The problems with your analysis is that you're trying to use the 3DS as a price comparison, while I'm using a Shield tablet, that's way closer in cost. 3DS is 5-year old tech (being generous, it's actually way more than that), so ir ends up being a bit overpriced because components are made exclusively for it. Who else uses a 240p glass-free 3D screen? Nobody. And a 720p 6.5' multitouch screen? Millions of tablets.

You are considering that the techs packed on the Switch are expensive, when they indeed are not. Basically, it just has a port to output video (tablets usually have HDMIs) and ports to slide the controllers. It's not like those costed 100 bucks.

The Tegra X1 is indeed close in performance to an Athlon 5150, that is a quad core version of the Jaguar used on PS4, so we are probably still dealing with a decent gap in the single components that is more bottlenecked on the PS4.

You pixel count still doesn't match. Tegra X1 is weaker than you think. It packs around 40% of the punch of a GTX 750, the weaker desktop Maxwell GPU (https://www.quora.com/How-do-modern-mobile-GPU-compare-to-desktop-ones). A GTX 750ti, that's a good equivalent for the GPU on the PS4 is 10% more powerful, so PS4 probably is 2.75X stronger while X1 is around 1.8X. The pixel increase from 720p to 900p is roughly 50%. In a very, very approximated math, we can assume that this isn't matching plain and simple: 80% more power pushing 50% more pixels, so it is closer to running a 900p game with X1 settings at 540p or less. We dould have to turn quite a bit of settings down to reach 720p. These benchmarks were made using the Tegra X1 on ARM boards or Shield console, so they can be more similar to "docked" performance than mobile one. My main worry is that it can end up being to weak to receive ports without some reengineering, instead of the current PC/X1/PS4 situation where you just tune some settings or the resolution and call it a day.

Regarding the Wii U, maybe if 3rd party games were successful, it would keep receving all games and the console would sell. Software sells hardware. You install base argument doesn't work when you consider that PS4 in its first few months had a smaller install base than the Wii U and 3rd parties still sold several times more copies than on the Wii U. The kind of consumer that buys Nintendo home consoles is the one that has other console or a PC os simply doesn't care about 3rd parties. Those are a minority, specially the first case that is for consumers with more cash. The single console owner are the majority and they gravitated towards Sony and MS to get 3rd party games.

This leaves Nintendo with two options:

- Create a 3rd party friendly console: risky, since it would have to match PS4/X1 in power and they would still have to assure that 3rd parties would support it instead of consoles with 25 or 50M units on the wild. They would have to basically pay for support and it could still backfire.

- Create a cheaper console, to be used by casuals or as second console to play Nintendo games: it has to be cheap. Less than 200 bucks. As the Switch is significantly weaker, they are going this route. If they combine it with a high price, it won't work.

1st Middle-ware makes x86 to ARM seemless? Either you know something I don't or that's one of the most incorrect statements I have seen on this thread. Can you link me to any resources that prove that point? Edit: just to clarify, I want some resources pointing to how seemless this code transition between x86 and ARM is, especially for game development

2nd - BF4 and Hardline, COD Ghosts, Dead Rising 3, Golf Club, MGS5, SW Battlefront, Quantum Break, BF1, FFXV (between 765 to 900p), Watch Dogs, COD BO3 (1280 x 900), Halo 5 (1152 x 810), Titanfall (792p)... at what resolutions exactly would such games run on a console around 3 times weaker and in a different architecture? (sources: IGN, Otakugame.fr, eurogamer and google in general)

 On your last comment regarding the PCs, they are also X86 architecture based. Actually the games are developed originally on PC and then ported to the respective consoles.

Adding to this, the memory bandwidth limitations on the Switch compared to the other 2 consoles will just make it a nightmare to use the same assets across platforms.

Basically what I'm saying is, this is a handheld and will need to have games designed for it specifically, which Nintendo will do. Some 3rd parties will probably try to port a few more recent games, but there's no way the big AAA games will make it there in any way. I just don't see it, I'm sorry.

1) These both focus on building ARM into x86, but they give the images and overview of building your game in the differing architectures. This is the entire point of using middleware, to remove the biggest pains when porting between difference systems. Its not zero-sum work, but it isn't as hard as some of you think.

https://software.intel.com/sites/default/files/managed/e4/d9/porting-guide-for-unity-game-on-intel-architecture-for-china-market.pdf
https://software.intel.com/en-us/articles/Unreal-Engine-4-with-x86-Support

2) All of those games are still using the same level of textures, AA and other visual settings but at lower resolution so that XO and PS4 are as similar as possible. What do you think is required once you drop textures or AA down? NS will be able to keep either high settings and lower resolution OR lower settings and higher resolution (likely XBO level).

3) Completely disagree with you on pretty much everything.

I think you're reaching SuperChunk. There's going to have to be some significant compromises for ports. The memory bandwidth alone is a big problem. 

I doubt NS is going to match these off-the-shelf configuration. After all, where did the 500 man-years of effort go to? Writing the vague announcement of support?

Probably a large chunk went to ensuring even 3 hour battery life. 

The Tegra X1 Maxwell chip eats energy like crazy. It runs at 20 watts when really being pushed ... that would kill even a massive battery in like 45 minutes. Likely Nintendo and Nvidia had to work a ton to get the chip to operate at far more efficiency for game related tasks than that. Also fine tuning it from an all-purpose chip to a game centric chip likely took a while, as did building the dev tools.

Likely Nintnedo also made a lot of specific requests like "absorbing" the Wii U architecture as much as possible likely to make it easier on their developers and allow for easy Wii U engine transfer. 

Every chip takes a while to develop even Wii U did. 

Where are you getting this information from?

Correction it's 10 watts for the GPU (this is still a lot for a mobile device), the Nvidia Shield console which has the Tegra X1 pushes 20 watts total as a system. 

Keep in mind though the Switch can't just power the GPU and nothing but. The CPU, WiFi, RAM, LCD screen among other components consume electricity too.

Treyarch had to build a completely new version, completely from the ground up, for the Wii. Using completely different assets and technology. This was the story with basically every HD "port" to the Wii, with the exception of sidescrollers like Rayman. It was basically a de-make.

Same would be required for Dark Souls 3, or any any game that was built to run exclusively on current gen hardware. The 360, PS3, and Wii U can't just "run" character models and environments using way more polygons, textures, and details than they were ever capable of running. They would require all new models and environments designed to run on them. This is why certain games like Assassin's Creed Unity were not released for Xbox 360 or PS3, which instead received their own Assassins Creed games released at the same time. Because a whole new game would basically need to be built whether it was a port or something else.

So it is very much a question of power as well as cost. If anything, the two go hand in hand: less powerful systems can be more costly to port to because a whole new game needs to be built, with all new assets and changes in design to accommodate a difference in power. More powerful systems that can receive simple ports of another system's games, same assets and everything, are less of an investment. But of course, it's still not always worth the investment even when a direct port is possible, but it at least makes it more likely.