I agree that 809/Kalydeco won't be on the market until mid-2015. As Qdelfan states, it will be a 6 or 12 month trial, but I think 12 months is more likely. That's what was required for Kalydeco approval. A trial looking at 12 months of efficacy per patient is going to take ~18 months assuming quick enrollment (likely, IMO). Filing the NDA and waiting for approval puts you somewhere mid-2015.
But that's where our agreement ends. Your Kalydeco estimates are definitely low. There are a significant number of G551D patients still receiving Kalydeco for free as part of a long-term follow up trial. When those patients start paying in the next few months, I'm estimating we get to $57 million/quarter. Then Europe starts to kick in. The number of patients is estimated to be the same as in the US, but let's say pricing is 50% of US levels. That's an extra $28 million per quarter, which should be in place for Q1 2013. Let's say Canada and Australia are on board too, worth $5 million. That gets you to $90 million/quarter, or $360 million in 2013 before any market expansion happens.
The R117H trial (3% of population, close to doubling G551D's 4%) finishes in July 2013. When those results are announced, there will likely be a huge ramp in off-label usage for those patients. Assuming they're worth 75% of the G551D revenues, that's an extra $43 million/quarter, just in the US. That would put Kalydeco at a ~$533 million/year run rate at the end of 2013.
Then there are your Incivek estimates, which extrapolate the recent decline in revenues as if the end destination is $0. That's completely unreasonable. Until the all-oral treatments get approved, there's going to be some steady state reached. I think it's unlikely we'll see more than another 20% drop from here, which puts the annual run-rate at $800 million. Combined with the Kalydeco estimates above, that puts 2013 almost flat with 2014. And as we've all more or less agreed, Kalydeco moves up from there in 2014.
On 809/Kalydeco you claim "one misstep with the combo and the company is gone". That's nonsense. They have solid proof of concept. The trial and its endpoints are straightforward, and approval is highly likely, unless major safety issues arise.
On VX-509, they may not have the resources to develop it fully alone. That just means they're likely to partner. A program like that with 2 phase II trials done will bring multiple 100's of millions in up-front cash plus development money, while retaining US rights.
On VX-787, "Vertex will stop development of VX 787 once the finish the current trial." Sure, maybe, if the results are bad. If they're impressive, Vertex will certainly either partner it or increase their burn rate to push it forward. I don't see how you can write it off completely based on speculation.
"you have to assume that every single drug works well in the clinic ... something unrealistic ..." No, you just have to assume a 25% success rate... something entirely realistic.
"Vertex has been very inefficient developing drugs from their own R&D, that is why I propose that they reduce it dramatically" Kalydeco, Incivek, VX-509, VX-787, VX-809, VX-661... inefficient? Looks pretty great to me. And cutting research completely would probably save a grand total of $50 million a year. That's a rounding error in their R&D spend. They're burning tons of cash because they aggressively moving multiple programs through the clinic. If you want to cut the burn, pick a drug to drop from development. Myself, I'm comfortable watching the burn and waiting for the payoff.