InsidersTradesSigma
V15 has ranked signals since 2026-07-05. Its observed history is still too short to publish a live return. Results below are versioned historical simulations.
Each month, the simulated portfolio buys the 10 highest-scored insider BUY signals (EU markets) in equal weights, then sells each position 90 days later. Result is net of 0.6% round-trip fees.
Pick a market and a period: every figure below is reproducible from the committed dataset, cell by cell. Cells with fewer than 30 observations say so instead of showing a fragile number.
Monthly top-10 backtest of the served ranking under strict point-in-time discipline, hold T+90, net of costs (audit 200).
Gold dot: 2025-2026 ran on an exceptional regime, never a guaranteed base rate.
Net of a 0.6% round-trip cost, returns winsorized at +-50%, T+90 hold. Figures were deliberately revised down in audit 200 to remove look-ahead bias: every input is point-in-time. Source: SCENARIO_RETURNS (audit 200). Backtest on public filings, not investment advice. Past performance does not predict future returns. Full methodology →
Top 5 / 10 / 20
What is a "top N signals" portfolio, exactly?
Each month we keep the filings best scored by the composite score (0 to 100), buy them in equal weights, hold each 90 days, then rotate. Top 5, 10 or 20 is simply how many of those monthly best you follow.
Full backtest window, 51-month backtest (2022-2026Q1, point-in-time strict), no cherry-picked sub-period. Each return is shown with its max drawdown and its sample size.
| Signals / month | Index base reaches | Annualized | Max drawdown | signals |
|---|---|---|---|---|
| Top 5 / month | €56,970 | +50.6% | -31.2% | 255 |
| Top 10 / monthServed configuration | €27,069 | +26.4% | -26.6% | 510 |
| Top 20 / month | €13,802 | +7.9% | -31.0% | 1,020 |
This window leans on the strong 2025-26 tape. Treat it as a single favorable regime, never an annualized promise.
Last updated: 01 Sept 2026, 09:18 · Period covered: 119 months
Regime-dependent simulation on real data, not a live return. Point-in-time strict: the figures were deliberately revised down (audit 200) to remove look-ahead bias, and re-measured on the repaired corpus (audit 216). The live ranking has run this construction since 2026-08-16, but these figures remain a backtest: no live window exists yet, and the served book covers more venues than the backtest universe. Past performance does not predict future returns.
Every day, we freeze the BUY ranking and its entry price. Choose a rank group and a fixed horizon to see realized results with no retrospective reshuffling.
Executives buy their own company's shares when they believe in it. We monitor public filings across 43 markets and score eligible purchases from 0 to 100. Each purchase is ranked within its own market. No market currently meets the historical allocation criteria. Signals remain visible as watch items with no suggested allocation.
How the returns are produced
Regime-dependent simulation on real data, not a live return. Point-in-time strict: the figures were deliberately revised down (audit 200) to remove look-ahead bias, and re-measured on the repaired corpus (audit 216). The live ranking has run this construction since 2026-08-16, but these figures remain a backtest: no live window exists yet, and the served book covers more venues than the backtest universe. Past performance does not predict future returns. Detailed methodology ↗
Internal backtest on public regulator filings (43 markets). Window Mar 2023 · Aug 2025. T+365 returns from Yahoo Finance EOD prices.
Historical backtest. Returns are indicative; past performance does not guarantee future results. This is not investment advice. CI95 and DSR documented in the methodology.
Backtest coverage: XPAR, XTKS and BVMF prioritised (T+365 returns computed). SEC (US) and LSE (UK) still being processed.
V14e archive, the former out-of-sample test
V14e EU OOS · archived strategy (2026-05 to 2026-07) · test window 2025-01-01 to 2026-03-31 · net of 0.6% per round-trip
Sigma v1.0, in-sample discovery cohort: n=3,864 of 391,106 observations (43 markets), 2022-2025.
On the 1st of every month, the V14e ensemble is refit on a 36-month window. The OOS Sharpe is measured on the following 6 full calendar months, shifted back 93 days so every T+90 return is realized, with a bootstrap 95% CI and the number of signals tested. A Sharpe below 0 (amber badge) flags a drift to watch. Fewer than 6 full months: the row is marked insufficient and no figure is published.
| Version | Fitted at | Sharpe OOS | CI95 | Months | Signals | CAGR | MaxDD | Deflated |
|---|---|---|---|---|---|---|---|---|
| V13.202609DRIFT | 01 Sept 2026 | -5.76 | [-38.62, -3.01] | 6 | 23,859 | -80.0% | -55.3% | -9.08 |
| V13.202608DRIFT | 21 Aug 2026 | -1.93 | [-18.36, 0.73] | 6 | 19,147 | -49.2% | -35.9% | -5.24 |
| V13.202607DRIFT | 21 Aug 2026 | -2.27 | [-7.41, 0.28] | 6 | 15,502 | -67.2% | -42.7% | -5.58 |
| V14e | 04 Jun 2026 | 0.83 | [-0.94, 3.22] | 15 | · | +27.1% | -25.8% | 0.21 |
The deflated Sharpe (Bailey-Lopez de Prado) corrects for multiple-testing selection bias. A negative value means the raw Sharpe is not statistically distinguishable from zero over this window.
A different exercise with different rules: a conservative Top 5 reinvested basket, fees included, not comparable to the Top 10 backtest shown above.
180 unlevered configurations with 0.6% round-trip costs. Parameters selected on 2022-2024, then frozen before the 2025-2026Q1 test.
Monthly top 5, 90-day hold. The rule was selected on training data, then evaluated unchanged on the test period.
Historical research simulation. Not expected return, investment advice or a guarantee. Market cap and FX are not yet fully point-in-time.
This measures what a reader captures AFTER publication, not the strategy's performance.
Key question for a retail investor: how much of the stock's total return happens between the insider trade and the regulatory filing (leak) vs. after (exploitable by you)? We measured the Yahoo price at transactionDate, at pubDate+1, and at tx+90d for 76,401 trades.
Local regulations (MAR 596/2014 in the EU for AMF/BaFin, SEC Form 4 in the US, SIX SER in Switzerland) require executives to file their transactions within a short window (3 business days in the EU, 2 in the US). But do they comply? On our sample of 103,702 filings:
Conclusion: the majority (70%) comply with MAR. The median is 1.5 days, meaning that when you see a signal on the site, the transaction already occurred 1.5 days ago on average. The market has had time to price the information.
returnFromPub30d/90d/365d (entry at pubDate+1) in addition to insider-view returns (transactionDate). Future backtests will be more honest by default.Sigma reduces a global public-filings dataset to a watchlist worth examining. The ranking prioritises research. It currently provides neither a validated allocation nor an expected return.