Backtest: Strategy vs BTC Buy & Hold
Veem v4 Strategy
BTC Buy & Hold
Returns are gross of fees and taxes.
Ulysses contract
IN ENVELOPEA backtest says what a strategy did on past data. The real question, once actual money is running, is whether it still behaves the way that backtest said it would. This is what the contract watches.
It is a written pre-commitment, signed before the money went in, named after Ulysses tying himself to the mast: the decision to hold is made while the sea is calm. Lagging an index, a losing year, a drawdown inside the known range are not reasons to stop, they are the price of admission. Only the four limits below can open a review, and a review is never an automatic exit.
| Measure | Live | Opens a review |
|---|---|---|
| Drawdown from peak | 0.0% | -55% |
| 2026 so far | +1.1% | -20% at close |
| Losing years in a row | 0 | 3 |
| Trades, trailing 12 months | 0 | > 15 |
Live figures are time-weighted, so deposits and withdrawals cannot fake a gain or a drawdown. The drawdown limit sits at 1.5× the worst the backtest ever showed (-41.4%). Running live since 2026-03-16, measured to 2026-07-24.
Yearly Returns
| Year | Strategy | BTC B&H |
|---|---|---|
| 2015 | +23.5% | +47.6% |
| 2016 | +27.1% | +112% |
| 2017 | +604.6% | +1479.6% |
| 2018 | -2.7% | -72.7% |
| 2019 | +29.2% | +92% |
| 2020 | +143.2% | +253.9% |
| 2021 | +129.4% | +93.4% |
| 2022 | +0% | -66.9% |
| 2023 | +25.1% | +151% |
| 2024 | +82.5% | +121.3% |
| 2025 | -7.8% | -6.1% |
| 2026 | +0% | -26.4% |
Recent Trades
How it works
Weekly trend following signal on BTC. Three technical filters — measuring momentum, trend direction, and market structure — must all be bullish simultaneously to hold BTC. If any one fails, the position moves to cash.
The AND logic is deliberately conservative. Each filter catches a different aspect of the market. When all three agree, the probability of a real trend is high. When they disagree, something is ambiguous — and in BTC, ambiguity usually resolves to the downside.
Why these filters
The indicator selection comes from an exhaustive search across 1.2M+ indicator combinations on BTC weekly data. The goal was not the highest backtest return (that's overfitting) — it was stability: combinations where changing the parameters within a reasonable range doesn't break the strategy.
All three indicators use their original, published default parameters (1968, 1979, 1994) — zero optimization. Since the stability analysis showed the strategy works across a wide range of parameters, there's no reason to deviate from the defaults.
The trade-off
The strategy spends only ~35% of weeks in BTC. You miss some upside from late entries and early exits, but you avoid the -70% to -80% drawdowns that BTC does regularly. The result: roughly the same CAGR as buy & hold, but with a max drawdown of -35% instead of -83%.
Stress testing
Validated across 70 parameter configurations with a CV of Sharpe = 6% — very robust. The MACD parameters are essentially irrelevant (CV = 1%). The trend filter is the most sensitive but still stable (CV = 4%). Combined parameter shifts in either direction still produce a Sharpe above 1.1.
Crisis performance
How these numbers are computed
Curves and metrics are gross: no spread, commission, slippage or tax is deducted. Trading fees plus spread run roughly 25 to 50 bps per flip, a drag of about 2 to 4 CAGR points.
Sharpe here is (CAGR − 4%) / volatility: a compounded return, minus a fixed 4% risk-free rate, per unit of volatility. Published Sharpe figures usually take the arithmetic average return and the risk-free rate of the day, so the two are not directly comparable.
Win Rate is the share of positive weeks, not the share of winning trades. Trend following wins less often than that, and makes it up on the size of the winners.