I just reread the question, and spotted a rather important point (how I managed not to specifically point it out in my previous posts, I have no idea, it's rather obvious): you think that 75% of all cars are bad (ie. expected to be worth £1000) and 25% of all cars are good (ie. expected to be worth £5000). There is no sliding scale.
And, since a bad car for you is worth a maximum of £1000 and a good car for the seller is worth a minimum of £4000, it seems pretty unreasonable to me to expect the car to be worth £2000, as there are no cars worth £2000.
Of course, my example does not illustrate a real-world market for cars very well, however, that is not the intention of it. The intention is simply to reason logically based on the information I get in the exercise, at least that's how I choose to see it. Oh, and maths is terribly important, don't get me wrong, but I don't think an admissions tutor would feel too impressed when you place a simple calculation in front of them, and neither do I think that the points of these exercises (nor that of the interviews) is to show basic skills already acquired.