Frankly I feel this a question that no one has really answered.
The fact is that people keep saying that 1st party games can't run on the device, but how far is the Switch from it's fellow consoles anyway. How Snake Pass and Lego City Undercover ran on the Switch compared to the other two makes me question just how big the difference is.
The Switch is stronger than the Wii U with better archeticture after all, so I can't help but feel people are exaggerating the gap.
You cannot really compare the power of different devices. Especially with different architectures. If you put benchmarks on systems (which is possible on PC, but on consoles we cannot load own programs and the makers never tell us such stuff), you see that there are different stuff. Maybe integer operations run on one machine better than on the other, but for floating point it is different. Maybe one benchmark performs badly, but if you reorder commands it goes much better. Basically you have areas like: integer, floating point, memory bandwith, transfer data to graphic memory, shaders, ... In each of the areas the performance can vary from other areas, so machine A performs better on one area, machine B on the other. But seeing how much difference it even makes if you compile a code with one compiler and then with another - you can always say this benchmark is putting one machine at disadvantage. So don't expect 'exactly'.