There are a number of different models that are used by ROSCOs in structuring deals with TOCs / FOCs.
These range from straight "dry" leasing, where the original purchase cost is amortised over the expected (remaining) life of the train, and a rental produced that reflects the period of use relative to that life.
A "soggy" lease, where in addition to the purchase price, a cost is added to reflect large maintenance required during the train life (eg major overhauls / refurbishments/modifications, and the rental reflects the relevent portion of the purchase price and a contribution towards the maintenance events)
A "wet" lease is a package of train supply covering the full purchase and maintenance of the trains and the provison of guarantees of a certain train provision each day to cover set diagrams.
Effectively, you can also have any shade in between - e.g a "dry" lease, but could add some specific modifications into the lease that an opeartor requires. As the OP surmises, there have also been sale and leaseback deals, primarily with freight operators, that work just as he imagines.
On all leases, there is a cost of finance (interest) just as there is in any commercial deal in any sector of business. When ROSCOs have been owned by banks, they are likely to access lower finance costs.
All specific deals are commercially confidential, which is why you will not find too much detail on the internet, but Roger Ford in Modern Railways has, over the yerars, provided good detail on ROSCO principles, and worked mathematical models that are not a million miles from reality.
A ROSCO will always price over the expected life of the train, therefore taking a "residual" risk beyond the end of the current signed lease. A Section 54 provision is given (by DfT) to cover uncertainty in some cases where an initial lease is unlikely to be long enough to permit funders to take the residual risk at a sensible cost. It is NOT available on any of the original BR rolling stock and does not apply to all new stock (eg Chiltern 168s will not have had a Section 54, and a section 54 never normally covers more than half the expected life of a train. Therefore, a ROSCO will take considerable risk if a train turns out to be poor, or DfT policy leaves them with no use. That will be why Porterbrook will have gone to such extreme lengths to adapt the Gatwick Expresses, and Angel will have spent large sums trying to make 180s work.
Despite Goatboy's comment, BR stock was purchased off the Governement at a price, and the initial leases reflected a mixture of the value paid and the remaining life of those vehicles. Added to that will have been significant maintenance costs as all initial leases were "soggy" with the ROSCOs being responsible for all heavy maintenance. Had the "buy" price been higher, then so would the rentals. The rentals were also based on an "indifference pricing" model devised by the Government which was designed to ensure that TOCs were not disadvantaged by either having to have old or having to have new stock in the early years. It worked during that period very successfully.
Goatboy - An asset always has a value - high when new, reduced to the point of scrap value at the end of life, but you simply don't get given, or are able to rent a house for free, just because it is 80 years old, and the same applies to trains.
Unfortunately, because a few individuals made money out of ROSCOs, many people do have a very blinkered view of ROSCOs, missing the fact that there was no better way at the time to manage the existing fleets, and missing some of the excellent engineering development work done to the train fleets by ROSCOs over the years. Additionally in the first ten years, the investment in new trains by ROSCOs exceeded the income made by them from the lease costs of the existing trains, meaning as a financing vehicle, the principle was working well. The cost of buying a new train for a ROSCO was relatively low, as they were not bound by EC Procurement Rules, and could effectively negotaite and put a deal to bed in about nine months. I'm sure there are examples of where the ROSCOs weren't always effective, but as Roger Ford has said, they were the one true success of privatisation.
After that, DfT decided that it was jealous of the money made in the early years by some of the ROSCOs and hounded them for "repayment". When that didn't work, they sought external audits, reviews and competition committee enquiries to try to illustrate that the ROSCOs were making too much money. Each time they were rebuffed with independent evidence that the ROSCOs were NOT making unreasonable returns on their investments. Upset with this outcome, DfT effectively took procurement of rolling stock away from ROSCOs by setting up their own procurement operation with trains to be financed after procurement, and there we have IEP, Thameslink and others. Over £100,000,000 has been spent by DfT on train procurements that have dragged on for years, without so much a vehicle being built, and have produced deals that threaten to cripple the railways for years. Value for money? Not in my mind.
I leave forum members to decide which methodology they think best, and whether or not they think ROSCOs to be the ogre so many paint.