My take on this is that you absolutely could break this out into smaller methods -you just refuse to do the work. It’s not really about code being called multiple times. As you’ve discovered it’s about the compiler and having too much on the stack. Really, who cares if you end up having many methods for the calculation?
I can argue, strongly, that shorter methods help in debugging and aiding in figuring out where exceptions occur. I’ve had to debug exceptions in very large methods before. A Nil Object Exception with hundreds of objects or an Out of Bounds Exception with dozens of arrays and/or loops makes it darned near impossible to find the cause.
I’ve literally made methods with names like this:
Calculation_Part1
Calculation_Part2
Calculation_Part3
This is harder if it’s a big overall loop but still not impossible.
Shorter methods make for easier debugging. Shorter methods make for easier documentation. Shorter methods lend themselves better to unit testing. Shorter methods mean the Code Editor stays fairly responsive. I’m sure there are other, additional benefits.