Skip to content

One method. Evidence tested market by market.

Coverage tells us where filings exist. Local evidence tells us whether a ranked signal can become an actionable recommendation. They are not the same claim.

2 Oct 2026
Public sourceRegulator or exchange filing
NormalizeRole, amount, currency and dates
Score 0-100Six publication-time factors
Local rankCompared only with its market
Evidence gateWatch or actionable
Test periods2023-2024 calibration · 2025 validation · 2026 Q1 shadow · T+90 net

A market needs two positive disjoint confidence bounds and sufficient price coverage before its signals become actionable recommendations.

See 43 covered markets and their public sources
FRAMF
USSEC (Form 4)
INSEBI (PIT + SAST)
KRDART (FSS)
DEBaFin
GBRNS (LSE)
CHSIX SER
CASEDI
AUASX
JPEDINET (Japan FSA)
ATFMA (OeKB OAM)
BALNasdaq Baltic
BEFSMA
BGBSE-Sofia (FSC)
BRCVM (Brazil)
CLCMF (Chile)
CNSSE / SZSE (CSRC)
CZCNB (Czech)
DKFinanstilsynet DK
EEFinantsinspektsioon (OAM)
ESCNMV
FIFIN-FSA
HKHKEX
HRHANFA / Zagreb SE
IDOJK / IDX
IEEuronext Dublin
ILTASE (ISA)
ITConsob
MYBursa Malaysia (SC)
NGNigerian Exchange (NGX)
NLAFM
NOOslo Børs
NZNZX
PHPSE
PKPSX (SECP)
PLKNF
ROASF / Bucharest SE
SATadawul (CMA)
SGSGX
THThai SEC (Form 59)
TRKAP / SPK (Borsa Istanbul)
TWTWSE / TPEx (MOPS)
ZAJSE
Σ · 0-100

How the ranking picks

  1. 1. Screen the universe
    Companies above 300 M€ market cap, insider trades under 1% of float.
  2. 2. Score eligible filings
    6 equal-weighted point-in-time signals, ranked cross-sectionally each month.
  3. 3. Hold the top names
    Top 10 per month, equal weight, held 90 days (T+90).
The 6 signals

The signals carry equal weights, none fitted on returns. The signal set and the screen were picked among tested variants, which biases any backtest upward.

0-100
Performance and transparency

We do not sell the ranking as a way to beat the market. Compare the corrected backtest with the index, then check the live track record.

Forward tracker: maturing, returns published from the threshold onwards.

See the full performance report
Corrected backtest · audit 231

What the ranking returned, next to the index

Sigma ranking
10.2%/yr
Sharpe 0.99 (CI95 -0.16 to 2.26)
STOXX Europe 600, same dates
9.9%/yr
Sharpe 0.65

The ranking clearly beats the pool of all insider purchases, but it does not demonstrably beat the index.

What Sigma puts together.

OpenInsider covers US Form 4 filings for free. Quiver sells a proprietary API. Bloomberg covers everything inside an enterprise terminal. Sigma combines multi-market filings, a public point-in-time methodology, a score broken down factor by factor, and a REST API plus an MCP server. Other multi-market trackers exist: see the full comparison.

CapabilitySigmaOpenInsiderQuiverBloomberg
US Form 4 (SEC)✓✓✓✓
Non-US filings (42 other covered markets)✓✗✗✓ (paywall)
FX-normalised cross-market✓✗✗✓
Composite score (open formula)✓✗proprietary✗
Free interactive backtest viewer✓✗paidpaid
Public REST API✓ (free tier)✗paidenterprise
MCP server (AI-native)✓✗✗✗
Entry price€0 / €39$0$25+/mo$24k/yr
Real examples

17 real captured signals, documented with their public filing and T+365 return (window Mar 2023 · Aug 2024).

See the case studies

Free account

Start reading the filings today

The free account opens the full read-only interface, no card required. Pro (€39/mo) is on an honest waitlist until payments open.

Create free accountSee pricing and the Pro waitlist