Shock Engine · Performance
Performance dashboard
Bitcoin and Ethereum, half of the capital each, 2018-09-01 → 2026-09-30. Every figure on this page comes from the published research files and can be downloaded below.
Historical simulation after modeled transaction costs. Not live performance.
Equity
Growth of 100
Each market starts with half of the capital and compounds only its own half; the portfolio is the sum of the two, without rebalancing. Commission of 0.045 % on every order.
Equity, indexed to 100
Log scale · daily closes
Dashed line: buy & hold BTC/ETH 50/50 over the same period (half in each coin on 2018-09-01, no rebalancing, no costs): 100 → 1,073, CAGR 34.1%, max drawdown −76.9%, Sharpe 0.78, Sortino 1.11.
Metrics
Key figures
Daily UTC close-to-close returns, annualized with √365.25, risk-free rate 0. The Sortino ratio counts only losing days as risk: mean / downside deviation of daily returns × √365.25 (target 0: squared negative returns averaged over all days). BTC, ETH and the portfolio cover exactly the same 2,952 days.
| BTC | ETH | Portfolio | |
|---|---|---|---|
| Total return | 1381% | 11490% | 6436% |
| CAGR | 39.6% | 80.0% | 67.7% |
| Annualized volatility | 25.0% | 33.3% | 27.4% |
| Sharpe ratio | 1.46 | 1.93 | 2.02 |
| Sortino ratio | 3.23 | 4.62 | 4.73 |
| Max drawdown (daily closes) | −24.8% | −24.1% | −17.8% |
| Max drawdown (15-min marks) | −26.5% | −26.7% | −19.6% |
| Calmar ratio | 1.60 | 3.31 | 3.79 |
| Skewness of daily returns | 4.25 | 4.09 | 3.65 |
| Profit factor (trades) | 1.93 | 2.73 | 2.59 |
| Time in market | 15% | 17% | 25% |
| Transaction costs per year | 4.0% | 3.9% | 3.9% |
| Trades | 358 | 358 | 716 |
| Winning trades | 30% | 33% | 31% |
| Payoff ratio (avg win / avg loss) | 5.18 | 5.52 | 5.58 |
| Best month | 30.3% (Jun 2022) | 62.1% (Nov 2018) | 45.8% (Nov 2018) |
| Worst month | −8.8% (Feb 2022) | −15.2% (Feb 2022) | −13.9% (Feb 2022) |
| Worst year | +1.2% (2025) | −9.0% (2019) | +8.0% (2019) |
When gains and losses are spread alike, the Sortino is about 1.4 times the Sharpe: 1.43 times for buy & hold BTC/ETH 50/50 (Sortino 1.11). For the portfolio it is 2.34 times: most of its day-to-day volatility comes from gains, not losses.
Portfolio trades are the 716 trades of both markets pooled; the payoff ratio uses their contribution to portfolio equity. The first and last years are partial (2018-09-01 → 2018-12-31 and 2026-01-01 → 2026-09-30).
Risk
Drawdowns
Depth, duration and recovery of every fall from a previous peak, on daily closes.
Drawdown from the previous peak
Underwater curve · daily closes
Five deepest drawdowns of the portfolio
| Start | Trough | Recovered | Depth | Duration | BTC / ETH at the time |
|---|---|---|---|---|---|
| 2022-01-22 | 2022-03-03 | 2022-06-12 | −17.8% | 141 days | −17.4% / −18.0% |
| 2019-09-18 | 2020-01-11 | 2020-02-12 | −17.8% | 147 days | −4.2% / −24.1% |
| 2026-02-05 | 2026-05-31 | 2026-08-19 | −14.3% | 195 days | −13.0% / −14.5% |
| 2022-07-18 | 2022-10-21 | 2023-01-11 | −13.1% | 177 days | 1.9% / −16.2% |
| 2020-09-01 | 2020-10-08 | 2021-03-29 | −12.8% | 209 days | −3.4% / −16.5% |
Rolling 12-month Sharpe ratio
365-day windows of daily returns
Portfolio: 100% of 12-month windows positive, 93% above 1, lowest 0.27, median 1.94.
Calendar
Years and months
Annual returns
*partial years
Portfolio monthly returns (%)
Green: gain · red: loss · color steps of 4 %, capped at 20 %
| Year | BTC | ETH | Portfolio | Portfolio Sharpe | Portfolio Sortino | Portfolio max DD |
|---|---|---|---|---|---|---|
| 2018* | +16.2% | +224.3% | +120.3% | 5.22 | 24.38 | −6.4% |
| 2019 | +55.5% | −9.0% | +8.0% | 0.45 | 0.76 | −14.3% |
| 2020 | +21.2% | +77.5% | +56.1% | 1.82 | 3.27 | −12.8% |
| 2021 | +28.3% | +60.4% | +50.9% | 1.80 | 4.02 | −11.1% |
| 2022 | +40.1% | +94.5% | +80.9% | 2.00 | 5.17 | −17.8% |
| 2023 | +59.3% | +22.3% | +29.5% | 1.35 | 2.98 | −7.1% |
| 2024 | +105.2% | +66.3% | +75.6% | 3.13 | 10.00 | −4.7% |
| 2025 | +1.2% | +112.7% | +81.6% | 2.24 | 5.49 | −8.6% |
| 2026* | +13.8% | +63.9% | +56.1% | 1.95 | 4.60 | −14.3% |
Portfolio
Diversification
Bitcoin and Ethereum prices are correlated at 0.82. The two strategies only at 0.36: each is out of the market most of the time, and on most active days only one of them moves.
| Measure | Value |
|---|---|
| Sharpe: portfolio / better market | 1.050 (+5.0%) |
| Volatility: portfolio / weighted markets | 0.939 (−6.1%) |
| Max drawdown: portfolio vs BTC, ETH | −17.8% vs −24.8%, −24.1% |
| Diversification ratio | 1.21 |
| Effective independent bets | 1.46 of 2 |
| Share of risk at 50/50: BTC / ETH | 40% / 60% |
| Correlation of strategy returns | Value |
|---|---|
| Daily, Pearson | 0.36 |
| Daily, Spearman | 0.27 |
| Weekly | 0.39 |
| Monthly | 0.37 |
| Days when both were exposed | 0.54 |
| Rolling 90 days: median (P10 to P90) | 0.35 (0.02 to 0.66) |
Rolling 90-day correlation
BTC vs ETH strategy returns
50/50 in capital is not 50/50 in risk: ETH is the more volatile market (33% against 25% a year) and carries 60% of the variance. Without rebalancing the weights drift; BTC averaged 25% of the portfolio and ended at 11%.
Stress
Behavior in market stress
Stress is defined from spot prices only, never from strategy results. A constant-dependence benchmark (same daily returns, Gaussian copula fitted to the full-period correlation) shows what a selection of days would produce without contagion.
| Episode | Window | BTC spot | ETH spot | Portfolio |
|---|---|---|---|---|
| 2018 Q4 capitulation | 2018-11-08 → 2018-12-15 | −51.1% | −61.8% | +41.7% |
| COVID-19 crash | 2020-03-01 → 2020-03-31 | −24.7% | −38.9% | −0.3% |
| May–July 2021 sell-off | 2021-05-10 → 2021-07-20 | −48.9% | −54.5% | +9.0% |
| Terra/LUNA and 3AC | 2022-05-05 → 2022-06-30 | −49.8% | −63.6% | +30.2% |
| FTX collapse | 2022-11-06 → 2022-11-30 | −19.4% | −20.4% | +0.5% |
| August 2024 carry unwind | 2024-08-01 → 2024-08-10 | −5.7% | −19.3% | +1.2% |
| Oct–Nov 2025 deleveraging | 2025-10-06 → 2025-11-30 | −26.8% | −33.7% | +7.6% |
Episode windows were fixed from known market events. The portfolio was positive in 6 of 7.
| Days used | Days | Correlation | Benchmark |
|---|---|---|---|
| All days | 2,952 | 0.36 | — |
| Non-stress days | 1,459 | 0.32 | — |
| Market crash days (worst 5 % of BTC+ETH spot) | 148 | 0.53 | — |
| Crypto bear market (spot > 30 % below 1-year high) | 1,370 | 0.38 | — |
| Named stress episodes | 289 | 0.62 | — |
| Both markets lost on the day | 153 | 0.56 | 0.20 [0.06, 0.36] |
| P(ETH loses | BTC loses) | — | 39% | 42% [38%, 45%] |
Correlation is 0.53 on market crash days and 0.38 in bear markets, against 0.32 on non-stress days, within the range of 90-day rolling correlations. Joint losses are not more frequent than constant dependence implies. On days when both markets lose, the size of their losses is more aligned than the benchmark, consistent with some trades being the same shock traded on both markets (23% of ETH entries are within 15 minutes of a same-direction BTC entry).
Robustness
How fragile are these numbers?
Resampled histories
5,000 histories built from 97 calendar months drawn at random, the same months for BTC and ETH.
| Statistic | Historical | 90 % interval |
|---|---|---|
| Portfolio Sharpe | 2.02 | 1.53 to 2.51 |
| Portfolio CAGR | 67.7% | 40.1% to 107.1% |
| Portfolio max drawdown | −17.8% | −30.6% to −13.7% |
P(Sharpe > 1) 100.0%. A bootstrap resamples history; it is not a forecast.
Sub-periods
The second half is weaker than the first, and the two markets are more correlated in it.
| Period | Sharpe | Sortino | CAGR | Correlation |
|---|---|---|---|---|
| 2018-09-01 → 2022-09-15 | 2.12 | 4.89 | 78.7% | 0.31 |
| 2022-09-16 → 2026-09-30 | 1.98 | 4.73 | 50.9% | 0.44 |
Costs, slippage and funding
| Commission per order | Sharpe | CAGR | Max DD |
|---|---|---|---|
| 0.000 % (×0) | 2.17 | 74.5% | −17.2% |
| 0.045 % (×1) | 2.02 | 67.7% | −17.8% |
| 0.090 % (×2) | 1.88 | 61.2% | −19.7% |
| Added cost (not in headline figures) | Sharpe | CAGR |
|---|---|---|
| Slippage 0.01 % per order | 1.99 | — |
| Slippage 0.02 % per order | 1.96 | — |
| Slippage 0.05 % per order | 1.86 | — |
| Funding: historical Binance perpetual funding (2020 onwards; constant +0.01 % per 8 h before 2020) | 1.97 | 64.7% |
| Funding: same, with Hyperliquid funding from May 2023 | 1.95 | 64.0% |
| Funding: stress: shorts pay 0.01 % per 8 h at all times | 2.00 | 66.7% |
Turnover is about 43× the capital a year (entries only), so commissions take about 7% of CAGR. Funding uses the historical rates of the perpetual market, with the correct sign for long and short positions.
Concentration
The result depends on the right tail: the best 5 % of trades account for 99% of the portfolio's log growth, and without them the CAGR would be −0.3%. This is the profile of a momentum strategy: many small losses, few large gains.
Rebalancing
The portfolio has no rebalancing; this was decided before any result. As a diagnostic only, monthly rebalancing to 50/50 gave a Sharpe of 2.08 (against 2.02). It is recorded as a hypothesis for forward testing, not adopted.
Method
Method and limitations
- Fixed rules. One 15-minute strategy, with parameters selected on Bitcoin. Ethereum runs it unchanged, without any ETH calibration. The portfolio reproduces the validated backtests exactly before any calculation; 40 of 40 consistency checks pass.
- Calm regime. New positions are only opened when the daily volatility regime is calm; an open position keeps its exits if the regime changes.
- Short entries. Opened only when the daily trend regime is bearish.
- Data. BTC/USD · Bitstamp spot; ETH/USDT · Binance spot; 15-minute bars, UTC.
- Capital. 50/50 at the start; each half trades 100 % of its own equity, one position at a time, without leverage. Average gross exposure 16% of the portfolio, maximum 104%.
- Costs. Commission 0.045 % per order on every fill. Slippage and perpetual funding are not included in the headline figures. All 1,150 executions since 2020 were replayed on real Binance perpetual-futures trades. With the bot as now coded, execution costs average about 2.3 bps per execution at $1M of capital, for a net Sharpe of 1.92 over 2020–2026 (resampled 90 %: 1.40 to 2.41, funding excluded), still 1.83 at $5M, the largest capital tested, in the central estimate. Capacity on Hyperliquid, the intended venue, and our own market impact are not yet measured. Institutional dossier (PDF).
- In-sample. The parameters were chosen on BTC data from 2017 to 2026, which contains the whole period; ETH is a transfer across markets, not across time. The two entry conditions (calm regime, short-side trend) were set in October 2026 with knowledge of both markets, so BTC and ETH are both in-sample for them.
- Venues. The data venues (Bitstamp, Binance spot) differ from a perpetual-futures execution venue: prices, fees, funding and stop mechanics differ.
The research page lists the studies behind the strategy.
Reproducibility
Downloads
The full report contains every table behind this page: reconciliation with the validated backtests, risk contributions, crisis correlations, trade overlap, drawdown episodes and all 40 consistency checks.
- Research paper (PDF)
pdf - Full research report (HTML)
html - Daily returns and equity (CSV)
csv - Monthly returns (CSV)
csv - Annual returns (CSV)
csv - Drawdown episodes (CSV)
csv - Rolling correlations (CSV)
csv - Machine-readable summary (JSON)
json - Research manifest: versions and hashes (JSON)
json
| Reproducibility | |
|---|---|
| Research version | shock-engine-2026.10.13 |
| Code commit | 997435b976491bffcc356ce8b3e916cbcd02a17b |
| Parameter hash (SHA-256) | c6ac30464bce14ddb6e5625cbcf08f68e69e3cc5e5454a0c2268583070a74435 |
| Generated | 2026-10-13 |
Research
How the numbers were tested
Placebo entries, delayed entries, parameter neighborhoods, cost stress, and the markets where the rules failed.
Important information. All performance figures on this site are hypothetical: they come from a historical simulation of fixed rules, after modeled commissions, and do not represent actual trading. Simulated results have inherent limitations. They benefit from hindsight in the design of the rules, they do not reflect slippage, funding, market impact or the liquidity of a real account, and past simulated performance does not predict future results.
Nothing on this site is investment advice, a recommendation, or an offer or solicitation to buy or sell any financial instrument or to provide investment management services.