The Imperal College London model seems to have been predicting pretty crazy death rates for Covid, about 550% off for global deaths based on a 10 week prediction period based on one paper I read which was 2 orders of magnitude worse than nearly every other model the paper was comparing.
Does anyone with a clue about modelling have any ideas why this thing was so eager to predict the apocalypse?
I'm not an expert on epidemiology but I'm working on a masters thesis based on it atm so I've looked around the literature a little.
The biggest suspicion I have is based on the number of parameters the model uses. There are literally hundreds of free parameters and I don't know how accurate the estimates for these could possibly be. I haven't come across any other models that try to estimate so many factors that could affect spread, but as I said I'm not an expert. Is this a normal thing to do? Over fitting on this scale seems like too obvious an answer. I feel like respected experts should be this bad at parsimony without good reason.

Pic related, comparison between error in predicted cumulative mortality for Covid models from MIT (Delphi), Youyang Gu (YYG), the Los Alamos National Laboratory (LANL), Imperial College London (Imperial)the USC Data Science Lab (SIKJalpha), and the Institute for Health Metrics and Evaluation (IHME).