The density of the seat is only marginally relevant to the number of J seats that there will be on the plane. What is much more relevant is the mix of J vs Y (and vs F and PE) demand, because the square footage allocated to different cabins is not fixed. That's why UA has a high-J 767 with 46 J seats that take up over half the square footage for seats and used primarily for routes like EWR/ORD - LHR/ZRH/GVA, where there's disproportionately high J demand from business passengers with big travel budgets relative to Y demand from low travel budget business fliers / tourists / VFR. Or even more extreme, SQ has an A350-900 ULR with 67 J, 94 PE and 0 Y for SIN-LAX, SIN-JFK and SIN-EWR routes, both because the economics of the flight only work with high-yield passengers and because there's an enormous amount of premium demand on those routes willing to pay for the non-stop (enough that SIN-NYC went from 1x daily to 2x daily). Conversely, AF has a Caribbean configuration 777 with just 14 angle-flat J seats, 32 PE seats and 422 regular Y seats for Caribbean routes that are almost entirely filled with tourists and VFR passengers that are not willing to pay for J (and UA has something similar for Hawaii routes). The amount and proportion of seats in different cabins can be matched to the demand of the routes flown, and it's clear that DL on average perceives significantly less J demand on the routes that it flies than UA does on its own routes. The debate above is about why that's the case. Density differences will explain a couple seats more or less on the margins, but they do not explain a 50% difference in J seat count.