I don't have any data on whether a multiplayer game at launch would be better or worse than a singleplayer game at launch. I share your bias about exciting and fresh experiences at launch over familiarity, though I also don't have any data on that. I guess we could argue the launches of the Wii, Wii U, and Switch and their flagship launch games (respectively the new and fresh Wii Sports, the familiar New Super Mario Bros. U, and the new and fresh The Legend of Zelda: Breath of the Wild) are data points here, but probably very biased, because one could argue for non-Nintendo platforms that Killzone: Shadow Fall, for example, wasn't really "new and fresh", but still a good-enough launch game for the PS4.
What I tried to prove was that multiplayer games during their launch period of a console aren't necessarily a bad idea. You say you're not convinced that Mario Kart and Super Smash Bros. appeal to the early adopter group. We can still argue about the exact definition of the early adopter group, but I'll go with "everyone who buys a console in its first year", and I'll extend that group to "everyone who buys a console unit shipped in the console's first year" for sake of my argument. In case of the Nintendo Switch, I define its early adopter group size as: all units shipped in CY 2017, and half of the units shipped in Q1 2018, which sums up to 16.34 million units. The install base of the Switch at the end of CY 2017 was 14.87 million.
I want to thank
@acezam for his extensive post with tons of charts with
quarterly Nintendo hardware and software shipments.
Other sources:
Supplementary Information for the Nine-Month Period Ended December 2017
Supplementary Information for the Nine-Month Period Ended December 2018
We have quarterly shipment data of Nintendo Switch hardware since launch, and Mario Kart 8 Deluxe was released only 2 months after launch. I have plotted the quarterly shipments of Mario Kart 8 Deluxe and Super Smash Bros. Ultimate (which was released in Q4 2018, at the very beginning of the 8th quarter of the Switch's life) against hardware shipments.
As you can see, at the end of 2017, Mario Kart 8 Deluxe's attach rate was 50% at a time where there was only an early adopter audience (7.33m copies). To compare (not in my graph), The Legend of Zelda: Breath of the Wild had shipped 6.70 million units as 31st Dec 2017, which means its attach rate was 45%, despite being a launch title with an attach rate of >100% in its first month.
Splatoon 2 had shipped 4.91 million units as of 31st Dec 2017, which is an attach rate of 33%, so lower than BOTW.
2018
At the end of CY 2018, the Switch install base was 32.27 million units. Based on my definition earlier, 16.34 million of those (50.6%) are early adopters.
Mario Kart 8 Deluxe's attach rate had only shrunk 3.4 percent-points to 46.54%.
The Legend of Zelda: Breath of the Wild LTD shipments were 11.68 million, corresponding to an attach rate of 36.2%.
Super Smash Bros. Ultimate had an attach rate of 37.43% at the end of 2018, despite being on the market for only 1 month(!!). Of course the game was the top choice to buy alongside a new console purchase during that Christmas season, but when 50.6% of the audience is early adopter, that's an achievement that simply is not possible if the game doesn't resonate with that audience.
Splatoon 2's attach rate had shrunk from 33% at the end of 2017 to 26% at the end of 2018, which shows it isn't as strong of an evergreen as the other games mentioned, but doesn't say much about the difference in appeal of that game between early adopters and non-early adopters.