de facto, the choice of a lower autobrake setting may involve many assumptions, e.g. condition of the runway surface, brake/tyre condition, thrust reverse contribution to deceleration, or any additional safety margin for flying inaccuracies.
How and when is it possible to realise that the lower setting might be insufficient?
Then what, and will any change be sufficient at a late stage of the rollout.
.
Indeed the choice of Autobrake is dependent on many factors including runway condition.
In the case of a DRY runway,no wind,and a 13000 feet runway,the use of AB 2 would me more than sufficient for stopping distance at Max landing weight flaps up and with a sufficient margin from brake temp limits...have a look at the QRH for NORMAL landing distance with Flaps 40 and add approach speed adjustment of 55 kts,it will give you a good idea of runway you will need if other parameters such as wind,runway condition,touchdown point,reversers ....stay the same.
Have a look at brake cooling temp at a speed of vref40 max landing weight plus speed adjustement at touchdown.
By the way the AFM landing distances include a ONE second delay to account for Max manual braking application,not the case when using autobrakes,so to answer the OP,if after all calculations for landing distance with max braking and factors such rwy condition,wind,weight.....and you find out there is not One meter to spare at the end and you cant divert,well the Setting of MAX autobrake will give you a second of MAX auto brake lead on the AFM over going directly to Max manual braking.
In the case of dry and wet,the AFM distances account for normally worn out tires,however the Fully modulated Antiskid system does NOT need to be checked at Max landing weight for certification....so you are on your own there and hope the antiskid perform properly.