I wanted to see how the Switch would have to perform for Nintendo to break their curse so I threw numbers in a graph. For obvious reasons I can't post them so if some user can just dump them in here I'd sure appreciate it.
http:// s000.tinyupload.com/index.php?file_id=24099414769387421219
http:// s000.tinyupload.com/index.php?file_id=90746372862243661930
http:// s000.tinyupload.com/index.php?file_id=41391770279179166440
Darker shades represent hardware and lighter shades represent software. Trend curves ignore the Wii and DS sales on the assumption that they were one-offs, and those curves are then extrapolated to get the numbers for the Switch. The first two plots shows what the trend says the Switch and an assumed new pure handheld launched in 2019 should sell. The third just stacks the two previous ones for the scenario where the Switch represents a new unified format and there will be no more handhelds. While the plot for the home consoles is fairly representative the one for handhelds has two caveats.
The first is it ignores the time each system was the top model. The GB/GBC numbers are for arguably two system that were the top models for about 12 years, whereas the GBA didn't even get 4 years before the DS jumped in and became the hot shit.
The second is that the 3DS numbers are the current ones and not estimates for what its final tally will be. Right now it looks like it will equal the GBA before it is discontinued, which would give a trend line that has started to go back up, but since that would essentially make the graph the opposite of what I think it should be I didn't bother.
So what would be more accurate? I was making some more graphs and putting some weighting on the numbers to compensate for the things listed above but the results shifted greatly even with light weightings so I decided it's not worth doing with rough numbers. The quick takeaway is that the GBA seems to have had stronger sales than the GB and GBC combined, but had its life cut short by the DS leading to a low total. Same goes for the GBC in comparison to the GB, given how soon after the GBA was released and collapsed its sales. Compensating for this would give the overall trend the shape of an arc, peaking with the GBA and the starting to fall, suggesting a new handheld would sell much worse than the 3DS rather than about the same or better than the 3DS as the simplistic approach says. It's interesting enough that I'm going to do it properly but that will naturally take some time.
TL;DR: Just look at the numbers in the graphs and say what you think will actually happen.
Posting some guts because I need to embed something.