I like data. Recently I used a speedo app that outputs GPS speeds to a spreadsheet. Then I make a specific run around my neighborhood, it's a few long flats from a stop. Usually also tryn time myself from a stop to a specific landmark. make small changes. +/- deg timing or +/- jet size, needle clip etc. plot em all overlapping and you start to see hey I got better pickup but crappier top. i'll try keep the needle and see if I can up the main or whatever.
But to go from experimentally validated trial and error to theoretical models, and just assuming away the lots of finicky little real world variances, I have a hard tie getting on board. sure as a starting point but not without validating w/ some tweaks
I also have a combustion analyzer, you can see when things are way off by the CO or off-stoich O2 but its not super responsive