Trading-strategy family
REGIME_BULL_SURVIVE_V1Bull-market regime filter (survive bears, earn in bulls)FAIL
tested Oct 27, 2026 4 variants tried
What was tested
Long-only, 1×, spot-style: hold BTC only while a long trend filter says the market is rising, otherwise sit in cash. Four filters were fixed in advance: the 200-day average; the same with volatility-targeted size; a 50/200-day average cross with a 20-day-low exit; and the volatility-targeted version with a 25% drawdown brake. ETH was tested as a secondary asset.
The goal behind it: stay alive in bear markets and actually make money when a real bull market comes. Binance daily prices, costs 15 bp per side. IS Mar 5, 2018 to Dec 31, 2022, OOS Jan 1, 2023 to Sep 30, 2026, which contains one big bull and one bear.
Pre-registered pass bar and result
The bar was written and hashed before the run. Results are on data the selection step never saw (out-of-sample).
| Criterion | Bar | Result | Verdict |
|---|---|---|---|
| Fee gate (IS): gross ≥ 2× costs | ≥ 2× | 37.2× | PASS |
| C1 Survive: OOS max drawdown no worse than −35% and shallower than buy-and-hold | ≥ −35%, shallower than −53.0% | −40.7% vs −53.0% | FAIL |
| C2 Bull-market capture | ≥ 50% | 29% | FAIL |
| C3 CAGR ≥ 70% of buy-and-hold, or Calmar above buy-and-hold | ≥ 70% or Calmar > 1.02 | 8.0% vs 54.0%; Calmar 0.20 vs 1.02 | FAIL |
| C4 Calmar above random timing (discounted 98.75th percentile, 1000 seeds) | > 2.73 | 0.20 (38th percentile) | FAIL |
Why it failed
Half of the goal held and half did not. “Survive bears” is true: every filter had a shallower worst drawdown than buy-and-hold (OOS −27% to −41% against −53%), and 1× spot cannot be liquidated. “Earn in bulls” did not hold for the variant the rules picked: it kept being shaken out by false breakouts in a choppy bull market and captured only 29% of the bull-market gains.
For information only, the three variants that were not selected earned +36% to +39% a year out of sample, but none beat random timing by a clear margin (85th to 87th percentile; 95th was needed even for a warning), and picking one now would be choosing with the answer in hand. ETH failed all four criteria too. The family is cut; the survival ideas continue as two forward paper tests, shown on the home page.
Variants tried
4 variants, fixed in advance: (a) 200-day average; (b) 200-day average with volatility-targeted size; (c) 50/200 cross with 20-day-low exit, selected on IS; (d) (b) plus a 25% drawdown brake. OOS CAGR (information; only (c) is judged): (a) 39.3%, (b) 35.5%, (c) 8.0%, (d) 35.5%.
What this test cannot tell
Only about three bull/bear cycles for two correlated assets. Daily bars hide intraday drawdowns. Binance USDT prices stand in for USD. Data ends Sep 30, 2026. Nothing here says anything about the Kraken evaluation account's own drawdown rules.
Rules lock
Rules file SHA-256:
92e019a0aee0cdd8f71635e469f99b87bcc7178e042ca90bc77cb1b6455f8ef0
Locked:
“not recorded” means the number or time was not written down at the time, and we have not filled it in afterwards.
Submit a rule
Have a rule with numbers? Send it to us on X @ProjectAnchorHQ or Telegram t.me/ProjectAnchorHQ (market, entry, exit, stop). Free; we only need the written rule, never keys, accounts or funds. We publish the result pass or fail, and credit you only if you say yes.