I don't know whose "estimate" you are talking about, but I do take issue with your anthropomorphic assertion above, that any virus MUST trade pathogenicity for transmissibility. There is no guarantee of that, and as far as I know, no serious researcher is saying that there is.
This virus, like other highly lethal viruses, may gain the ability to transmit easily between humans and retain its deadly characteristics. What's one clue? Well, entire chicken flocks have died of it, in the thousands, seemingly overnight. Why didn't it trade pathogenicity for transmissibility in chickens or wild birds? After all, it's been around since 1996 or 1997. Are you saying it will only make that benign change when it happens to humans? I have never heard any report that it only kills one out of every ten chickens it infects. If that were the case, it wouldn't pose a problem and poultry growers in the US would have not have the present level of concern. They'd just let it sweep the flock, remove the 10% dead, and have 90% of their chickens now immune forever. Or intentionally give it to some chickens, and make a vaccine from the survivors to innoculate the rest. But as we know, that is not the case. And what can happen in chickens can happen in people. I am finished with this topic. If you wish to reassure yourself, be my guest. But on this, I disagree with you: I don't think there are any guarantees that it will happen the way you project. What you propose is a best case scenario, and that is not useful for planning, because of the obvious pitfalls.
I agree this is not a model I am familiar with, generally speaking the highest virulence is evinced in the index case generation. As any virus optimizes itself to a new host as is the case with influenza being zoonotic, it becomes less virulent in every generation of it's lifecycle. This is not a result of evoloving a new or enhanced biological transmission mechanism