There's only one way you can avoid it, and I've put a LOT of thought into this subject, and here's your POD to avoid (or at least blunt) The Crash:
In November of 1979, Atari finally released it's two entries into the home computer market; the Atari 800 and the Atari 400.
Now, while the 800 was a decent enough computer, the 400 was nothing but a dressed up (and over-priced) game console.
Stands to reason though, as two of the chips used in the 400 and 800 computers (ANTIC and CTIA, the graphics and display chips) were actually being developed specifically for the 'perfected' Atari video game console, scheduled for a 1980 release. POKEY, the third new chip, was put together on the fly and handled both potentiometers and sound.
Now, while ANTIC (AlphaNumeric Television Interface Circuit), CTIA (Color Television Interface Adapter) and POKEY (POt KEYboard Integrated Circuit) weren't quite perfected (CTIA had some flaws to it and would, eventually, be replaced by an improved version GTIA and POKEY's wave table was quite limited, and ANTIC suffered from one flaw that's really only a flaw if it's a purpose built computer display chip in that it lacked 80 column capability) they WERE a fairly formidable package, capable of some remarkable feats for their time.
So much so that Warner Communications (Atari's parent company at the time) would have been MUCH better served to do this in 1979:
1. Design a dual purpose motherboard that would satisfy the requirements of both the computer guys AND the console guys.
2. Build the 800 as designed.
3. Rather than building 400s, make a console out of the 400 configured boards. (Complete with 6502 CPU, ANTIC, CTIA and POKEY, with 4-8 K of RAM, depending on how much they can put in a machine priced to sell at $200 dollars and four controller ports.)
4. Release both the 800 and...whatever they end up calling the new console in November of '79 (OTL when they released the 400/800) but with a little twist...
You see, Atari had acquired exclusive rights to produce a home version of the blockbuster arcade game of 1978: Taito's Space Invaders.
OTL, what few people realize, is that Space Invaders was the title that moved more 2600s than ANY title ever produced for the system. In fact, it helped move so many 2600s, it shackled Atari to a weak system for long past it's expiration date.
So, for your second POD, have Warner (in order to promote the new system) release Space Invaders as an exclusive title to the new console.
Basically, have Warner eat the loss on those final 250,000 or so unsold 2600s and go all in on the new system in 1979 and completely shift development and production to 800/New System in '78 and '79 for a fall of '79 launch.
Hell, if they abandon the 2600 by mid-'78, they might actually come out with even better hardware and a better overall system.
For one thing, if they use a dual purposed board for the 800 AND the new console, they can maintain compatibility between the two systems, meaning, they can give themselves a brilliant two pronged push into both the home computer market AND the console market, as total compatibility (uniform cartridge size and format) between the two systems give the consumer more flexibility at the point of purchase: The consumer who wants a computer but also wants to play video games has the option of buying an 800 that can run all Atari console games. The consumer who just wants a console gets a damn good console (light years better than the 2600 OR Intellivision).
So how does that impact The Crash?
While I don't know if it can kill ET, it puts a KILLER Pac Man in people's homes in time for Christmas 1981.
See, the 2600 just couldn't handle Pac Man. They TRIED to make a 'good' Pac Man for the 2600...and came to the realization that it simply couldn't be done. MEANWHILE, porting Pac Man for the Atari 8 bit computers and 5200 Super System, in OTL, was not only a cinch, the version ported for 8 bit/ 5200 stood as the undisputed BEST home port of the game until Tengen released a version for the NES in 1989.
So, at the very least, Pac Man's going to satisfy the throngs and make LOTS of dough for Atari in FY 1982. (Assuming that Atari begins working on a home port released for new system/800 in early to mid 1981 for either a Christmas 1981 release or spring 1982 release.)
THAT could possibly offset the ET flop. Think of it this way: Say Pac Man is driving sales of the 1979 Atari console, along with other Atari Games blockbusters like Tempest, Centipede, Defender, Space Invaders, Asteroids, Missile Command, etc... PLUS, a really slick looking (and playing) Raiders of The Lost Ark and other licensed games.
Further, sales are booming to satisfy the hunger for Pitfall! (released in 1982) and, a game that could have gone straight to console (had the console in question been a stripped down 400) called Miner 2049er.
Miner 2049er, by Big Five Software, was released for Atari 8 bit computers in 1982. (They would eventually release a crap version for the consoles of the day, but none of them could stack up to the original Atari 8 bit computer release.) If you're not familiar with the game, look it up, because it really was (for the time it was released) a magnificent game. 10 unique levels (that's 6 more than Donkey Kong) and a magnificent platformer that could have redefined the platformer in a manner similar to what Pitfall! would do...if had been available (in it's entirety and in it's 8 bit glory) for a home video game console.
Big bonus here is that Swordquest could yield four really good games and a franchise all it's own on a much more sophisticated platform, further offsetting the losses that ET would generate.
See, if Atari just goes with '400 as a console, rather than a cut rate computer in 1979' then that butterflys away at least TWO of The Four Great Failures That Doomed Atari in 1982:
1. The Atari 5200 Super System was a bust.
2. The Atari 1200XL home computer was a MEGA bust.
3. 2600 Pac Man was just putrid and left consumers feeling cheated.
4. ET was a rushed atrocity.
(The excess of Swordquest didn't help matters, but if the games are good, Atari could actually break even there.)
If Atari only has to recall the 1200XLs and swallow the losses of ET, it's dinged and FY 1983 ain't pretty, BUT, she's not killed by those two failures.
Just those two alone can't kill Atari and, frankly, I think the sophistication of games for a fully 8 bit Atari console that's been around since 1979 could nullify the glut of shitware that caused the other component of The Crash.
If the games being published by 'The Big Boys' (Atari Games, Activision, Imagic, Parker Bros.) and some of the better 'Up and Comers' (Epyx, Broderbund, Big Five, CBS Games) are setting the bar high (and sales on fire) the fly-by-nights and 'no business being in the damn industry' game developers don't even bother, because demand for the good product won't leave an inch of shelf space for the fly-by-nights and fad developers, thus, no over-saturation of the market and 'shelf-crowding' and thus, no Crash.
But that's just my opinion.