They'll need a price cut soon if they plan to do this, sales are falling fast. I imagine Switch is cannibalizing sales a bit as well

While a price cut on the 3DS is always possible, I don't think it is necessary to keep it in production. Nintendo could be content to just keep making 1-2 millions per year for a few years if they wanted to. Nintendo is also not inclined to lose money on hardware as a loss-leader.

The main point I was trying to make is that Nintendo themselves have made it clear that they do not see the switch as being ideal for young kids as a portable, and they remain committed to the 3DS for the time being. At the same time, "young kids" is not a large enough group to create a 3DS successor for, so the switch should most likely not have to deal with a 3DS successor anytime soon.

Interesting that the NES and GB sales curves were left off, but they are two of the more unusual ones. The Famicom launched to a slow start in 83, but by 85 was going strong in Japan. Then NES launched in one city only in 85, didn't reach a nationwide release until 87 and then peaked in the US in 89. Putting the two markets together it probably peaked in either year 5 or 6.

The original Gameboy has more of a roller coaster sales curve where it was high the first few years, dipped down a lot for a couple of years and then shot back up again. The first 6 years wouldn't really tell the whole story for the Gameboy.

Basically what I am saying is that unconventional sales curves can happen, especially when you recognize the console is operating under unusual circumstances. I believe the Switch is operating under unusual circumstances, so it won't clearly fit any of these sales curves. It will go down a little bit in its second year, but then really take off like the DS sales curve in later years. I don't think it will quite get as high as 200m+ in only 6 years (like the table suggests), but that is only because I don't think Nintendo will make that many consoles. Instead Switch will be supply constrained during years 3-5. The number of consoles they sell will be limited by the number of consoles they make.

Interesting analysis. Had no idea so many Nintendo systems peaked their first year. That obviously won't be the case with the Switch. Just like the Wii it was supply constrained for much of its first year, though probably not quite as long as the Wii was. I think year 2 will probably be right around year 1 for the Switch. Then at the beginning of year 3 I think we'll see a price cut, plus Pokemon that year and probably a couple other major AAA Nintendo games after a slow second year software-wise. Plus third party games should start arriving in serious quality on the Switch in year 3. I think year 3-5 will be the peak years with major software and price cuts and revisions, perhaps doing 60 million during those 3 years. Then dying off the next 2 to 3 years as the next Nintendo system releases and takes over sales.

They could? I don't see why they'd even bother at that point...

I've left them out intentionally as they're both too extreme for any average calculations. They were on the market for donkey years and some of the other Nintendo systems were discontinued before these two reached their peaks, meaning I'd have go past year 10. Also both of them launched in the 80s, with little to no competition. On the top of it, it took almost 4 years for the NES to be available worldwide and Nintendo consolidated GameBoy and GameBoy Color figures, which reflected in absolut explosion of sales between years 10 and 12 (plus Pokemon in years 8-9), not only outliving NES, but also SNES and N64! I mean, I could do it, but I think we can all agree that the Switch won't be on the market for next 14 years and next GB "situation" is very unlikely to happen.

Edit: But hey ... for the sake of clarity:

Surprised that many Nintendo consoles seem to peak year1 or 2.

It looks that way going by OP, unless Im reading that wrong?

Which of those consoles is closest to todays 299$ in term of price?

tbh I dont think switch does 100m+ atm... so Id probably fall into Group 1.

*edit:  Did you make a mistake with the Total of group1?
You avg 5 numbers, all below 70.9m but end up with a result of 73.1m?

GRP1 total  = (49,2 + 52,5 + 69,5 + 70,9 + 55,7) / 5 =  59,56M.

*no idea how to delete tables.

I think we can make the general conclusion that Nintendo systems usually start selling faster and peak earlier, but it's not really enough to go on to make any remotely accurate prediction. I don't think looking at sales curves of the previous systems is enough to come to any conclusion other than it will probably peak earlier than systems such as the ps4, maybe in it's 3rd year.

From the data in the OP, there's no telling how high that peak will be, nor how steep or shallow the decline after will be. The data while an interesting experiment, in reality isn't worth much.