Article 3 · Signals and contributions

Is there anything better than the 200-day average? Indicators, inverse ETFs, volatility and forums

No alternative signal clearly beat the simple 200-day average, and contributing every month doesn't keep you from losing.

Read the full article on Substack (in Spanish)

An educational tool, not investment advice. Historical figures, computed with the Modeler's engine; past performance does not guarantee future results.

If I put the same amount into a 3x every month, how often do I end up losing?

19.6%of windows, the 3x ends up losing money

2.0%the index's (1x)

18.6%median annual return of the 3x (IRR)

23.7%3 years

17.3%5 years

19.6%10 years

16.8%15 years

15.1%20 years

Years contributing → of windows, the 3x ends up losing money

Annual return on contributions (IRR; p5, p25, p50, p75, p95)

Contributing $1 a month doesn't protect you from a crisis at the end: what counts is where the window ends.

How it's computed

943 windows since Feb 1, 1938, only 7 independent. Data through Aug 31, 2026.

Main figure: matches the article From the article: With monthly contributions for 10 years, the 3x ends up losing money in 19.6% of windows since 1938 (the index, in 2.0%). The Modeler gives the same figure (19.6%).

Open it in the Modeler

Do an exponential average, MACD or Donchian beat the simple average?

17.3%a year with the simple average

15.2%with the exponential

11.1%with MACD

13.9%100

18.1%150

17.3%200

16.8%250

15.5%300

Sessions in the average → a year with the simple average

Annual return since 1938

All four rules hold the 3x or T-bills; the control changes the average of the first two.

How it's computed

Real history from February 1938 to July 2026, in whole months. Data through Jul 31, 2026.

Main figure: differs from the article From the article: Since 1938, the rule with the 200-day exponential average has returned 15.3% a year versus 17.4% for the simple one. The Modeler gives 15.2% with data through Jul 31, 2026 and whole months.

Open it in the Modeler