Your conclusion that BotW must sell 2.5 times that of Zelda on NES is completely nonsensical and is not backed up by your arguments. Your only numerical evidence for this is US population growth and Switch's worldwide appeal. The US population has only grown by a factor of 1.35 since 1986, and population growth would actually help Odyssey to a greater extent than it would help BotW because 3D Mario is more popular in the US than 3D Zelda. As for worldwide appeal, even if Switch becomes just as popular in Europe as it is in North America, that wouldn't let it reach 16.5 million.
4 million (Ocarina of Time, NA) * 1.35 (US population growth) * 2 (Europe turns into NA) + 1.7 million (Zelda NES, JP) + 1 million (Optimistic RoW) = 13.5 million
Of course, it doesn't even matter just how well Breath of the Wild does, because Odyssey is going to do better. In Japan, Odyssey is 700k ahead of BotW, which is an even greater lead than what Odyssey has in the US. The original Zelda on NES is still the best-selling Zelda game in Japan, and Odyssey is already ahead of that. You list Ocarina of Time as a previous Zelda game to have this kind of impact, but Super Mario 64 was still more popular than Ocarina of Time in every region. The US is BotW's best region, and Amazon shows that Odyssey is consistently outselling it there.
Finally, that Odyssey bundle sold nowhere near 3 million. If you look at the US Amazon charts, it was behind both of the $300 bundles in October, and then completely dropped out of the top 100 in November, never to return. It didn't even reach one third of Switch sales during its best month, so it is baffling that you think it accounted for more than 40% of Switch sales over the entire holiday quarter.