I forgot about 2019, my bad. We have indeed chosen different weeks for 2017 and 2018 but it doesn't really matter in the end since these are still higher than 2020. 

Of course the low stock for the new consoles are a factor but I also strongly believe that Xbox but especially Playstation were not agressive enough with XB1/PS4 and that were potential for better sales.

Nontheless I'm not too concerned about their fate for next year, especially if they have stronger software to support the consoles.

What I do in a situation like this is I come up with a hypothesis that explains the data, but I try to be willing to change it in light of new information, especially since there are several possible explanations.  Usually I just say what I am guessing without explicitly saying it is a hypothesis though.

In this case, my guess (hypothesis) would be that there are shortages in most regions, especially for the Switch base model.  Of course I will be very eager to see Nintendo's quarterly report to see if it has any information that would strengthen or weaken that hypothesis.  The second most likely explanation is that adjustments need to be made.  We'll just have to wait and see.

Btw, thank you for doing all of this research.  It is very helpful.

I don't think we have any info on how PS5 is going to do, yet.
All we know right now is that PS5 has greater supply than PS4 did for the same period.
Once PS5 starts popping up on retail shelves in Western markets, we'll know that the supply pipe is delivering enough to satisfy the demand velocity.

Japan is going to lose millions of sales based on the data. But I admit, this is a deduction based on: either demand isn't high in Japan OR Sony isn't factoring Japan heavily into their plan for this generation; both lead to the same conclusion, lower sales. And again, that's only where things currently stand.

Anyway, once PS5s sit on retail shelves, as opposed to flying off immediately to satisfy demand, then actual sales can be more accurately predicted.

I think market forces are going to work against it in the same way market forces worked against 3DS, but probably not to that extent. ARM-based devices with greater usability, like the Switch, are only going to increase in demand. There's a bit of a tech revolution going on right now for the casual/home market (as opposed to the professional market), and X86 is on the losing end.

Yes, it would most likely extend NS life as well as DSi did with DS, and just like that the huge existing user base of the original model will keep most new games compatible with it, not fully using the higher power available except for higher framerate, better AA and all the automatic scaling up allowed by game engines and graphics drivers. This will keep on dropping the number of 3rd party home console targeted AAA games, favouring Ninty ones, and also keeping the typical Ninty target as little overlapping with competing ones as possible. Making power whores mad and in denial as usual, but doing a lot of good for Ninty as usual too. and after all doing good for Sony too, as the times of PS2 overwhelming victory are long over, and avoiding a direct clash witn Ninty is most clearly the right way to go, succeed and avoid disruption.

Have Switch's physical software sales beat 3DS's physical software sales? I'm curious to see how cartridge sales evolve over time, especially since Nintendo hasn't solved the cartridge size problem for 3rd parties who want big games in retail.

The more I read, the more I think they piss AAA 3rd parties off on purpose. 

Maybe the increase of Nintendo gamers buying digitally also impact the physical software sales. 

I imagine that's been the case, both as digital becomes more and more accepted and because bigger games end up as incomplete releases in cartridges anyway. If you have to download 90% of the game to play it anyway, what's the point? Nintendo should've really solved it by now, though I imagine 3rd parties really want to push digital as much as possible and don't mind releasing half assed physical versions of their titles.

Switch continues to pull impressive numbers. Looking forward to see PS5 and X|S when we have better stock.

1) I think it's simply due to low demand in Japan resulting in low allocation. I'm pretty sure traditional home consoles in Japan will be even more irrelevant this gen. I don't see PS5 crossing 5 million there unless it gets exclusivity to mainline Monster Hunter, Final Fantasy, and Dragon Quest titles (or some surprising new system seller which is unlikely).

2) Countless predictions have been dooming various consoles since forever. I'm not necessarily saying that's what you're doing here. And I'm not saying your prediction about X86 won't come true. But how do you think this "losing end" will impact the longevity and sales of PS5? Correct me if I'm wrong but isn't PS5 (with its estimated 4.5 million sales in 6 weeks) one of the fastest selling consoles of all time? PS5 in Japan will most likely be down but besides from that, how do you think the "market forces working against PS5" will impact it?

The thing is not that people here are in denial, but the fact that we really don't have enough data. Not saying that is impossible for it to happen, but we really can not have an idea of how high will the PS5 fly until stock is easily available and we check out how it can do on a normal week. Right now, what we can tell, is that there are a lot of challenges ahead. The competition is way tougher now than what it was at the beggining of the PS4's life, it still going to be a while until more strong exclusives get released and the stock could be a big problem in 2021 if it's not solved soon enough. Besides the elephant in the room that are the japanese sales.

Here's the thing, right now we have no idea on how this can or cannot affect the console. If it will do better, worse or equal as the PS4, we really don't have any data to answer this. We can only look at what is happening right now and what is seens that this is showing us.