💼 My Portfolios

📚 What the past tests showed

15.1%

Annual return

S&P 500 6.2%

-25%

Max drawdown

S&P 500 -53.6%

$161,420

$10,000 became

S&P 500 $32,670

12/20

Years beating the index

17%

2001-2012

S&P 500 1.9%

12.1%

2013-2020

S&P 500 13.1%

The highest return of every formula tested over 20 years: 15.1%/yr vs 6.2% for the index.
Most of the edge came from protection in downturns: in 2008 the index fell 35% while the portfolio was almost flat (-0.1%), because the market filter moved it to cash. It barely lost in 2001-2002 either.
Over 2013-2020 (the blind test, data not used to build the formula) it made 12.1%/yr vs 13.1% for the index, roughly index-like.
Weakness: when the market turns fast, the market filter steps out at the wrong time. The portfolio lost 8.6% in 2016 and made only 6.9% in 2019, years in which the index rose 14% and 31%.
Beat the index in 12 of 20 years. With high costs (0.3% per trade) the return drops to 13.5%/yr.

✓ beat the index that year

How it was tested

Data: ~9,000 US stocks from 2001 to 2020, including delisted ones, so the test isn't limited to the "survivors".
Universe: the 500 most-traded stocks each month (similar to the S&P 500), using only what was known at the time.
Buy the day after the signal, 0.1% cost per trade, rebalanced monthly.
The formula was built on 2001-2012; 2013-2020 was held back as a blind test.

Hypothetical portfolios only, not investment advice. Past performance does not guarantee future results.