Something that worked for my INTJ brain is depicting all the factors in a spreadsheet.
YES :)
For a model with static relationships (income statement, for example), I also do years across and categories down. Even if there are lots of assumptions.
But like many financial calculations, this is more iterative and interactive, rather than a simple solve-for-x kind of thing. Of course if you fix all but a single item you *could* solve for x, but I don't think that's really what you want.
In such cases, I tend to put assumptions at the top, labeled so I can change them (say, spending levels, starting 'stache, tax rate(s), etc). Not saying this is "the" way to do it, just my habit ;) Then I do years going down.
Also in such cases I tend do several scenarios (that is, once I set up the model I copy it a few times over to the right and fill in the assumptions differently) - you could just change the assumptions without copying anything, but it's easier for me to be able to see/play with the impact of different assumptions (higher/lower spending, higher/lower/more volatile/less volatile assumed investment returns, etc.).
You can make it simple with as few as two columns (the year and your 'stache balance) - this means you use a single assumed rate of return, a static level of spending, same tax rate each year, ignore inflation. You could also make it quite complicated with many columns, then you can vary some or all of your assumptions year by year if you want to get crazy (withdrawals, spending, investment return, asset allocation, sun spot cycles ;) etc.). Taxes you may wish to do as column with a formula based on implied yearly investment income & capital gains (rather than a static assumption).
If you are torn between two approaches, try them both - as long as it doesn't take too long, I feel better if I can come to a similar conclusion via two separate paths!
*caveat* set a time limit, or the model can get more of your attention than actual FIRE ;) Be mindful & you can avoid analysis paralysis!