Strategy track record · 1 → 26 October 2026
Equity curves · indexed to 100 on 1 Oct (USD book returns)
Why RVOL is a flat line: its volume-spike gate closed on 3 Oct — the model holds zero exposure while the gate is off. A flat period is the strategy working as designed, not a missing month. The ensemble (ENS) halves RVOL's weight (h=0.5) and keeps TSI's risk capped, which is why ENS sits between the two.
Every observation
Account basis
Each book runs as a EUR 1,000 investor account: USD book returns converted at daily ECB EUR/USD (FX drift included and disclosed), all modeled friction (7 bps taker-in/out) and observed Binance funding already inside the equity path. Current FX 1 EUR ≈ USD. This is the same engine, panel and cost model as the audited backtest (2019→Sep 2026); forward days are appended only after the audited segment reproduces bit-for-bit (max abs diff ).
The separate 365-day v9 study
A stricter second study (fixed commitment, RFC3161 timestamp, tamper-evident public chain — live terminal) re-registers daily and has not yet completed its first registered fill; it publishes only evidence, no returns, until fills exist. 144.51% gate: UNMET there — by definition until the prospective year runs.
Prospective paper evaluation on live-collected market data (Binance USDT-M closes/funding, ECB FX). Not live trading; fills are modeled at daily closes with fixed friction; real executions, slippage regime shifts, capacity and liquidation are not certified. Sharpe annualized from 26 daily observations — a small sample, quoted for comparability, not significance. The +22.7% TSI day (26 Oct) is a single day; expect such swings to cut both ways. Past (audited or forward) performance does not predict future results. 144.51% gate: UNMET. · attempt & amendment log · full book dashboards