And anyone nursing the belief that current ESA assessments are an honest, well-meaning attempt to assess disabled people but haven't quite hit the mark, I can disabuse you of that notion.
I have been filling EB/ESA forms in for 10 years. I have watched the questions change to capture less and less relevant information, by changing the tick boxes.
For example.
The question about climbing stairs used to offer answers from*
I cannot go up even one step
through gradations such as
I can use stairs but have to rest halfway
I can use stairs but in a special way, eg going up sideways
to
I have no difficulty using stairs.
In 2012, you still had to fill in the question - but the answers had been cut to only two:
I cannot go up two steps
I have no difficulty using stairs.
In 2013, this changed again. You still had to fill in the question and there were only two answers. But your answer wouldn't be counted unless you scored more on it than on the walking-on-level-ground Q with which it had been combined. (Oh, and the walking Q wasn't about walking anymore: using a wheelchair was deemed to be the same as walking.)
This isn't an oversight.
This is a deliberate reduction in the information collected and use of that info - while still forcing the claimant to jump through the same hoops.
Ironically, if you know enough, and have the energy to spend researching or guessing what might be deemed relevant, you can write large amounts in the "anything else you want to say" boxes. This has been shown to significantly increase the chances of getting an award.
But you have to know to do this and what will be relevant. So the new forms hugely disadvantage new and very ill claimants, while having slightly less impact on experienced, mentally competent ones. (Guess into which category the few actual fraudsters fall?)
All this, when the original forms just asked the questions outright in an easy-to-answer format.
*from memory: exact wording may differ.