USPRMEXNAS
Cluster · 2 insiders
Filed
May 2025
Amount
130 k€
Sigma score
63 / 100
Return T+365
+130.2%
InsidersTradesSigma
Last updated: 14 Jul 2026, 10:50 · Period covered: 120 months
Live · Out-of-sample
Realized performance of live recommendations
Every day at 3 a.m. UTC we freeze the engine's top 10 BUY and top 10 SELL, tag the CAC 40 market regime, then measure realized returns at T+30 / T+90 / T+365. No retrospective revision.
TRACKERsince:27 May 2026 · snapshots:960 · T+30:pending · T+90:pending · win-rate:pending · regime:Range
Executive Summary
30-Second Takeaways
Avg. filing delay
2.5d
55% of filings within 3d (MAR)
Pre-publication info leak
10%
of the T+90 return happens BEFORE you see the signal
Validated selection (screen)
t=5.70 / 2.84
Score top-decile residual, 2 out-of-sample windows (audit 168) · n=3,864 qualified signals · below transaction costs: a screen, not alpha
Point-in-time backtest (audit 169)
3 windows
PIT Sharpe +0.05 (2022) · -1.99 (2023-2024, CI95 [-3.29, -0.76] negative) · +0.53 (2025-2026Q1) · Point-in-time backtest: negative in 2023-2024, positive in 2025-2026. The result is regime-dependent; no standalone alpha is demonstrated.
What Sigma is for, concretely.The system is not designed to replace an index ETF, it's a pre-screening tool that turns 585 companies into 10 to 15 files per week worth a real look. Over the period 2021-2026 (CAC at all-time highs), our best filtered strategy delivers +5.1% CAGR net of fees with a Max DD of -45%, against an index at +6.5%. Absolute alpha isn't there on this particularly bullish cycle; the signal's primary value remains: identifying companies where executives themselves are taking positions, to guide your fundamental research. Concretely, a file stands out when several executives of the same company buy their own stock within a few weeks of each other (a cluster): it is the signal that holds up best over time.
In-sample vs out-of-sample
Discovery (IS) vs out-of-sample validation (OOS)
| Metric | In-sample (2021-25) | Out-of-sample (live) |
|---|---|---|
| Sharpe (ann) | -0.05 ± 10.8 | Pending (n<30) |
| Win rate | 51.5% | Pending (n<30) |
| Mean 90d return | +2.60% | Pending (n<30) |
| Max DD | -45% | Pending (n<30) |
| n trades | 3,864 | 0 |
Reading this table correctly. All IS figures come from the discovery phase. Bootstrap CIs (§06, §08) describe estimation uncertainty, not overfit risk. The only grid-search-independent measure is the live OOS, published here once n=30 trades have settled.
How to read
How to read our return figures
All our return metrics follow the same convention. They are deliberately conservative, it's the return a retail investor can actually capture, not the insider's own return.
T+90
3-month return
Price change between the day after the regulatory filing (pubDate+1) and 90 calendar days later. Our recommended hold horizon.
T+365
1-year return
Same principle over 365 days. Useful to see if the signal holds long-term, not for annualising, it's already a 12-month absolute return.
Median vs Mean
Read the median
The median is the "middle" trade, more representative. The mean is dragged by outliers (a +400% trade artificially inflates the mean).
Worked example
On March 15, an LVMH executive files a €5M stock purchase with the AMF. On March 16 (next day open), the price is €700, that's your P0, the price YOU could have entered at.
On June 16, 90 days later, the price is €742. Return T+90 = (742 − 700) / 700 = +6.0%.
On March 16 the following year, 365 days later, the price is €840. Return T+365 = (840 − 700) / 700 = +20.0%.
On a table, if this trade were part of a 50-trade "cluster CEO/CFO" cohort, its +6.0% / +20.0% would count as one data point in the displayed median and mean.
Don't confuse the two. T+90 and T+365 are per-trade absolute returns. They are NOT annualised, NOT compounded. The strategy CAGR(section 3) is different: it's the annualised return of a portfolio that rebalances every 3 months and holds the top 20 current signals simultaneously. The two metrics answer different questions: "how much does a typical trade yield?" vs "how much does the whole portfolio yield over a year?".
Part 1 · freshness
How long between the trade and its filing?
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 392,505 filings:
Median
2.5d
p25
1.5d
p75
5.6d
p90
24.0d
≤ 3d (MAR)
55%
Conclusion: the majority (55%) comply with MAR. The median is 2.5 days, meaning that when you see a signal on the site, the transaction already occurred 2.5 days ago on average. The market has had time to price the information.
Part 2 · information leak
Has the market already moved before the filing?
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 224,286 trades.
Total T+90 return (insider view)
+14.01%
What the insider would have captured buying and selling on the exact day.
Leak tx → pubDate
+1.45%
The market moves without you. On our sample: 10% of total return.
Retail T+90 (your view)
+14.22%
What YOU capture entering at pubDate+1 and holding 90d.
Good news: the information leak is moderate (10% of total return). You don't lose much by waiting for the filing. The retail return (pubDate+1 → pubDate+90d) is nearly identical to the insider return. The real problem isn't freshness, it's signal selection (see part 3).
Part 3 · selection
Which strategies actually work?
We tested 6 strategies on the same universe, with the same rules:
- Top-20 portfolio · we buy the 20 best signals from the previous month, sorted by score.
- Monthly rebalancing · old positions out, new ones in.
- 3-month hold per position (matches the T+90 backtest window).
- Brokerage fees 0.6% round-trip (French retail price).
- Entry at
pubDate+1(the day after the regulatory filing).
Two numbers, two distinct populations. The +9.4% CAGR shown here (Passive row) is the naïve strategy on the full universe (top 20 each month, no filter). The Sigma v1.0 cohort (in-sample discovery, pre-V14e) tightens selection (composite score ≥ 50 + high-rank insider or cluster) and delivers +5.1% CAGR, Sharpe 0.38, max drawdown -44.7% over 2732 qualified signals.These are the historical discovery numbers. The production strategy is V14e EU_strict (OOS Sharpe 0.53, see hero above).
| Strategy | Signals | CAGR | Sharpe | Max DD | Win % | Beats CAC |
|---|---|---|---|---|---|---|
Passive · all buy signals Naïve strategy: buy the top-20 scores every month regardless of filters. | 428,725 | +9.42% | 0.68 | -58.4% | 59% | 53% |
Score filter ≥ 50 Buy only when our composite score (v2) exceeds 50. | 3,670 | +2.39% | 0.22 | -48.8% | 52% | 47% |
Cluster only Only trades where ≥ 2 insiders bought the same company within ±30 days · collective conviction signal. | 16,792 | +2.85% | 0.24 | -49.5% | 52% | 47% |
CEO / CFO only Filter by role: only trades by CEOs and CFOs (the best-informed insiders). | 83,911 | +10.57% | 0.67 | -44.5% | 57% | 51% |
Trade ≥ €500k + Cluster Material conviction: trades of at least €500,000 inside a cluster. | 1,361 | -5.34% | -0.29 | -49.7% | 48% | 46% |
Sigma v1.0 · narrow discovery (pre-V14e) In-sample discovery cohort, CEO/CFO filter only. Pre-V14e baseline. The production strategy is V14e EU_strict (OOS Sharpe 0.53, see hero). | 2,732 | +5.09% | 0.38 | -44.7% | 52% | 50% |
| CAC 40 (buy & hold) | passive | +6.47% | 0.48 | · | · | · |
Reading the table: the Signals column shows how many trades matched the filter over the full period. Sharpe = annualised return/risk ratio.Max DD = worst cumulative loss during the simulation.Beats CAC = % of months the strategy outperformed the CAC 40.
T+90 ↔ CAGR conversion. T+90 returns are per-trade, 90-day holds; CAGR is the compounded annual return of a portfolio that recycles the signal continuously. Example: a +2.8 % average T+90 compounded over ~4 cycles per year ≈ +12 % CAGR (formula: (1+0.028)365/90 − 1). The two numbers are consistent; they answer two different questions ("how much per trade?" vs "how much per portfolio-year?"). The published win rate is the % of signals closing in profit at T+90.
Reading: over 2021-2026, a period when the CAC 40 reached all-time highs (+6.5% annualised), the ★ Sigma strategy delivers +5.1% net CAGR with a Max DD of -45% and 50% of months beating the index. Filtered strategies serve as a selection foundation; their value shows most clearly during drawdowns where insider signals have historically held up better.
Real examples · case studies
Some real signals we caught
Internal backtest on public regulator filings (43 markets). Window Mar 2023 · Aug 2024. T+365 returns from Yahoo Finance EOD prices.
USLTRXXNAS
Cluster · 4 insiders
Filed
Mar 2025
Amount
37 k€
Sigma score
62 / 100
Return T+365
+110.4%
USHTLDXNAS
GERDIN MICHAEL J
Chief Executive Officer
Filed
May 2025
Amount
1 M€
Sigma score
65 / 100
Return T+365
+94.1%
USISBAXNAS
Cluster · 10 insiders
Filed
Jan 2025
Amount
14 k€
Sigma score
60 / 100
Return T+365
+90.5%
FRSMCP.PAXPAR
Cluster · 4 insiders
Filed
Apr 2025
Amount
26 k€
Sigma score
60 / 100
Return T+365
+90.1%
USTTIXNAS
SERRANO ELIJIO V
Sr. Vice President & CFO
Filed
Mar 2023
Amount
97 k€
Sigma score
52 / 100
Return T+365
+67.2%
USMMLPXNAS
Cluster · 6 insiders
Filed
Feb 2024
Amount
3 k€
Sigma score
64 / 100
Return T+365
+60.3%
USBOTJXNAS
Cluster · 4 insiders
Filed
May 2025
Amount
35 k€
Sigma score
61 / 100
Return T+365
+60.3%
USBOTJXNAS
UMBERGER HARRY P JR
EVP - Chief Credit Officer
Filed
Oct 2023
Amount
1 k€
Sigma score
52 / 100
Return T+365
+56.9%
FRFGA.PAXPAR
Jean-Claude MAILLARD
Chairman & CEO
Filed
Oct 2023
Amount
26 k€
Sigma score
54 / 100
Return T+365
+50.9%
ESPSG.MCXMAD
YIRAYIRA INTERNATIONAL, S.L.
HELENA REVOREDO DELVECCHIO - PRESIDENTE DEL CONSEJO DE ADMINISTRACIÓN
Filed
Apr 2024
Amount
71.1 M€
Sigma score
50 / 100
Return T+365
+50.0%
USMRBKXNAS
Lindsay Denise
EVP, Chief Financial Officer
Filed
May 2024
Amount
17 k€
Sigma score
57 / 100
Return T+365
+49.0%
USCBANXNAS
Fountain T Heath
President and CEO
Filed
Feb 2024
Amount
9 k€
Sigma score
58 / 100
Return T+365
+48.8%
USQCRHXNAS
Ziegler Marie Z.
Director
Filed
Apr 2023
Amount
6 k€
Sigma score
50 / 100
Return T+365
+47.5%
FRALTA.PAXPAR
ALAIN TARAVELLA
Chairman
Filed
Mar 2024
Amount
31 k€
Sigma score
53 / 100
Return T+365
+45.5%
USNEOXNAS
Sherman Jeffrey Scott
Chief Financial Officer
Filed
May 2025
Amount
133 k€
Sigma score
63 / 100
Return T+365
+44.5%
USMTDRXNAS
Foran Joseph Wm
Chairman and CEO
Filed
May 2023
Amount
77 k€
Sigma score
52 / 100
Return T+365
+44.3%
USGTLSXNAS
Evanko Jillian C.
President and CEO
Filed
Mar 2025
Amount
49 k€
Sigma score
58 / 100
Return T+365
+43.2%
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.
Part 3.b · risk profile
The risk, unvarnished
Beyond the Sharpe / CAGR pair, here are the six risk metrics we now publish on the production V14e EU_strict strategy. They answer the questions a real investor actually asks: how much can I lose, for how long, and whether my upside volatility pays for my downside.
Annualised vol
17.0%
Max DD (monthly)
-44.7%
Time-under-water
48 mo
Calmar
0.11
Sortino
0.47
Sharpe (recap)
0.38
Definitions. Annualised vol = std-dev of monthly returns × √12. Max DD = worst peak-to-trough loss on the monthly equity curve (not daily, see caveat below). Time-under-water = consecutive months below a prior peak. Calmar = CAGR / |Max DD|, more meaningful than Sharpe for tail-risk strategies. Sortino = return / downside deviation; distinguishes upside vol (welcome) from downside vol (suffered).
Granularity. Our equity curve is rebalanced monthly (T+90 holds, month-over-month aggregation of net returns). A true intra-month Max DD computed on a daily curve would necessarily be more adverse than the displayed figure, typically 1.3 to 1.7× depending on vol. We document this honestly rather than hide the limitation. A daily NAV series is on the roadmap for Q1 2026.
Hit ratio by sector
Win rate and average T+90 return broken down by Company.sectorTag, on the Sigma v1.0 historical discovery cohort (pre-V14e). Highlights where alpha concentrated and surfaces sector concentrations capped at 30 % by the risk-management caps.
| Sector | N | Win % | Avg T+90 return |
|---|---|---|---|
| Finance & Banque | 634 | 51% | +1.6% |
| Santé & Pharma | 435 | 36% | -7.9% |
| Énergie | 259 | 50% | +4.9% |
| Industrie | 221 | 42% | +2.3% |
| Médias & Communication | 208 | 55% | +4.1% |
| Technologie | 201 | 32% | -9.1% |
| Distribution & Commerce | 135 | 49% | +3.3% |
| Luxe & Mode | 123 | 70% | +36.1% |
| Immobilier | 115 | 48% | +0.2% |
| Services aux entreprises | 111 | 35% | -5.3% |
| Agroalimentaire | 99 | 62% | -0.4% |
| Défense & Aérospatial | 85 | 56% | +7.5% |
| Chimie & Matériaux | 73 | 63% | +11.9% |
| Construction & BTP | 25 | 20% | -12.3% |
| Inconnu | 6 | 83% | +66.1% |
Part 4 · practical guide
Max 20 positions? Here's how to use Sigma
No, you do NOT need to hold all 585 companiesto extract value from the system. It's all about selection and discipline. Here is our recommendation for an investor with 20 positions maximum and capital between €10,000 and €100,000.
1
Use Sigma as a filter, not a strategy
The Sigma system is not a market-beating machine. It's a pre-screening tool: instead of manually scanning 585 companies, you have 3 to 10 strong signals per week to study. Sigma does 90% of the identification work; you do the remaining qualitative analysis.
2
Prioritise clusters
When ≥ 2 executives buy the same company within 30 days, the signal is more robust, the most consistent block in our backtests (+2.8% vs CAC +6.5%, in-sample, not demonstrated alpha: audit 169). An isolated buy can be noise; coordinated buys from multiple insiders signal real collective conviction.
3
Filter by role and amount
Trades by CEOs and CFOs are more predictive than those of board members or supervisory council members. Also prefer amounts ≥ €100,000, a €5k buy by a board member is symbolic; a €500k buy by a CEO is a material conviction.
4
Adopt a long horizon
Insider signals work better over 6-12 months than over 3 months. If you buy a Sigma position, plan to hold it for at least 6 months, otherwise transaction costs (0.6% round-trip in France) eat all the potential alpha.
5
Limit concentration
Even if a signal is very strong, no more than 8% of portfolio on any single position (per-name cap). European and emerging-market small-caps are illiquid and can drop 30-50% rapidly. Spreading across at least 13 positions keeps you under that cap, close to the diversification optimum.
6
Diversify with an ETF
Our honest recommendation:60% of your portfolio in a broad ETF (CAC 40 or MSCI World), 40% in Sigma convictions. The ETF core captures passive market performance. The 40% Sigma is your “edge”, and if it underperforms, you haven't lost everything.
Concrete example: with €50,000 of capital, allocate €30,000 to a CAC 40 ETF (expected return ≈ +6.5%/yr) and €20,000 across 4 to 5 Sigma convictions at €4,000-5,000 each. You keep the market base AND test your ability to select the right signals. If your picks beat the CAC over 2 years, increase your Sigma allocation.
Part 5 · transparency
Limits and Biases to Know
- Backtest ≠ future. Our +5.1% are historical over 4 years. Markets and insider behaviours evolve. The future can be better… or worse.
- Survivorship bias. Our database contains companies still listed. Delisted or bankrupt ones are not counted, which could overstate real performance by 0.5 to 1 point.
- Slippage not modelled. On illiquid small-caps (typically EU and emerging mid-caps), the bid-ask spread can reach 1-3%. Our backtests assume execution at closing price. In practice, you will pay the ask and sell at the bid → subtract an extra -1 to -2% from returns.
- Limiting liquidity. Some small-caps have daily volume < €100k. Putting €5k into a micro-cap signal can move the market. Sigma has no value on these stocks for investors > €500k.
- Short backtest horizon. 4 years is insufficient to judge an investment system. Academics typically require 10-15 years. We continuously enrich the database to reach 10 years by 2029.
- Favourable CAC period.2021-2025 was particularly bullish for the CAC 40 (ATH in 2024). In a bear cycle, our low-drawdown signals could catch up with the benchmark, but we don't yet have the data to prove it.
- Not investment advice. This site is an information tool. It is neither a robo-advisor nor a fund manager. Your decisions are yours; we are not responsible for losses.
Position management
Why sell at short / medium-term horizon?
Insider signals carry a strong but time-limited informational edge. Respecting that window maximises returns and limits exogenous risk.
Horizon
Recommended action
Why
T+30 d
Hold
Signal not yet priced by the market, too early to exit.
T+60-90 d
Take profits ★
Optimal window per backtests: peak alpha, highest win rate.
T+180 d
Reassess
Check if a new active signal reinforces the thesis; otherwise trim.
T+365 d+
Exit
Insider signal is spent; macro and sector noise now dominate.
Practical rule
Target an exit between T+60 and T+90 days after the filing publication date. If a new active signal on the same stock appears before T+90, reassess before selling.
The 4 key reasons
Information window is temporary
When an insider buys, they act on a private edge (earnings, contract, acquisition…). That information gets published within 30-90 days, the edge disappears after that.
Signal fades over time
Beyond T+365, macro, sector, and global market factors progressively overwhelm the insider signal. Measured alpha drops sharply after 180 days.
Opportunity cost
Staying invested too long ties up capital that stops working. Every week on hold is a week you can't deploy on the next strong signal.
Insiders exit too
Public filings from our 43 regulators (AMF, SEC, BaFin, SIX, RNS, ASX, EDINET, SEBI, CVM and more) show insiders are often tactical traders, short entry window, selling 6 to 18 months later for tax or personal reasons.
Monthly walk-forward refit
Weights and OOS Sharpe history
On the 1st of every month, the V14e ensemble is refit on the trailing 36-month window. The OOS Sharpe is measured on the following month with a bootstrap 95% CI. A Sharpe below 0 (amber badge) flags a drift to watch.
| Version | Fitted at | Sharpe OOS | CI95 | Months | CAGR | MaxDD | Deflated |
|---|---|---|---|---|---|---|---|
| V13.202607 | 01 Jul 2026 | · | · | 0 | · | · | · |
| 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.
Continuous improvement
What we're doing to improve
- Scoring v3 deployed (April 2026). Methodology overhaul after audit: cluster detection is now strictly directional, added insider track record (Bayesian shrinkage), DCA bonus and analyst-contrarian signal. Staleness penalty removed from the stored score (applied at display time only). Weight redistributed from public information (fundamentals, composite) toward insider-specific signals.
- Bayesian shrinkage on recommendations.Under-sampled buckets (role × size) are now pulled toward the global mean with a size-adaptive prior (k=30 if n<20, k=15 if 20≤n<200, k=5 if n≥200). No more structural over-ranking of low-n buckets on noise. The recency half-life was raised from 21 to 45 days (aligned with the 90-day hold).
- New "Sweet spot" bucket (€300M-€1B). Split from Mid to isolate the zone where the grid search found alpha concentrates (liquid, under-covered). Taxonomy now in 6 sizes: Micro / Small / Sweet / Mid / Large / Mega.
- Real retail returns stored. Each backtest now contains
returnFromPub30d/90d/365d(entry at pubDate+1) in addition to insider-view returns (transactionDate). Future backtests will be more honest by default. - Automated monthly backtest. Every Sunday at 5h UTC, we recompute 300 missing backtests. The system becomes more accurate each week.
- Historical expansion. Back-filling filings 2015-2020 in progress (currently available: 2005 → 2026).
- Sell signals. Our engine supports BUY and SELL, useful for position management on a held stock.
Use Sigma for what it is:
an analysis tool, not a magic wand
43 markets covered, 1,518,187 aggregated filings, 708,254 backtests. Use it to pre-filter your universe and focus your analysis on the 10-20 companies that truly deserve your attention this week.