Yep. The thread is laughable.
Basically MSE or someone has highlighted that the CRA scores aren't exactly the same as what every financial company uses and somehow the internet has used that to mean they are useless and it keeps getting peddled on internet forums.
Barclays will for example know that X types of people with X1 and X2 credit histories end up defaulting Y1 and Y2 amounts. They can attempt to build a model of that for the many permutations.
The CRAs can do this for all companies. They have that exact same information from Barclays, but also HSBC, Natwest etc. They all upload that information to the CRA databases!
What these people will be surprised to hear is that the CRAs actually produce the best credit scores as they have the most information. Companies, know this and so don't inherently dismiss the CRA scores.
Smaller companies or companies that don't wish to invest heavily in analytics will in fact use these off the shelf scores.
The issue arises in that not all financial companies are looking for exactly the same things, and so any one size fits all score doesn't work. Certain factors may be weighted differently according to the needs of the company. Companies also have other information available to them that they don't necessarily reveal to CRAs (or CRAs aren't allowed to reveal to other companies).
This is where additional information can help and they have analytics teams to decide what is best for the company. Will it be massively different to the CRA scores? It will actually be very strongly correlated.
So why does the OP (and me) have vastly differing scores with the different Credit Agencies then? Given the exact same information, then surely a "good" score with one would mean a "good" score with all of them, rather than a "good" score with Experian and a "poor" score with Noddle?