Historical simulation - frozen and reproducible
The historical simulation record.
A concentrated quantitative growth strategy with systematic market defense, evaluated as one complete tradable strategy.
Fund status: Closed family fund · Founder and affiliated family capital
Evidence through
2026-08-04
Complete-strategy evidence
34.35% annualized return. -12.37% maximum drawdown.
Arbiter's concentrated growth strategy with systematic defense produced a Sharpe ratio of 1.718 across the 2018-01-09 through 2026-08-04 simulation, after modeled trading and ETF-borrow costs. Sharpe ratios use a 0% risk-free rate. Annualized return compounds at 252 trading sessions per year. Sortino downside deviation uses a 0% target averaged over all sessions. On the 2018-01-09 through 2022-12-30 discovery window, the 10 bps trading / 75 bps borrow scenario had an observed Sharpe of 1.648 and a 99.96% Deflated Sharpe probability after 183 searched formulas were deduplicated into 169 economic paths and every path was treated as independent. Under the declared model, that is strong evidence that the discovery-window Sharpe exceeds the selection-adjusted hurdle. Historical simulation, not investor performance or a forecast.
Historical simulation - frozen and reproducible
Growth of $1
Arbiter $12.46 | NASDAQ-100 Index Proxy $4.70
Scroll horizontally to explore the full chart.
2018-01-09 through 2026-08-04 · 2,153 aligned trading sessions · Arbiter 12.46x · Benchmark 4.70x
Historical simulation - frozen and reproducible
Drawdown
Arbiter -12.37% | NASDAQ-100 Index Proxy -35.12%
Scroll horizontally to explore the full chart.
Arbiter and NASDAQ-100 Index Proxy monthly drawdown · Arbiter maximum -12.37% · Benchmark maximum -35.12%
What the historical evidence means
A pristine out-of-sample test - where the strategy is fixed before results are observed - is the strongest form of historical validation.
Arbiter's later historical period was reviewed during final strategy development, so we deliberately classify it as extension and robustness evidence, not pristine out-of-sample evidence. It shows that the strategy's behavior persisted across additional market regimes, but it is not treated as independent confirmation.
We research, execute and monitor Arbiter daily. The evidence presented here is historical research; continued private account monitoring does not establish live validation of the strategy.
| Metric | Arbiter | NASDAQ-100 Index Proxy |
|---|---|---|
| Annualized return | 34.35% | 19.87% |
| Sharpe ratio | 1.718 | 0.877 |
| Maximum drawdown | -12.37% | -35.12% |
| Maximum underwater | 142 sessions | 493 sessions |
| Ending growth of $1 | $12.46 | $4.70 |
| Annualized volatility | 18.16% | 23.94% |
| Sortino ratio | 2.657 | 1.249 |
| Calmar ratio | 2.777 | 0.566 |
| Beta to benchmark | 0.456 | 1.000 |
| Correlation to benchmark | 0.602 | 1.000 |
| Positive sessions | 54.48% | 56.06% |
NASDAQ-100 Index Proxy · QQQ ETF · Buy-and-hold total return with dividends reinvested (dividend-adjusted closes), rebased to $1.00 on 2018-01-09 and aligned to every Arbiter session through 2026-08-04. Sharpe ratios use a 0% risk-free rate. Annualized return compounds at 252 trading sessions per year. Sortino downside deviation uses a 0% target averaged over all sessions.
Calendar record
Year-by-year compounding.
| Period | Arbiter | NASDAQ-100 Index Proxy | Difference vs NASDAQ-100 |
|---|---|---|---|
| 2018 | +1.85% | -4.29% | +6.14% |
| 2019 | +12.69% | +38.96% | -26.27% |
| 2020 | +75.03% | +48.62% | +26.42% |
| 2021 | +65.74% | +27.42% | +38.32% |
| 2022 | +14.10% | -32.58% | +46.68% |
| 2023 | +34.18% | +54.85% | -20.68% |
| 2024 | +39.76% | +25.58% | +14.19% |
| 2025 | +36.14% | +20.77% | +15.38% |
| 2026 through August 4 | +28.48% | +17.98% | +10.51% |
Every displayed calendar period is positive. The weakest relative calendar period was 2019: Arbiter returned +12.69% while the NASDAQ-100 Index Proxy returned +38.96%.
Evidence & validation
The result survives the search.
Arbiter's complete strategy clears the Bailey–López de Prado multiple-testing check under both a dependence-adjusted trial estimate and the deliberately punitive assumption that every deduplicated economic path was independent.
Research windows
| Window | Period | Annualized return | Sharpe | Max drawdown | Max underwater | Status |
|---|---|---|---|---|---|---|
| Discovery | 2018-01-09 through 2022-12-30 | 30.76% | 1.648 | -11.65% | 142 sessions | Selection window for the registered simplified complete strategy within the frozen defense-formula family. |
| Reused extension | 2023-01-03 through 2026-08-04 | 39.51% | 1.810 | -12.37% | 117 sessions | Already inspected during final selection; descriptive continuation, not pristine out-of-sample evidence. |
| Full record | 2018-01-09 through 2026-08-04 | 34.35% | 1.718 | -12.37% | 142 sessions | Combined discovery and reused-extension record at the registered 10 bps trading-cost and 75 bps borrow frame. |
The strategy was selected on the discovery window. The later extension remained strong, but it was reviewed during final selection and is therefore labeled reused evidence - not a pristine holdout. This page presents no new out-of-sample confirmation. Any future strategy evaluation requires separately validated evidence; private account monitoring does not validate the historical simulation.
Deflated Sharpe Ratio
The Deflated Sharpe Ratio adjusts the selected complete-strategy Sharpe for the measured search family, sample length, cross-path dispersion, skewness, and kurtosis. It validates the complete strategy within the search family and does not independently validate the defensive overlay.
| Scenario | Trial treatment | Observed Sharpe | Deflated Sharpe probability |
|---|---|---|---|
| Registered punitive | All 169 distinct economic paths independent | 1.648 | 99.96% |
| Registered correlation adjusted | Average-correlation implied independent trials | 1.648 | 99.98% |
| Selection punitive | All 169 distinct economic paths independent | 1.735 | 99.98% |
| Selection correlation adjusted | Average-correlation implied independent trials | 1.735 | 99.99% |
Sharpe ratios use a 0% risk-free rate. Annualized return compounds at 252 trading sessions per year. Sortino downside deviation uses a 0% target averaged over all sessions. 183 evaluated formulas deduplicated into 169 economic paths over 1,254 discovery sessions; average path correlation 0.99295, and the correlation-adjusted scenario uses 2.24 effective trials. The hurdle also reflects the record length and the measured non-normal shape of the returns. Read the Bailey and López de Prado paper ↗
Robustness
| Cost model | CAGR | Sharpe | Maximum DD | CAGR/DD | Maximum UW |
|---|---|---|---|---|---|
| Gross frame · no trading or borrow costs | 41.51% | 2.006 | -11.51% | 3.605 | 127 sessions |
| 10 bps trading / 75 bps borrow · registered net frame | 34.35% | 1.718 | -12.37% | 2.777 | 142 sessions |
| Gross frame · discovery window | 37.50% | 1.942 | -11.02% | 3.402 | 127 sessions |
| 10 bps trading / 75 bps borrow · discovery window | 30.76% | 1.648 | -11.65% | 2.640 | 142 sessions |
The full record remains viable through a realistic acceptance cost and a materially harsher stress frame. The calendar record above shows the same path year by year.
Strategy structure
| Component | Rules | Description |
|---|---|---|
| Stress signal | 1 | Aggregates the declared breadth-stress inputs into one bounded activation signal. |
| Cyclic controller | 2 | Uses a bounded active interval and a bounded inactive interval while the stress signal persists. |
| Funded defense action | 2 | Caps the liquid ETF short and funds it only from available defensive collateral. |
Declared complexity score 5.
What the evidence establishes
Under the stated DSR model, the complete strategy has a 99.96% probability of exceeding the selection-adjusted Sharpe hurdle after the measured search, record length, and return-distribution shape are included.
Incremental recovery diagnostic: Does not pass 95% incremental DSR - The complete simplified strategy passes the stage-specific DSR within its immediate search family. The defensive overlay does not independently clear the 95% threshold; its case remains bounded risk logic, drawdown behavior, temporal diagnostics, and prospective observation rather than a standalone-alpha claim.
- The 2023–2026 extension was already inspected during final strategy selection and is not a pristine out-of-sample holdout.
- The complete-strategy DSR does not independently validate the defensive overlay or establish standalone defensive alpha.
- The calculation covers the immediate 183-formula defense family and 169 distinct economic paths, not the full research lineage.
- The frozen record does not include the 2 bps daily selection path; its scenarios conservatively reuse the registered discovery skewness and kurtosis with the reported 0.03491 selection-frame Sharpe dispersion.
- Cost sensitivity is reported from the two published frames - the gross frame and the registered 10 bps trading / 75 bps borrow net frame - over the full record and the discovery window; intermediate cost levels are not separately simulated.
- Results are gross of prospective investor fees, vehicle expenses, taxes, financing outside the declared borrow assumption, and other implementation differences.
Historical simulation cannot establish execution against real liquidity under real operations. We research, execute and monitor Arbiter daily. The historical record is research evidence, not live validation or a promise of future results.
Execution
Market-on-open portfolio rebalance timing
Trading cost
10 basis points per unit of turnover
Financing
75 basis points annualized for modeled ETF borrow exposure
Important information
Arbiter is a closed family fund currently focused on the Founder's and affiliated family capital. No interests are offered and no capital is solicited on this page. Nothing on this page constitutes SEC registration, review, approval, or endorsement.
The results on this page are a historical simulation, not performance from an actual client account. Fundreporting reproduced the displayed growth multiple, annualized return, Sharpe ratio, maximum drawdown, maximum underwater duration, and expanded ratios directly from every return in the frozen daily series and checked the release hashes. Sharpe ratios use a 0% risk-free rate. Annualized return compounds at 252 trading sessions per year. Sortino downside deviation uses a 0% target averaged over all sessions. This is reproducibility, not independent audit, actual performance, or regulatory approval.
The discovery window is 2018-01-09 through 2022-12-30. The 2023-01-03 through 2026-08-04 extension was reviewed during final strategy selection and is not represented as an untouched holdout. DSR is a probability that the selected strategy's Sharpe exceeds the estimated selection hurdle under the declared trial-family and non-normality assumptions; it is not a forecast, a guaranteed return, or the probability of making money.
The simulation covers 2018-01-09 through 2026-08-04 and assumes Market-on-open portfolio rebalance timing; 10 basis points per unit of turnover; and 75 basis points annualized for modeled ETF borrow exposure. Benchmark means the QQQ ETF used as a total-return proxy (dividends reinvested) for the Nasdaq-100; QQQ is not the index itself. It is a same-window reference, not an investable account result or a statement that Arbiter will outperform the benchmark. Arbiter's simulated returns exclude dividends on its long holdings while the benchmark includes them; the methodologies differ in this respect. No fund management fee or incentive allocation is applied in the simulation; taxes and account-specific charges are not modeled. The strategy is concentrated and can incur substantial loss. Model selection, reused research windows, data availability, survivorship and look-ahead controls, execution timing, liquidity, short exposure, financing, and market-regime change can cause actual results to differ materially. We research, execute and monitor Arbiter daily. Ongoing account monitoring does not validate historical simulation. Past performance does not guarantee future results.