Timeform try to be objective and, as indicated by the name, time is a significant factor in that if a horse can get from A to B in a certain time (taking account, as much as possible, of all positive and negative factors) then that horse has proved beyond all doubt that it can get from A to B in that time. And generally there are other races to compare with, particularly on the same day at the same course. Other horse haven't got from A to B in that time, some because they haven't needed to; some because they can't. When there is enough form, it is probably relatively easy to identify those that can't. The problem is the ones that haven't because they haven't needed to. Just how much faster could they go if pushed? I imagine Timeform try to assess this by the ease of their victories. However, winning easily in a slow time does not mean they could go any faster over the full distance (eg if the race was run at a faster pace from outset that horse could be struggling and actually return a slower time than when it won easily).
It isn't an exact science which, by definition, leaves room for an element of subjectivity. When comparing horses of different generations the element of subjectivity becomes even greater. Hence there will always be disagreements about ratings.
You only have to look at Timeform ratings for major races (ie where one could be excused for thinking all have been trained for the race and will be trying), and then look at the results, to appreciate that Timeform ratings are far from infallible; and that's just across a single generation.
