Between 4 August and 14 September 2026 the live strategy book lost 246.5 signal R, or −89.9 R as independent bets. When that happens, a team needs to know whether the market changed or the system broke. NOVAQ's infrastructure made it possible to separate the two.
Attribution
Classic strategies were recomputed with production code on clean data. The waterfall reads: historical expectation +59.6, market path −214.7, residual −43.6, classic execution +22.0, trades outside the backtest +16.2, new legs and copies −86. The execution of the classic book was not the cause. The newly added legs were a material part of the loss.
Drawdown attribution, 4 August to 14 September 2026 (signal R)
Mechanism
Losses did not grow. Wins disappeared: the win rate fell from 49.6% to 37.1% with the stop rate almost unchanged, and an index of market favourability reached the 0.4th percentile of 2010 to 2025.
A market-level picture
| Measure | Norm | 1 Jun to 14 Sep |
|---|---|---|
| Share of NQ variance in the first 30 minutes of RTH | 19.9% | 27.1% |
| NQ beta to ES | 1.15 | 1.52 |
| Overnight share of NQ daily range | 58.5% | 69.9% |
| Days reaching neither ATR target | 47% | 58% |
The shortfall came mainly from a change in the mix of market states, rather than worse results within the same state.
Market structure: norm against June to mid-September 2026
Limits
The study is experimental. With an interval of −291 to +177 R at book level, edge decay can be neither proven nor excluded. The market table comes from a separate study. All figures are aggregated statistics.