Research
We test what retail traders believe against 12 years of data — and publish what holds, what doesn't, and what the data actually says. Methodology, null results, and corrections included. Context, not signals.
Positioning brief
Where the crowd and futures positioning are stretched heading into the week — the setups worth watching, read for you.
Read it in the Member Briefs panel →COT report brief
The week's Commitment of Traders shifts across FX and the majors — what changed in dealer and large-spec positioning, and why it matters.
Read it in the Member Briefs panel →Monthly Context Review
The month in price, extremes, episodes and what followed them, regime and volatility — the full context layer, reviewed.
Read it in the Member Briefs panel →Published Papers
Full research papers, free to read on FXE Research.
Why the Data Is Different: Provenance, Methodology, and What We’re Building
SSI does not Granger-cause price in any of the seven major pairs tested. The founding piece: what 12 years of hourly positioning data can and cannot tell you.
The Anatomy of Retail FX Positioning: 12 Years of Hourly Crowd Behavior
Persistence, extremes, causal ordering, and structural evolution of retail positioning across 28 currency pairs.
Session Volatility in FX: Where Price Discovery Happens
The intraday volatility structure is remarkably stable — hourly profiles correlate r=0.987 across two decades.
Price Path Dynamics After Retail Positioning Extremes
75,864 extreme events against a random baseline: the directional fade is a coin flip. 91% still produce a 15+ pip move — extremes detect volatility, not direction.
The Complete FX Volatility Predictability Surface
How far ahead realized volatility can be predicted, pair by pair and horizon by horizon (AUC 0.69–0.77). The model behind the platform’s predicted-vol surface.
Public Research
5 foundational notes — open to everyone.
Price Path Dynamics After Retail Positioning Extremes
When the crowd reaches maximum conviction, does the contrarian reversal materialize? Forward excursion analysis of 75,864 SSI extreme events against a random baseline finds no systematic price path advantage. The null holds across z-score magnitudes, sessions, and eras. The only non-random signal is a modest directional asymmetry favoring crowd-short fades.
Volatility Regime Effects on Event Reliability
The finding this platform is built on: positioning extremes are primarily a volatility detector — 91% produce a 15+ pip move regardless of direction, while the directional read is near a coin flip (53.7%, corrected data). Moderate-volatility regimes produce the most reliable events.
Statistical Transparency: How We Report Our Numbers
Every number on this platform represents a deliberate choice. This report documents those choices — why we use medians over means, how we compute MFE:MAE ratios in different contexts, what we publish that most providers hide, and what common practices we intentionally avoid.
Crowd Positioning as a Contrarian Signal: What We Tested
The founding question: when retail positioning hits an extreme, does fading the crowd work? The original statistical framework across 28 pairs and 7 years of dual-timeframe crowd extremes — the starting point for the corrections that followed. Read alongside RN-004: the directional edge does not survive honest out-of-sample testing.
Anatomy of Unfavorable Outcomes: Loss Mechanics After Positioning Extremes
47% of events that ultimately resolved negatively reached +30 pips first. What the loss mechanics reveal about market microstructure: the move usually happens — direction and timing are the hard part.
Subscriber Research
6 additional research notes — market structure and context, plus the archive of claims we tested and corrected.
Access the Full Research Library
Subscribers get full access to all 11 research notes, raw data appendices, and monthly/quarterly intelligence updates.