Normal
100%Hold the full Dynamic Growth Bucket. Strategic weights drift until the next scheduled rebalance.
03 · Optional drawdown insurance
Stops work.
They are not free.
10.3–11.3 pp
Maximum-drawdown improvement from the existing tested rules—paired with 1.1–2.0 percentage points less CAGR and far more trading.
Plain-English mechanics
Three possible
risk settings.
The portfolio never borrows. When a rule reduces a growth position, the released capital moves to SGOV. Signals are observed first; trades happen later.
Soft
50%Weak growth sleeves are cut in half. The freed allocation moves to SGOV until the lagged trend recovers.
Hard
0%Weak growth sleeves are removed. A conditional volatility brake may also scale the full portfolio toward SGOV.
ETF gates + weekly brake
01At each month-end, compare every active growth ETF with its own 200-day moving average.
02Below trend means 50% exposure in the soft version or 0% in the hard version; the difference moves to SGOV.
03At each week-end, the volatility brake activates only when SPY is below its own 200-day average.
04Every signal uses information through the prior close. The close-t trade first affects the next session.
Slower composite governor
01Observe the whole Growth Bucket and SPY against their own 200-day averages once per month.
02A two-month positive confirmation is required before restoring full exposure, reducing whipsaw.
03The simple soft rule added only eight state-change trade dates over the decade.
04Weekly and monthly portfolio brakes were rejected: trading exploded without improving full-period drawdown beyond the simpler state rule.
Real historical episodes
Press play.
Follow the money.
The chart does not merely highlight a crisis. It shows the actual lagged model state, target exposure, SGOV balance, trade markers and two-leg execution cost.
REAL BACKTEST EPISODE · LAGGED EXECUTION
Watch the rule, trade and consequence
Selected strategy with Risk GovernorSame-holdings static control · quarterly, no stops
These are portfolio-performance lines. The 200-day moving averages are signal inputs and are not drawn on this chart.
Dec 2, 2019NORMAL
Dec 2, 2019
GROWTH EXPOSURE NOW41.7%
SGOV NOW2.5%
LATEST TRADENone
TRADE COST0 bp
Shock-to-trough: this rule avoided 10.5 percentage points of loss versus the same-holdings static control, with 4 trade days and 147% gross bought + sold inside the shock window. Orange markers are real modeled trade dates.
Existing governor · 10 bp per side
A large risk benefit.
A visible return premium.
These are same-holdings comparisons. The only difference is the rule. Costs charge each dollar sold and each dollar purchased.
| Profile / version | CAGR | Volatility | Max drawdown | Gross traded / yr |
|---|---|---|---|---|
| Conservative · Static control | 11.94% | 10.68% | -24.16% | 16.98% |
| Conservative · Soft gate + brake | 10.67% | 8.27% | -12.87% | 167.49% |
| Conservative · Hard gate + brake | 10.74% | 8.12% | -12.87% | 145.90% |
| Moderate · Static control | 13.84% | 12.16% | -24.89% | 20.73% |
| Moderate · Soft gate + brake | 12.34% | 9.47% | -14.19% | 153.09% |
| Moderate · Hard gate + brake | 12.27% | 8.95% | -14.11% | 140.81% |
| Aggressive · Static control | 16.71% | 14.83% | -27.11% | 23.02% |
| Aggressive · Soft gate + brake | 14.92% | 11.71% | -16.69% | 151.06% |
| Aggressive · Hard gate + brake | 14.74% | 10.94% | -16.61% | 162.01% |
How to read gross traded / year: 167% means that a $1 million portfolio executed about $1.67 million of total sell-plus-buy notional in an average year across the entire portfolio. It is not 1.67 times the Growth Bucket and it is not limited to drawdown years.
Full history uses proxy-spliced series where live ETF history is unavailable. Exact stop-rule mechanics reconcile to the legacy Strategy Lab at machine precision before repricing. Taxes, bid/ask spread, market impact and investor-specific constraints are not included.
RESEARCH RECOMMENDATIONKeep Soft as an option. Keep Hard as a stress reference.
The new low-frequency Soft governor is the only overlay worth paper-testing: modest protection, only eight state-change dates and no leverage. The existing hard/weekly system remains valuable to demonstrate maximum protection, but its 141%–188% annual gross trading and rebound lag make it too costly to be the default return strategy.