Guys seems to be forgetting what legs means...
If the game sold 1.2M in 2 weeks and them on the next 6 weeks it only sold 50k (not exact numbers) that is a definition of weak legs... sure it could go to sell 6M after 3 years and the guy would be wrong. But so far the only accusation of being wrong is because you don't like his conclusion.

No.  The accusation of being wrong is because the article falsely represents 2 weeks worth of sales as being 6 weeks worth of sales.  The report that gives those numbers only covers until the end of June, and the article presents it as including almost all of July.  The numbers for July, the numbers that would prove or disprove the point, don't exist yet.  It's like someone only having box office data for the first two weekends Wonder Woman was in theaters and concluding that the film flopped when he doesn't have the information on the rest of the movie's run.

That the numbers are wrong is obvious. But look at all the answers in the thread... Sorry to inform you most are saying it's a success for selling 1.2M in 2 weeks. Legs have nothing to do with how much it sold on launch, but how much post launch sales were.

The thread is pretty bad, but Kotaku is worse.

ARMS shipped 1.18m units in June, but that's obviously not sell-through which is the normal measurement to determine the strength of legs. Actual sell-through is much lower than the shipped numbers in the first two weeks, because retail stores want to have copies of a heavily advertised game on shelf, after all. That ARMS launched at the end of a fiscal quarter led to a bigger discrepancy between shipped and sold than usual. I estimate sell-through in June to be ~600k which is only half what was shipped. Most of the remaining stock should sell during this quarter, so LTD shipments will see a slight increase, because once again, retailers keep copies on shelves. The gap between shipments and sell-through will shrink by the time we get Nintendo's next update.

The only sell-through data for July comes from Japan where ARMS is doing decently. Everything else is guesswork at this point, but assuming bad legs when the only evidence contradicts such a conclusion is silly.

When most games don't reach 1 million sales it is a success for a new IP to do so on a system with an install base which is less than 5 million.  While it would be disturbing if sales were to drop off a cliff, let's not pretend that Arms was expected to do Zelda or Mario Kart numbers or to have a 50% attach ratio.  Even if Splatoon and Mario overshadow it, those 2 weeks have established Arms as a viable franchise going forward.

I don't think I can disagree with any of your points Rol.

Man... I'm not discussing this point. Sure a 1M sell on a new IP would usually mean it's successfull (we would need if it's making good profit or not to be sure).

But let's say new IP sells 5M on first month and 6M LTD after 3 years... probably a big success, but with weak legs.

Another game sell 1M first month and end up 3M LTD after 3 years... may not be a success, but had very good legs.

I'd consider both a success, though in different ways.  You're right that success and legs are not the same thing, though they're related.  Right now the only thing we have to go on with Arms is the initial succes since there is no information to even start talking about its legs.  Context matters.  A COD game selling 3 million units would be a flop given past success, expectations, being split over at least 3 platforms (XBOX, PS4, PC), and the likely high development and marketing budget for the game.  On the other hand Fire Emblem was considered a big success and a breakout hit when Awakening sold over 1 million after the series saw consistently declining sales.  Nintendo games have less people working on them and smaller budgets that most big AAA games like Mass Effect or GTA, and Arms would have had a smaller budget than other Nintendo games like BOTW and Mario Oddessey.  And it's on a system that still only has an install base of about 5 million.  It's hard to see how it could have possibly been expected to do better,