It seems his model was highly flawed, plenty of experts in software engineering that have reviewed the code are agreeing that it produces different outcomes using the same data on separate runs - not good science.
Obviously a model like this would use some sort of fuzzy logic but the disparities from run to run are huge.
https://tinyurl.com/yavxw4b8
https://www.telegraph.co.uk/technology/2020/05/16/neil-fergusons-imperial-model-could-devastating-software-mistake/