Modeling from scratch help

$30 book by Staiger.  Goes through chapter by chapter.  I learned this way.  it seems cumbersome in the beginning, but I guarantee it is a great way to learn.

Once you've done enough practice models, you'll start to have your own preferences for formats and formulas. It should come with time and repetition. 

I would say just learn by doing. Take a set of cash flows you produced for appraisal and go model it out line by line. Use the cash flows you have to see if you are correct. Another way would be take the Accelerator models you have and re-build those models from scratch. 
 

Just a heads up, as you know, they will make you build from scratch to pass your modeling test but once you’re at the firm, there is very little which will be done from scratch. It’s inefficient. Most firms have templates for anything. And if there is some tweaking to a model, that’s common, but that’s about it. 

I don't want to minimize your frustration, as virtually every one of us has been in your shoes or still is, but here's my sincere advice: "Keep it simple". Ever heard of the "keep it simple stupid" principle? I'm glad you mentioned you did the REFM and ACRE accelerator courses. Because I can now tell you with confidence that you're adequately trained, but just lost in the sauce right now. All the shit about the preferred return hurdles and the expense recoveries and lease structures blah blah blah... take a step back, big picture.

What I mean is, when you're evaluating a property, don't start with the big fancy company pre-built waterfall model with all its bells and whistles. My working rule when I do underwritings (if time permits) is take a blank excel sheet and do "napkin math" on the deal. Pretending you're being interviewed is actually a great concept. Literally as simple as "ok i'm going to buy it at a 4.5% going-in cap, renovate spending roughly $3m, stabilize to a 6.5% yield on cost, and sell at a 5.75% cap rate" and determine no more than a handful of other big picture assumptions and you can produce a 5-year annual cashflow with a ballpark IRR and multiple.

Now, you know what the deal could look like from 30,000 feet in the air. You've produced a gut check and a quick vetting of whether the deal is even worth zooming into from 5,000 feet or whether the seller's guidance price is dog shit. Most importantly, you've developed an uncluttered view of the deal's major economics. With the right frame of mine, you can then get granular and more confidently underwrite the finer details. Buildings are living breathing businesses in and of themselves... step 1 should always be "do i understand the business model and can i articulate it in a few sentences and a napkin calculation?" then actually punching numbers into an actual model will be a piece of cake (generally... lol)