Hi Juramentado,
The question that you are asking is far deeper and more complicated than it might appear.
Regarding locks, reliabilty and strength are two very different questions.
As mentioned, reliability is more important and can be built into any lock design if the design is properly manufactured out of good materials.
Ultimate lock strength is essentially how much force will the lock resist before it yields to the force (breaks).
Most locks can be made stronger or weaker depending on the material strength and engineering geometry. When we create a model, we decide how much strength we seek. We break a proto, see what breaks, and stengthen the part that failed, break it again, increase strength of failed part, etc. until we reach the desired strength. The "weakest link" of the collection of parts that make up a lock will usually fail first.
No opinions involved. Our breaking machine is hooked up to a computer that measures ultimate force and inch / lbs per inch of force. Our standards listed below are "in-house" standards. the computer prints our graphs of the break. Current prodcution models are regularly broken to make sure "nothing has changed".
Heavy duty is 100 - 199 inch / lbs of force per inch of blade to break the lock. This means that a 4" blade (eg: Spyderco Endura) must resist over 400 inch / lbs of force before yielding, in order to be classified as heavy duty.
Medium duty is 50 -99 inch / lbs of force per inch of blade.
Light duty is 25 - 49 inch / lbs of force per inch of blade.
We also rate very heavy duty or what we call MBC (Martial Blade Craft) rating, which is over 200 inch / lbs. per inch of blade length (eg: A Spyderco Chinook is MBC rated. The blade is 3-3/4" long X 200 inch lbs per inch = 750+ inch / lbs of force to break the lock).
Fixed blades are usually stronger and more reliable against folding than folders (if made properly), but are often more difficult to carry due to their length. ("They call them "fixed blades" cuz they aint broke

)
SD is another very deep question with many factors involved, but that's another discussion.
sal