Every sleeve stands aside when its index falls below a moving average. We tried to beat those windows with everything we could implement — bands, crossovers, cross-market gates, the strategy’s own equity curve, seven families of volatility filter, Barroso & Santa-Clara, Daniel & Moskowitz. Two markets had nothing better. One had a single candidate worth 0.09 of Sharpe, found after 130 attempts. Re-optimising annually lost in all three. Includes two of our own results that turned out to be look-ahead.
The production BTC signal (RSI14 > 54 · price > MA155 · vol < 100%) against everything we could throw at it: 360 parameter variants, 5 timeframes, 66 challenger systems, long/short overlays — all net of live-measured costs (19bps round-trip). It ranked #1 of 360 on both windows. 45% CAGR net since 2018, Sharpe 1.32, MaxDD −32% vs −77% for Buy & Hold. Verdict: do not touch the signal.
The BTC signal applied to 10 other coins, plus 753 coin-specific challengers with a strict train/test split. What travels is the protection, not the alpha: every freshly-fitted champion collapsed 51–110% out of sample while the un-refitted BTC signal held Sharpe 1.37. Every multi-coin blend lost to BTC alone. The book stays BTC-only.
Rebuilt from the series each strategy publishes. At 35/40/25 the book returns +29.5% with Sharpe 1.44 and MaxDD -19.5%, against -26.0% for the best sleeve alone. Correlations 0.26–0.30, and they fall rather than rise in the worst months. 65 of 231 weightings are statistically tied for best — and re-optimising every year loses 0.18 of Sharpe to simply holding a fixed weight.
r=0.35 full-sample, but it swings from −0.73 to +0.85 across rolling 12-month windows. Six crisis periods, a check for cash-timing artifacts, and where 50/50 actually sits versus the efficient frontier (spoiler: not on it).
Five one-at-a-time sweeps behind the TSX momentum strategy — lookback, market cap floor, regime window, portfolio size, fundamental filter — plus a full extreme-return data-quality scan. N=10 isn't a strict winner on every metric, and we say so.
Southeast Asia's dominant superapp at $3.50. First full-year profit in 2025, $5B net cash, 17 consecutive EBITDA growth quarters. The SPAC scar is still suppressing the multiple.
Systematic analysis of the Brazilian real carry trade. Interest rate differential, FX volatility, and historical risk-adjusted returns across rate cycles.
China's largest direct retailer trading at deep discount to intrinsic value. Supply chain advantages, logistics network, and shareholder return analysis.
Deep-dive on Sea Limited (NYSE: SE). Southeast Asia's largest digital economy — gaming, e-commerce, and fintech. Valuation, competitive moat, and 3-year P&L forecast including Brazil expansion.
Value screening (VC1/VC2/VC3) combined with 6-month momentum. Buy the cheapest 20%, pick the 25 with strongest momentum. TV3 (value + dividend yield + momentum) at >$5B monthly delivers +13.8% CAGR, +5.4% alpha, Sharpe 0.82. Momentum reduces drawdowns by 10–20pp vs pure value.
O'Shaughnessy's Value Composite promised 18–21% CAGR. Three versions tested with quarterly TTM fundamentals (2005–2026). With annual rebalance (12 portfolios averaged): VC2 delivers +5.6% to +9.8% depending on market cap. Monthly rebalance improves to +10–12%. Real alpha exists but is a fraction of the book's claims.
P/E, P/S, P/B, P/FCF, EV/EBITDA, Shareholder Yield, and Dividend Yield — each tested independently on quarterly TTM data (2005–2026). P/S is the only factor that works consistently (+6.1% spread). P/B is inverted (−3.8%). Shareholder Yield is strongly inverted (−7.4%). Year-by-year breakdowns, rolling 3- and 5-year windows, regime analysis.
Jegadeesh-Titman 12-1, 6-1, 12-2, and Novy-Marx intermediate tested at >$200M, >$1B, and >$5B with monthly and annual rebalance. Plus value exclusion filters and MA200 overlay. Mom 6-1 >$5B delivers +13.5% CAGR. With exclusion filters and MA200: Sharpe 0.87, max drawdown −28%. Early-stage exploration — not the configuration running live today, see the current USA momentum strategy report below.
Systematic sweep across six parameter dimensions — MA period, concentration, lookback, market cap, skip month, and filter ablation. 6-month pure momentum (6-0) dominates conventional 9-1 across every metric. Not a Mag7 proxy: only 0.6 of 20 positions per month are Magnificent Seven stocks. 8 charts, 16 sections. Early-stage exploration — not the configuration running live today, see the current USA momentum strategy report below.
The MA drives returns. The VIX protects against tail risk. Adding a VIX > 25 & rising exit cuts max drawdown from −31% to −22% at less than 0.5% CAGR cost. 28 variants tested across thresholds, moving averages, percentile ranks, and combined filters.