The additional problem with 'bolt on' keels is the concentration of
stress at the root of the keel and the mating structure. The Bavaria
line is the most stunning example of this stress anomaly wherin new
boats are losing their keels ... not as a failure of the keel nor its
attachment bolting but rather the mating FRG structure. This, to me,
is simply an unforseen 'concentration' of stresses which lead to a
very weak, prone to failure structure.
With an encapsulated keel system the lines of stress are allowed to
follow a 'more open' or less concentrated pathway over larger cross
section of the structure --- inherently safer but at a greater cost of
excess material/weight. My engineering eye usually is in an extreme
'wince' whenever I see a sharp inside corner anywhere near a
cantilever structure --- as thats the prime location of concentrated
lines of stress. Encapsulated keels usually always have 'smooth
transitions' in this critical area and thus avoid this 'stress
concentration'.

The real problem with critical stress design is that the 'calcs'
sometimes dont match with the actual alignment of stress, causing
unforseen 'stress risers' in the design that inherently weaken it.
Anytime the lines of force/stress come 'close together' (concentrated)
the basic values of ultimate tensile, etc. strength simply 'go out the
window'. All the 'strength values' have to be amended whenever there
is a 'discontinuity' in the surface of a structure ... as the lines of
stress are more located in outer surface of a 'shape'. The only way
to prevent such anomalies is to do (very expensive and very long term)
dynamic load testing to failure in a test rig -- as is done on
critical components (wing roots, etc.) on advanced aircraft, cranes,
gantrys, etc. etc. Obviously this extremely expensive process of
dynamically validating a design to actual failure is well beyond the
financial limits of a boat builder.