Hardness is by definition extremely closely related to strength; resistance to deformation. The trouble comes when people try to use a single term to convey something that is the result of multiple properties. In the quench/temper processes in steels minimizing temper keeps hardness and strength higher. What is lost is toughness (area under the curve in a stress/strain plot). In fact, with no temper at all after a quenched thru-hardening, steel in the martensite phase would produce the highest possible yield strength IF you were able to get a pure uniaxial tensile load on it. The reality of the situation, however, is that it's basically glass, and any uneven loading will cause it to crack because there is almost no yield before break to 'soak up' the uneven loading.
Toughness, on the other hand, is increased by the balance between the height and width of the stress/strain curve. Enough ductility to have some give while retaining as high a yield/tensile strength as possible. The Charpy test measures toughness. Tempering in steel allows a percentage of the martensite to reform into pearlite, which is much softer and weaker, but has ductility. The matrix of pearlite/martensite is what makes steel very tough, and there is a very wide band of physical properties that can be attained with the same material based purely on the heat treat. 4140, for example can have yield strengths in the 60-80ksi range all the way up to 200+ksi.
Another big contributor is fatigue life. The harder you make steel, the higher the ductile-to-brittle-transition temperature is. We live in a relatively narrow band of temperatures and most hardened steels qualify as "brittle" at those temps. Brittle metals fail by fracture and crack propagation. Fatigue failure can happen in parts that are cyclically loaded, even if the loading is UNDER the yield strength of the material. Luckily most steels have a fatigue limit-- a stress level much lower than the yield strength that if you never exceed, the part will never fail by fatigue. If I had to guess I'd say most AR bolts fail by fatigue (micro-crack propagation) more than simply over-loading and outright failing the material. Hence the reason why even if you don't see pressure signs on the cases you shouldn't exceed book loads in PPC variants, and why different bolt designs/materials have popped up.