Useful posts from Slasher and GotTheTShirt. This has been my point all along.
The AFM data is quite fine - if the boundary conditions are constrained to approximate those pertaining to the certification tests.
In the majority of cases, overall, accel stop is not limiting and, depending on the extent of pad, we can afford to be a little relaxed. (Fortunately, with little aeroplanes, we usually get to use runways built for big aeroplanes)
In the real world limiting runway case, however, it is quite silly to presume that you will stop in a critical situation - historical data and simulator tests clearly suggest than the typical line pilot is going to be in an awkward situation with a high probability of a runway departure.
Hence the need for considered assessment by the crew on a takeoff by takeoff basis, with due consideration of AFM and OM guidance. I note again that Boeing (quite surprisingly to this observer considering the potential for consequent liability) recommends very strongly in training material that the go/no go decision is, in effect, made before V1.
I have no legal competence. However, the situation for the courtroom is that you had better have a very good and convincing story with answers for each and every argument presented by opposition mouthpieces. While I have survived to date, I have seen a number of very technically competent people made to look like idiots by tactically clever barristers - and this is precisely their job - to win.