Alternative filter

On 25 August and from 28 August to 2 September 2026 the alt strategies' filter flipped OFF while Bitcoin was in a bull regime, the alt index CCI was strongly positive and the index sat above every moving average. The KAMA leg tripped: a single down candle crosses it. We measured alternatives and one stood out: CCI(40) > 0 and price above the SMA50 of the alt index. It ran on this page from 4 September.

Since 16 September 2026 that rule is the live alt filter. Through September the KAMA rule kept switching the alt strategies on and off almost every other day while the index moved sideways, and each switch was a round of entry or exit emails. Bitcoin's leg did not change: it is still CCI(120) > 0. The old KAMA rule stays published here as the previous rule, so the two can still be compared day by day.

Since 6 September the Lab rule is the same rule on both legs — Bitcoin: CCI(120) > 0 and price above the SMA50; alt index: CCI(40) > 0 and price above the SMA50. Until now only the alt leg changed and Bitcoin kept its bare CCI(120). The reason for adding the SMA50 to Bitcoin as well is drawdown: inside a bull regime the SMA50 leaves earlier when price rolls over. Measured on BTC Active from 2014, the deepest fall in dollars goes from −63 % to −48 %, and the price of that is CAGR dropping from 65 % to 58 %. It is a worse compounding rate for a materially shallower hole — better Calmar in dollars from 2014, 2018 and 2022 alike. It costs return; it is not free.

On the window published below, it does not hold. These curves start in July 2018, are charged 0.30 % a leg and measure drawdown as a percentage of equity rather than in dollars — and on that basis the both-legs rule makes BTC Active plainly worse: 58.9 % CAGR falls to 51.6 %, the deepest fall goes the wrong way too (−34.1 % to −38.3 %), and Calmar drops from 1.73 to 1.35. The 2014 measurement and this one disagree, and the disagreement is the point of publishing both: read the table, not the argument.

The chart below switches between the previous rule and the both-legs rule, and the metrics table shows current, previous and both-legs side by side. The both-legs rule does not drive the main page, alerts or emails. Review date: 1 December 2026. Measurement details available on request.