Is placing a fire safe inside a larger fire safe helpful? If I put a small fire safe (say, one rated for 1 hour at 1700°F) inside a larger, equivalently-rated fire safe, to what degree is the fire/heat protection improved?
 A: The short answer is: yes it's helpful but it's hard to quantify exactly how improved the protection would be.
One problem is the idea of a "fire rating" or a "fire-resistance level". These are typically given in minutes. What this means is that the product passed a standardised test which exposes it in some way to a standardised time-temperature relationship for a certain period of time.
Failure criteria are different for different standard tests. An example of a failure criteria may be "ignites with flames for more than 5 seconds" or "temperature on unexposed side of more than 400 degrees above ambient". If this doesn't happen within 20, 30, 60 or however many minutes, this is the "fire rating" stamped on the product.
One things to bare in mind that this is performance to a very specific time-temperature curve and a very specific heating regime. It may or may not (most likely not) be representative of the heating regime your boxes may experience if there is a fire in your house.
An example of a standardised test is the use of a furnace. Propane is pumped into the furnace and burnt. The assembly is within this furnace. The thermal impact upon the assembly could actually be different from furnace to furnace. Only the gas temperature is controlled. For example, thermal irradiation could be much higher if the furnace walls were close to the assembly or made of a more reflective material.
So you could ask two questions in fact: (1) how would two fire safes, one inside the other, perform when tested to the standard? or (2) how much better is the protection provided to items in two Russian dolled fire safes compared to one when exposed to a real house fire?
For (1); you would have to test them. It has been demonstrated time and time again that you cannot just add fire ratings. Assemblies just don't work like that. Issues related to pre-heating, chemical changes, release of pyrolysates, dehydration all mean that two "one hour" boxes won't gain a "two hour" rating. In fact, sometimes it may be more, sometimes less.
For (2); it's even more uncertain. That's why standardised fire tests were invented in the first place! To predict what sort of fire your fire would be, where it would start, what the ventilation would be, what the fuel would be is a massive undertaking and currently we just don't have good quality data to do this. You could definitely carry out calculations for the response of a certain fire. But how do you know this would be your fire (the answer in practice is, you try to pick the worst credible fire)?
