I've written the accuracy review to `output/accuracy-note.md`, and the script's figures are in `output/accuracy.json`. The script confirmed that every closed month ties to the P&L actuals and that Budget equals the P&L budget.

**Revenue:** The 9+3 reforecast was the most accurate. It was off by $3K for the year, or 0.02%, with a MAPE of 0.63%.
- **Budget:** Over-forecast revenue by $362K, or 3.2%. Its worst month was 2025-07, with $1.0M forecast against $878K actual.
- **3+9 and 6+6:** Neither fixed the over-forecast. 3+9 still had a bias of 2.1%. 6+6 missed 2025-09 by $151K, or 15.0%.
- **9+3:** It had only three months left to forecast, so its accuracy doesn't show that the forecasting improved.

**Operating expenses:** Budget landed within ($649) for the year, or (0.03%), but its monthly MAPE was 4.1%. Offsetting misses hid the error. The reforecasts tightened steadily, to a MAPE of 0.30% by 9+3.

The data doesn't show why the revenue misses happened. To find out, ask the forecast owners what they assumed for 2025-07 and 2025-09.