Walk-forward validated · No curve fitting

ETF rotation that
re-optimizes itself
every two weeks

StratScout detects the market regime, picks the best ETFs for it, and re-tunes its own parameters every 14 days on trailing data, then trades blind. Every result below is fully out-of-sample.

Join the waitlist See the results
$253k
Median final NAV (2 runs)
46%
Median CAGR (OOS)
34%
Median max drawdown
7.3×
vs SPY buy-and-hold
59%
Active trade win rate

How it works

Four steps, fully automated, zero discretion.

Step 1

Detect the regime

Compare AGG vs BIL (90-day return) to determine risk-on or risk-off. Compare TLT vs BIL to split risk-off into rising-rate vs falling-rate. No prediction - just measurement.

Step 2

Optimize parameters

Every 14 days, a random-search optimizer runs 300 trials on the trailing 12 months of data, split across 3 sub-windows for overfit resistance. It finds the ETF mix and sizing that maximizes risk-adjusted returns (Calmar ratio).

Step 3

Deploy blind

The winning parameters trade the next 14-day period with no further adjustment. No peeking at future data. Every period in the results below was traded this way.

Step 4

Repeat

The cycle restarts. Old data rolls off, new data comes in. The optimizer never uses the same period twice as training and validation.


Three regimes

The signal is two ratios. The regime drives the entire ETF selection.

Risk-On

Leveraged equity

When AGG > BIL: rotate into top-ranked leveraged ETFs - SOXL, TQQQ, TECL and peers, weighted by momentum + inverse-vol blend.

Risk-Off · Rising Rates

Cash or inverse

When AGG < BIL and TLT < BIL: hold cash (BIL) or inverse ETFs. Backtests show BIL wins 145 of 218 periods in this regime - cash beats decaying inverse funds.

Risk-Off · Falling Rates

Defensives & managed futures

When AGG < BIL and TLT > BIL: rotate into UGL, GLD, KMLM, DBMF, TAIL, XLP. Near-zero equity correlation, positive in drawdowns.


Results

2018-01-01 → present. $10,000 starting capital. Every period is out-of-sample, meaning parameters were optimized only on data prior to that period. Runs are statistically independent (different random seeds).

9x
vs SPY buy-and-hold (HoF Seeded)
59%
Active trade win rate
14,000+
Optimizer trials per period
220
Out-of-sample periods per run
0
Parameters fit to future data
14 days
Re-optimizes automatically
Configuration Median NAV Median CAGR Median Max DD Median Calmar
Random Universe $114,700 +31.1% 36.5% 0.92
Curated Universe $214,897 +44.4% 37.4% 1.14
Curated + HoF Seeded $314,970+50.5%29.5%1.70
SPY buy-and-hold $34,798+13.1%34%0.39

More runs completing. Table will be updated as r3–r5 finish. Prior 5-run benchmark (pool expansion): mean $231k, 41.2% CAGR, 35.1% MaxDD.


Honest limitations

Things you should know before trusting these numbers.

* One extended run covers 2007-present (494 periods). During the 2008-2009 crisis window, the strategy grew $10k to $29k with a 34.5% max drawdown, as the regime signal rotated out of equities early. That run uses a partially different ETF universe since several current instruments did not exist pre-2010.


Get early access

StratScout is launching as a SaaS - hosted walk-forwards, live signals,
strategy inspector. Join the waitlist for early pricing.

No spam. Unsubscribe any time. Early members get lifetime discount.