The odyssey of the reclaimed NK-33's is intriguing as hell, but I can only imagine that everyone involved in their resurrection knew the odds of failure would be higher than with modern engines. Soviet metallurgy in the 1970's likely didn't have the consistency and precision we have today. So while the NK-33 was (is) a remarkably advanced design, it was likely pushing the limits of the materials of its time and place–even before sitting in storage in uncontrolled conditions for decades.

Odds of failure would perhaps be higher, given all the rest equal. Which is not - Soviet rocket engine school was at its golden age at the time, there were a lot of people both academically educated and who participated in large scale experiments with engines. Granted, today's capabilities allow us to surpass the level of sophistication achieved during the space race. However, we still need to try hard enough - and we know the support from governments is not as great.

Limits of materials were pushed indeed. Some safety margins for materials were claimed as under 1.0 - i.e., 0.97 - that is, materials were pushed further than their limits. The idea was that since the engine is used once, we can use it in the manner that greatly reduces the life of the engine - and nobody would care, since a stage will be done with in a few short minutes. A lot of things were done quite unconventionally - remember SR-71, for example, or oxygen-rich pre-burners which were considered rather impossible.

NK-33 remains unsurpassed even today by many criteria. Now it includes even the ability to work after 40 years of shelf line - which probably left the trace on engines reliability.

>with tests like throwing metal chips (nuts?) into pumps

SpaceX did this test with the Merlin engines too. http://www.airspacemag.com/space/is-spacex-changing-the-rock...

The chamber pressure (and therefore Isp) are both dialed back for the sake of reusability. They're stated multiple times that they're not running those engines anywhere near their physical limits.

It's certainly good that SpaceX uses many ways to test hardware, including those approaches successfully used before. During development of F-1 some interesting debugging techniques were invented, some could be well applied in SpaceX.

Gas generator cycle in general doesn't allow reaching effective Isp as high as closed loop cycle. On the other hand, smaller pressure reduces stress and generally considered safer. NK-15 were single use engines, but NK-33 and most other modern engines are all reusable.