Arrow
Get your free stagnation score
Arrow
Find Your Solution
Logo

Atari’s E.T. Failure Was Priced In Before Anyone Wrote Code

by | Sep 10, 2026

FICTION · An Unfair Advantage Intervention

Fiction. Real history. Sourced below.

The programmer who supposedly made the worst game ever calculated that he wrote finished, debugged, documented code at thirteen times the industry average. He more than doubled the maximum output multiplier any human being can theoretically sustain — and it still didn’t matter. Because the schedule wasn’t set by what the work required. It was set by a licensing fee, and every failure that followed was priced in before anyone wrote a line of code.

I’m Maria Gonzalez. I came off a shop floor before I ever ran a process, so when people tell me the Atari story I notice that everybody talks about the game and almost nobody talks about the production decision that made five million cartridges. That’s the part I’d be fired for. Let me run this case the way I’d run a line audit — backwards from the failure to the decision that determined it — because the answer isn’t where the legend puts it.

Full fiction disclosure and sourcing note

This is a work of fiction. Maria Gonzalez is a fictional character from The Unfair Advantage by Todd Hagopian, and Cartwell is the book’s fictional company. Maria analyzes the real, documented history of Atari’s E.T. cartridge and the 1983 video game crash from the present day; she never met, advised, or spoke with Howard Scott Warshaw, Ray Kassar, or any other real person, and no dialogue is invented for them or anyone else. The only words attributed to real people are their genuine, publicly documented statements. Manufacturing and sales figures vary somewhat across sources and are given here as reported ranges. The historical facts are accurate and sourced as of 2026; the framework analysis is invented.

Run the Chain Backwards

Start at the landfill and work upstream, because the sequence is the finding. Atari buried inventory because it had millions of unsold cartridges. It had millions of unsold cartridges because it manufactured around five million. It manufactured five million because, having paid $20 to $25 million for the E.T. license, it needed to sell roughly four to five million cartridges simply to break even.

Keep going. Because five million cartridges had to move in a single Christmas season, the Christmas date became immovable. And because the licensing deal was not finalized until late July, an immovable December date left the developer five and a half weeks.

So the schedule that produced the famously bad game was not a scheduling decision. It was a consequence of a pricing decision made months earlier by people who never spoke to the developer. The license fee set the breakeven, the breakeven set the production run, the production run set the deadline, and the deadline set the development time.

That is what a line audit looks like when it goes well: you keep asking what determined the constraint until you reach a decision somebody actually made freely. Here that decision is the license fee, and every downstream failure was already inside it.

The Karelin Method: He Beat the Ceiling and It Didn’t Help

The Karelin Method multiplies Activity, Efficiency, and Focus: 1.20 × 1.20 × 4.0, for a theoretical output multiplier of 5.76x. Activity is 48 focused hours against 40. Efficiency is twenty percent more output per hour. Focus — the dominant term at 4.0 — is spending eighty percent of your time on the top twenty percent of activities. There is also a 50-Hour Sustainability Boundary, because the model describes what a person can hold, not what they can survive once.

Now measure what Atari demanded. Game development at the time ran several months; a six-to-nine-month cycle compressed into five and a half weeks is roughly a five to sevenfold compression, asked of one person, primarily by extending hours — which is the 1.20 term, the weakest lever in the model.

Here is what makes this case unusual. Warshaw calculated that he produced finished, debugged, and documented game code at thirteen times the industry average. Take that at face value and he did not fall short of the Karelin ceiling. He roughly doubled it. He delivered one of the largest documented personal output multipliers in the industry’s history.

And the product still failed, which is the entire lesson. Warshaw was a respected developer — he had built Raiders of the Lost Ark and Yars’ Revenge — and by his own measurement he performed at a level no framework predicts is sustainable. The constraint was never his throughput. It was scope: a deliverable sized for months, dropped into a container sized by a licensing fee, with essentially no time left for testing.

That is the correction the legend most needs. Karelin describes what a person can multiply. It does not describe what a mandate can compress. When leadership sets a schedule from a financial constraint rather than from the work, no personal multiplier closes the gap — and blaming the person who achieved thirteen times the industry average is not merely unfair, it points the post-mortem at the one part of the system that performed.

The Sentence That Explains Five Million Cartridges

Magnificent Obsessions is a customer and competitor intelligence discipline: end-user research, total-cost analysis, pain-point work, and a standing allocation of capacity to intelligence rather than a pre-launch box to tick. Atari’s leadership had a substitute for all of it, and stated it plainly.

Ray Kassar, Atari’s chief executive, is reported to have responded to internal concern about the game by saying that E.T. was such a hot property that whatever they put out would sell.

That is a demand forecast with no research behind it, delivered by the person who controlled the production run. It is Magnificent Obsessions inverted into a single sentence — brand certainty substituting for customer intelligence — and it is the most useful artifact in the whole case, because it is exactly what the failure sounds like from the inside. It does not sound reckless. It sounds like confidence earned by a long winning streak.

And the evidence against it was already on the books. Atari’s 1982 Pac-Man port had shipped into a market of roughly ten million Atari 2600 owners — and Atari manufactured twelve million cartridges. More cartridges than there were consoles in existence to play them. That production run was the warning, it was in the same building, it was months earlier, and E.T. repeated the pattern anyway.

Note what this does to the “worst game ever” framing. E.T. was not a sales failure in any ordinary sense: it sold 2.6 million copies by the end of 1982, which would be a strong season for almost any title. More than a quarter came back in 1983 as word spread, netting somewhere near 1.5 million against roughly five million made. A product can sell millions and still bankrupt you if you manufactured against a sentence rather than a forecast.

The Burial Was a Warehouse Clear-Out

The most mythologized part of this story is the most ordinary, and correcting it matters because the myth is what makes the case feel like it happened to someone else.

The legend is millions of E.T. cartridges buried in secret shame under poured concrete. What was actually buried in the Alamogordo landfill in September 1983 was roughly 728,000 cartridges of many different titles, as Atari cleared out its El Paso facility. The 2014 excavation — attended by Warshaw himself and by James Heller, the Atari logistics manager who ran the 1983 operation — recovered around 1,300 cartridges, of which only about a tenth were E.T. The rest were Pac-Man, Centipede, Defender, and others, along with consoles and computer parts.

The motive was mundane: disposing of excess, broken, and returned inventory, with a strong likelihood that writing it off served a tax purpose. The site was chosen because it prohibited scavenging and crushed its garbage nightly. Practical disposal, not a dramatic gesture.

Which is the point. This is what the end of an overproduction cycle actually looks like — not a cinematic act of shame, but a logistics manager, a landfill contract, and a write-off. It is unglamorous, it is repeatable, and it is happening in someone’s warehouse right now.

What Actually Happened

Warner Communications paid an estimated $20 to $25 million for the E.T. license in 1982, an unprecedented sum for a film tie-in, and Atari reportedly needed to sell four to five million cartridges to turn a profit. With the deal not finalized until late July and the Christmas season fixed, Howard Scott Warshaw was given roughly five and a half weeks to develop the game, against industry norms of several months, leaving little or no time for testing. Atari planned a $5 million advertising campaign, with a company executive publicly predicting the game would match or beat Pac-Man.

The game shipped in December 1982 and sold 2.6 million copies by year end. More than a quarter were returned in 1983 as word spread about frustrating gameplay, leaving net sales near 1.5 million against roughly five million manufactured.

In September 1983 Atari cleared its El Paso warehouse into a landfill in Alamogordo, New Mexico — approximately 728,000 cartridges across many titles, plus hardware. The event was dismissed as urban legend for three decades until a 2014 excavation confirmed it, recovering about 1,300 cartridges of which roughly ten percent were E.T. The recovered cartridges were auctioned to benefit local civic institutions, and several entered museum collections.

The broader crash was systemic: market saturation across competing consoles, a flood of low-quality third-party software, no industry quality standards, industry-wide overproduction, and competition from cheap home computers. United States console sales fell roughly 95 percent between 1983 and 1985, and industry revenue collapsed from about $3.2 billion to roughly $100 million. Warner sold Atari at a substantial loss.

From the Operator’s Chair

I’m Todd Hagopian, and overproduction, demand forecasting, and the danger of confusing a boom for a permanent condition are things I’ve managed directly across Berkshire Hathaway, Illinois Tool Works, Whirlpool, and now as VP of Product Strategy and Marketing at JBT Marel. Three disciplines out of this case.

First, trace every schedule back to the decision that set it. Nobody at Atari decided that five and a half weeks was the right amount of time to build a game. They decided to pay twenty-odd million for a license, and the schedule fell out of that arithmetic four steps later. I have seen this in manufacturing repeatedly: an impossible timeline that everyone treats as a scheduling problem is almost always a commercial decision wearing a calendar. If your team is being asked for something that cannot be done, find the decision upstream that determined it, because that is the only place it can be fixed.

Second, never let a personal multiplier substitute for a scope decision. Warshaw delivered roughly thirteen times industry-standard output and the product still failed. That is worth sitting with, because the reflex when a plan is tight is to look for more from the people doing the work. There is a real multiplier available in focus and method — meaningfully more than most organizations extract — and it is nowhere near large enough to absorb a scope error. When the gap between what is asked and what fits is larger than a few multiples, no amount of talent or hours closes it, and asking anyway guarantees you will blame the wrong person afterward.

Third, be suspicious of your own confidence exactly when it’s been earned. “It’s such a hot property, whatever we put out will sell” is not the sentence of a careless executive. It is the sentence of one who has been right for several years running. That is what makes it dangerous, and it is why demand validation has to be a standing process rather than a judgment call — because the moment your judgment feels most reliable is precisely the moment it stops being checked. Build to validated demand and take the conservative side of the error, because excess inventory is the one mistake you can see from orbit.

The Monday Morning Playbook

  • Trace impossible deadlines to their upstream decision. A schedule nobody can meet is usually a commercial commitment in disguise. Find the decision that set the constraint; the calendar is a symptom.
  • Size scope to the container before you assign it. If the gap between the work and the time is more than a few multiples, it is a scope problem, not an effort problem, and no personal output multiplier will close it.
  • Check your production run against your last one. Atari had already made twelve million cartridges for ten million consoles months earlier. The warning was in-house. Before committing a run, look at what your previous optimism actually returned.
  • Write down the demand assumption and who validated it. If the honest answer is “the brand is hot,” you have a sentence, not a forecast. Name the person accountable for the number and what evidence it rests on.
  • Post-mortem the mandate before the execution. When something fails, examine the constraints imposed before you examine the people who worked inside them, or you will fix the one part of the system that performed.

The Contrarian Reappraisal

The consensus reading — Atari made the worst game ever, buried millions of copies in shame, and crashed the industry — is wrong in nearly every particular. Three corrections.

It was a production failure, not a product failure. E.T. sold 2.6 million copies by the end of 1982. More than a quarter came back, netting around 1.5 million against roughly five million manufactured. And the five million was not free-floating optimism: a $20–25 million license had set a breakeven near four to five million units. The company had to produce that many to have any chance of profit, which means the license price, not the gameplay, was the load-bearing error.

The developer outperformed, by his own measurement dramatically. Warshaw calculated that he produced finished, debugged, documented code at thirteen times the industry average, in five and a half weeks, on a job sized for months, with almost no testing time. Scapegoating him points the post-mortem at the only component that exceeded specification.

The burial and the crash are both smaller and larger than the myth. Roughly 728,000 cartridges across many titles were disposed of in a routine warehouse clear-out, only about a tenth of the recovered ones being E.T. And the 1983 crash was systemic — saturation, shovelware, no quality standards, industry-wide overproduction, cheap home computers — driving revenue from about $3.2 billion to roughly $100 million. E.T. was the vivid symbol. Blaming it let everyone else skip the reckoning.

The Atari E.T. Timeline

  • March 1982 — Atari’s Pac-Man port ships; the company manufactures roughly twelve million cartridges against about ten million consoles in existence, leaving millions unsold.
  • Mid-1982 — Warner pays an estimated $20–25 million for the E.T. license; the breakeven requires selling roughly four to five million cartridges.
  • Late July 1982 — with the deal finally signed and Christmas fixed, Howard Scott Warshaw is given about five and a half weeks to build the game, against industry norms of several months.
  • November–December 1982 — a $5 million advertising campaign launches; the game ships and sells 2.6 million copies by year end.
  • 1983 — more than a quarter of copies are returned, netting near 1.5 million against roughly five million manufactured.
  • September 1983 — Atari clears its El Paso warehouse into an Alamogordo landfill: approximately 728,000 cartridges across many titles, plus hardware.
  • 1983–1985 — the systemic crash drives U.S. console sales down roughly 95% and industry revenue from about $3.2 billion to roughly $100 million; Warner sells Atari at a loss.
  • 2014 — an excavation confirms the long-dismissed legend, recovering about 1,300 cartridges of which roughly ten percent are E.T.

By the Numbers

  • 13x industry average — Warshaw’s own calculation of his finished, debugged, documented code output, against a Karelin theoretical ceiling of 5.76x. He beat the maximum and it still wasn’t enough.
  • 12 million cartridges / 10 million consoles — Atari’s 1982 Pac-Man production run against the installed base: more cartridges than machines to play them, months before E.T.
  • $20–25 million license → 4–5 million unit breakeven — the pricing decision that determined the production run, the deadline, and the development time.
  • 2.6 million sold, 25%+ returned, ~5 million made — a genuine hit by unit sales, converted into a catastrophe by the manufacturing decision.
  • ~728,000 cartridges, ~10% E.T. — what was actually buried: a mixed-title warehouse clear-out, not a shame-burial of millions.
  • $3.2 billion → ~$100 million — the systemic industry collapse E.T. symbolized but did not cause.

Frequently Asked Questions

Was E.T. really the worst-selling game ever?

No. It sold 2.6 million copies by the end of 1982, a strong season for any title. More than a quarter were returned in 1983 as word spread about the gameplay, leaving net sales near 1.5 million against roughly five million manufactured. The disaster was a production and forecasting failure, not a sales failure — a product can sell millions and still ruin you if you built far more than the market could absorb.

Why did Atari only give the E.T. developer five weeks?

Because the schedule was determined upstream. Warner paid an estimated $20–25 million for the license, which meant Atari needed to sell roughly four to five million cartridges to break even. Moving that volume required the Christmas season, and since the licensing deal was not finalized until late July, an immovable December date left about five and a half weeks — with essentially no time for testing.

Was the E.T. game the programmer’s fault?

No, and his own numbers make the case. Howard Scott Warshaw was a respected Atari developer who had built Raiders of the Lost Ark and Yars’ Revenge, and he calculated that he produced finished, debugged, documented code at thirteen times the industry average on E.T. The constraint was never his throughput; it was a deliverable sized for months dropped into a container sized by a licensing fee.

Did Atari really bury E.T. cartridges in the desert?

Yes, but not as the legend describes. In September 1983 Atari disposed of roughly 728,000 cartridges of many different titles, plus hardware, in an Alamogordo landfill while clearing out its El Paso facility. The 2014 excavation recovered about 1,300 cartridges, only around ten percent of them E.T. It was routine disposal of excess and returned inventory, likely serving a tax write-off, rather than an act of shame.

Did E.T. cause the 1983 video game crash?

No. The crash was systemic: market saturation across competing consoles, a flood of low-quality third-party software, no industry quality standards, overproduction across many titles, and competition from cheap home computers. U.S. console sales fell roughly 95 percent between 1983 and 1985 and industry revenue collapsed from about $3.2 billion to roughly $100 million. E.T. was the vivid symbol, not the cause.

About the Stagnation Assassin

Todd Hagopian is a Fortune 500 transformation executive who has generated $3B+ in shareholder value across Berkshire Hathaway, Illinois Tool Works, Whirlpool, and JBT Marel, where he serves as VP of Product Strategy and Marketing. Known as The Stagnation Assassin, he turns proprietary frameworks — including the Karelin Method and Magnificent Obsessions — into field-tested tools for operators who need to tell a scope problem from an effort problem before they blame the wrong person. He is the author of The Unfair Advantage and the Stagnation Assassin series, a Forbes contributor, and an award-winning keynote speaker. Connect with Todd on LinkedIn or bring him to your stage via his speaking page.

He delivered thirteen times the industry average and it still failed, because the schedule was set by a licensing fee four decisions upstream. When a deadline is impossible, the problem is never the person working inside it. Find the decision that set the constraint, because that’s the only place it can be fixed.

The discipline of telling a scope problem from an effort problem is in The Unfair Advantage.

Connect on LinkedIn · Book a keynote.