District forecasts one to three months ahead from the DGHS RMNCH Scorecard, scored out of sample against three naive baselines under a protocol committed before any scoring. A method counts only if it beats the best baseline with calibrated intervals. The data in the twin stop in August 2025, so the backtests and the as-of board are of past months; the one live part is the NMR nowcasts of months DGHS had not printed when they were logged.
Loading the health backtest…