They needed Mario 64 that's for sure, but that year also allowed time for manufacturing more units and let them cut the launch price from $249 to $199. 

Rushed games, supply constraint and price were also all contributing factors in the Saturn's poor start and early troubles.

In what way? Yes, there are some legendary games on the N64 but, from a business standpoint, it got MURDERED. And when I say murdered, I mean it. It lost by 70 million units and is responsible for Nintendo's first WW lose in home consoles. It literally signifies the end of Nintendo's reign.

I don't think it turned out great at all.

I don't think it turned out great either.  Even so, if you're looking at it from a business standpoint the numbers that count the most are not the console sales numbers but the profit numbers.  And the N64 did make them money.  

Look at it this way,

Nintendo won with it's first two consoles and came in 2nd place with the N64 and made some money.  

Sony won with it's first two consoles and came in 2nd place with the PS3 and lost 3.5 Billion USD.

Technically that is what is getting murdered from a business standpoint.

SGI turned out great for the N64.

Cartridge media held it back...

The technology behind the Virtual Boy wasn't bad in itself. It was the way Nintendo executed it that made it fail. Nintendo made several mistakes that doomed the system even before its launch.

1. THE NAME VIRTUAL BOY
Having the word "Virtual" in its name created expectations that the system's technology never had the chance to live up to. A more suitable name would have included "3D" or "Hologram" instead of "Virtual".

2. IMPRACTICLE DESIGN
The design of the Virtual Boy, with its "legs", made it very impractible. Making it seem more like a prototype than a finished product. Nintendo should have spent an additional 3-6 months tweeking the design, trying to reduce size and weight so that you could have been able to wear the system on your head.

3. BAD GAMES
The Virtual Boy was released with too few games, and they didn't really show the performance of the system. The release of the system was rushed to steal attention from Sony PlayStation and Sega Saturn, and in retrospective that rush didn't reach its objectives at all. Nintendo could have waited an additional 6 months and released the system with far better and more deversed software.

I still believe the Virtual Boy could have been a commercial success with total life time sales of 8-12 million systems.

However, an even better solution would have been to scrap the system on the drawing board, and focused entirely on Nintendo 64 development. Back in 1994 and 1995 the Virtual Boy did draw resourses away from valuable Nintendo 64 software development.

I agree with most of what you said except for point number 1.

I think the name Virtual Boy is a really good marketing name.  

You can tell that it's a Nintendo product because it uses the "Boy" moniker popularized by the Game Boy. 

The word Virtual is used to tell you that it can give a physical sense to an image that is not through the illusion of depth, basically 3D.

And most the ads for the Virtual Boy had the words 3D in them.  

Virtual is broadly defined and though the word gets the imagination going that's exactly what makes good marketing.

I agree that from a marketing perspective the name is good, but it requires that the product delivers what the  marketing promises. I can remember head lines in magazines back in 1994 and 1995 saying things like "Virtual Boy is the worst Virtual Reality system you can imagine!" and Nintendo couldn't recover from that. This was back when people had seen Lawnmover Man, they had seen the early 25 000 dollar Virtual Reality systems, they had seen arcade games such as Virtua Racing and Virtua Fighter. And Virtual Boy was quickly associated to those products becuase of its name, and eventually also compared to them. A comparison Virtual Boy couldn't handle from a technical perspective. Even though I personally consider it far more entertaining to play Warioland and Mario Tennis than it was running around in the early true VR-systems :P 

I would also like to dispute the claim that Sega talked to SGI first because Nintendo already had a working relationship with SGI.  

sources - are linked where available

Here's a look at Rareware and Nintendo and their previous experience with Silicon Graphics (SGI) and how Sega fit in the timeline:

• Rare's Tim Stamper recently posted pictures of their first SGI machine (nicknamed VOSTOK) which was a SGI Power Series IRIS 4D/310VGX.

• The 4D/310VGX was released in the early part of 1991 with a price of 55,000 UKP (United Kingdom Pounds).
• Rare was assumed to have bought this machine sometime in 1991 or 1992
• According to David Wise, Rare could initially only invest in 2 SGI machines in the early 90's (the above 310VGX being the first)

• On Mar 13, 1992 - Silicon Graphics announced that it was buying MIPS
• On May 28, 1992 - At CES, Nintendo displayed a Real Time Mario generated on a Silicon Graphics 420 VGX Workstation

Real Time Mario demo at the 1992 CES

• The Mario demo was done on a Silicon Graphics Workstation courtesy of SGI and must've taken Nintendo some months to make.

This means that means that Nintendo and SGI had some sort of working relationship prior to SGI's buyout of MIPS in March.

• Sega's Tom Kalinske claimed that they were approached sometime after the buyout of MIPS by SGI, but is fuzzy on the timeline.

So, SGI may have been shopping the chip around but were almost certainly talking to Nintendo at the same time.

• In August 1993 Nintendo/Silicon Graphics announce details about Project Reality to the public, which was well under way.
• In parallel, sometime in 1992/93 Rare shows it's working demo of a boxing game with fully SGI rendered graphics to Nintendo.
• In March 1994 Nintendo announces that Rare would make games for the Project Reality video game system.
• In October 1994 Nintendo buys a 25% stake in the developer Rare (later increased to 49%)
• In November 1994 Donkey Kong Country is released

• According to the ex-head of software at Rare (Martin Hollis), Rare had also amassed around 50 SGI machines by 1998 and by contrast had around 3 in 1993 before Nintendo's investment.