AFAICT the model seems fitted to historical data, and fit well. I'd be curious to see if they or anyone else kept tracing events using the model to check its predictive ability.