First g computation sections

[tgerke]

The broader chapter still in progress, but would be good to get some feedback on this presentation ordering (i.e. a preference to show Monte Carlo first, with the "data stacking" technique introduced later as a special case)

[malcolmbarrett]

Ok, I think I'm on board with this because I think I see the pedagogical advantage. Maybe the data stacking can even just be a longer callout box, e.g. for these simple cases you can use this one weird trick

I think we should address the question of sample size, like how does the sample from the model relate to the number for the monte carlo and the ultimate variance, e.g. the one from the bootstrap or whatever? I think this might relate to how g-comp can technically be more efficient while still being accurate but I don't remember the nitty gritty of our conversation about this

[tgerke]

Cool, good call re: sample size. I was planning to be more intentional about this is the subsequent section that uses bootstrap for CIs etc. Will remember to circle back to the initial section to pick a resample size that's more deliberate. I have a ref for this somewhere, stay tuned :)

[malcolmbarrett]

I didn't dig into the writing as much but overall very happy with the code pattern, so I think you can continue on!