We’ve been meaning to post an article about what newspapers say Dominic Cummings said about HS2 on his £10-a-month-to-subscribe blog.
Luckily, the particular post in question – Systems politics, systems management– is currently free to read. So here is what Cummings said:
…Few public or private organisations (including No10 before we set up the A.P.O) now do the basics of data right therefore they are not able to deploy advanced tools such as machine learning or causal models.
Example… In January 2020, the PM was presented with ‘evidence’ showing exponential increase in HS2 demand — demand that, if taken seriously, would have meant the entire country either travelling on HS rail or building HS3/HS4/HS5… in ever shorter periods after ~2040. I.e a garbage model/graph was fed to a PM making a >£100 billion decision (which he blew despite us pointing out such absurdities in the ‘evidence’). Officials responsible were also responsible for dealing with exponential covid curves shortly after. E.g the original ‘single peak / herd immunity optimal strategy’ ignored the fact that once hospitals collapsed the jump in deaths would be nonlinear. We escaped total disaster partly because we had started to bring into No10 people who could reason quantitatively and talk technically with the modellers, cutting out the layers of management and powerpoint-creators that had obscured reality. We could then put accurate graphs in front of the PM and improve decisions.
We’ve long argued that the passenger growth figures presented by HS2 Ltd are nonsense. Changes to commuting patterns had already started before 2020, and currently rail usage is about two thirds of pre-pandemic levels. Far from concerns about passenger growth, train companies are now more worried about a lack of passengers.