A seller's package showed adjusted EBITDA of $1,349,772. The number that survived an independent review was $1,118,468. The gap was not a lie; every line tied to a document. It came from three habits of add-back schedules: a "one-time" expense that recurs in the bank data every year, an owner who works for free, and a related-party lease carried at market when it is not. Each is testable with material you already have, and each has a ten-minute habit behind it. At the 3.5x asking multiple, the three add up to $809,564 of purchase price. I took the deal apart in Issue 2 of The Workup Letter (composite deal, planted findings): redacted. The 1.25x calculator at getworkup.ai/tools has a slider that shows what a 17% haircut does to the loan, and there is a three-minute walkthrough of it here: redacted. Happy to answer questions on any of the three tests.