔬 Synthetic
2007+58.0%+18.5%+39.5 pp🔬 Synthetic
2006−0.6%+4.6%−5.2 pp🔬 Synthetic
2005+35.5%+2.6%+32.9 pp🔬 Synthetic
2004+61.5%+10.8%+50.7 pp🔬 Synthetic
2003+63.0%+42.8%+20.2 pp🔬 Synthetic
2002+24.2%−38.9%+63.1 pp🔬 Synthetic
2001+20.1%−25.9%+46.0 pp🔬 Synthetic
2000+36.8%−38.2%+75.0 pp🔬 Synthetic
1999+56.5%+99.9%−43.4 pp🔬 Synthetic
1998+246.3%+82.1%+164.2 pp🔬 Synthetic
1997+86.3%+21.5%+64.8 pp🔬 Synthetic
1996+86.4%+40.2%+46.2 pp🔬 Synthetic
1995+164.5%+44.8%+119.7 pp🔬 Synthetic
1994+15.1%+2.2%+12.9 pp🔬 Synthetic
1993−9.3%+11.8%−21.1 pp🔬 Synthetic
1992+21.7%+8.0%+13.7 pp🔬 Synthetic
1991+136.5%+65.6%+70.9 pp🔬 Synthetic
1990+32.1%−11.9%+44.0 pp🔬 Synthetic
1989+112.8%+28.2%+84.6 pp🔬 Synthetic
1988−5.3%+9.6%−14.9 pp🔬 Synthetic
1987+37.7%+9.4%+28.3 pp🔬 Synthetic
1986+4.7%+7.7%−3.0 pp🔬 Synthetic
1985+0.0%+18.0%−18.0 pp🔬 Synthetic · Warmup
Methodology & data notes
How the backtest was constructed, data sources, and assumptions

1985–2010 uses synthetic TQQQ/SQQQ data modeled as a 3× daily-rebalanced leveraged exposure to the Nasdaq-100 index, including ETF expense ratios and daily compounding effects.

2010–2025 uses actual historical TQQQ/SQQQ prices.

Backtests assume 0.05% slippage per trade and commission-free execution.

Backtests are generated using the same engine that produces live system positioning.

All performance figures are derived from backtested simulations and do not represent actual trading results. Backtests are based on historical data and assumptions that may not reflect real market conditions. Past performance does not guarantee future results.

Validated expressions

Cross-asset results

Same system, validated across markets and leverage levels.

Every row is backed by a documented test in the validation suite.

Test Instrument Period CAGR Sharpe Max DD
12× NDX MF (UOPIX)2000–2013 worst window+31.5%1.71−17%
22× NDX MF (UOPIX)1998–2026 (27y)+38.8%2.01−17%
33× NDX (TQQQ/SQQQ)2010–2026 real-only+77.5%2.56−27.2%
43× SPX (UPRO/SPXU)2009–2026 (17y)+45.5%2.08−22.1%
52× S&P MF (ULPIX)1997–2026 (28y)+24.1%1.52−22%
61× SPX (SPY/SH)2006–2026 (20y)+21.2%2.43−8.7%
7DAX / EWG (Germany)1996–2026 (30y)+13.6%0.98−24.5%
8Nikkei / EWJ (Japan)1996–2026 (30y)+10.5%0.74−24.6%
93× NDX · OOS slice2025–2026 (12 mo)+79% total3.01−9.9%

Non-US Sharpe is lower than US because the parameter set is calibrated on US indicators. Direction is positive on every test — not the “collapses on a different asset” pattern of overfitting.

What Would $10K Have Become?
Run the interactive simulator across any starting year (1985–2025)

What Would $10K Have Become?

Based on 41 years of backtested data (1985–2025). Adapts across bull, bear, and crisis markets.

Explore different market cycles:

$

Access Live System Positioning

See how the system is positioned today — using the same decision framework that drives these results.

Get Started Free

JEDI AI provides algorithmic signals and user-authorized trade instructions. It does not provide investment advice or manage client funds.