License to Ill: How Atari's Biggest Arcade Licensing Coup Became Its Worst Adaptation

By Scott Stilphen

If the Atari VCS ever had a 'jump the shark' moment, this was it, for the VCS looked hopelessly obsolete immediately upon VCS Pac-Man's release.  It's Pac-Man, just not as you remembered it from the arcade.  Certainly being the most-anticipated cart release of 1982 helped programmer Tod Frye earn a reported $1 million in sales royalties, since we can pretty much rule out 'word-of-mouth' helping sales of what was soon considered to be not only a bad knock-off, but one of the worst arcade ports ever done (although him and Howard Scott Warshaw threatening to leave forced Atari's hand to finally make good on their oft-broken promises of paying royalties to programmers).  When both Frye and Bob Polaro were given the choice of porting either Defender or Pac-Man to the VCS, Polaro chose Defender, as he couldn't see how Pac-Man could be done, given the 4K ROM allocated for both, instead of the 8K previously used for Asteroids.  Pac-Man now has an eye and only faces left or right, plus he's constantly 'jawing' whether he's moving or not.  He literally looks and acts like he's been lobotomized... which isn't far from the truth.  Also, fruit was condensed into a single "multi-vitamin" (and only worth 100 points), dots ballooned into large "wafers", energizers became "power pills", etc.  The monsters also became similarly-colored flickering ghosts, making them very hard to see, and they're only worth 20, 40, 80, and 160 points.  The ghost moniker was probably Atari's Marketing department's effort of trying to explain away the terrible flickering (a year later, Namco started calling them ghosts as well.  See this article for more information.).  Their eyes also constantly spin around, as though they're already zonked out on pills themselves before the game even starts.  The maze looks more like Atari's Gotcha, or one of those cheap plastic maze toys, and is about as fun as walking through a long, winding queue line, or being a vendor at a stadium.  The tunnels are gone, replaced with exits on the top and bottom that, when entered, cause you to disappear from the screen for a second or two before reappearing on the other side.  Collision-detection is brutally unforgiving.  Gone is the generous overlap offered in the arcade version.  So much as brush up against a ghost here, and you become one yourself.  Eating wafers or power pills, however, now require you to completely pass over them in order to eat them, and it's quite possible to nearly cover a power pill and yet be killed by a ghost just around the corner from you.  The garish sounds (the "gobble" sound f/x is the same from Street Racer) and colors only added to the nightmare.  In other words, it looks, sounds, and plays like it was programmed by someone who never played the original.

The 4K was blamed by Frye for many liberties being taken, but subsequent homebrew efforts in the years since prove ROM size wasn't to blame.  Years later, Frye claimed he never asked for more ROM during the project but Frye's co-worker and office mate Rob Zdybel claimed as far back as the late 1990s in Warshaw's Once Upon Atari that Frye absolutely asked for 8K during the project.  In this June 2016 podcast, Zdybel briefly talks about Frye and mentions, "Tod's changed a lot over the years.  Tod's now denying stories that, I don't know, man... like he's saying he never asked for more than 4K for Pac-Man.  I know that's not true, Tod.  I was there when you did", to which Frye apparently replied, "That's okay, maybe I don't remember that one."

Atari licensed Namco's future arcade titles years earlier, in 1978 (for $1 million, which also included Galaxian), thanks to Joe Robbins, and given Frye started programming his version in mid-1981, "Pac-Mania" was already in full swing, so anyone would have assumed Atari would go all out and do their very best to translate this to the VCS.  Frye claims the arcade game wasn't a massive hit until the end of 1981 (somehow the fact Pac-Man was the world's most-popular game at the time eluded him...), but both Coleco (with Pac-Man) and Entex (with PacMan2) had their hand-held versions out later that year, and Magnavox released K.C. Munchkin (also 4K) for their Odyssey2 just in time for Christmas of 1981.  Regardless, what difference would that make in regards to making an accurate version of the game?  Popular or not, that was his job to do, and he failed miserably.  For someone who claims they spent their first 3 months at Atari playing VCS Space Invaders, you would have thought the reasons Rick Maurer's arcade port was such a success wouldn't have been lost on him... but they were.  He would actually go on to 'blame' that game for his poor color choices.

Masaya Nakamura and Joe Robbins renewing Atari's contract with Namco in 1978 for $1 million and rights to their future coin-op games.

LEFT: November 6th, 1981 ad for Odyssey2's K.C. Munchkin; RIGHT: March 18th, 1982 ad for Atari VCS Pac-Man.

When the game was released, and the critical responses started to mount, Frye says his superiors asked him about it, and he replied he could have done a better job with 8K.  According to him, Atari briefly considered releasing a revised version, but that likely would have involved recalling all the carts that were released - an even-bigger public relations disaster for sure.  Being a 'senior' VCS programmer, both he and Polaro could have asked for/demanded 8K ROMs be used, but apparently didn't.  To be fair, Atari could just as easily made 8K the new standard with their VCS games after seeing the sales success of Asteroids.  Cost certainly wasn't an issue for Atari at that point, but Atari (under Kassar) was very much in the mindset of minimizing cost/maximizing profit, even after their most-talented designers had already left to create Atari's competition.  The idea of spending more on development to make better games was never a priority.

Atari wildly over-produced Pac-Man, making 12 million carts, at a time when the installed user base was 7-10 million (the book Video Games by Daniel Cohen states on page 22, "Pac-Man is expected to sell nine million cartridges in 1982 alone.").  They were hoping for another "Space Invaders" effect of a game driving system sales to pick up the surplus, and it worked; nearly 8 million copies were sold at an astounding $37-$45 each, making it the best-selling VCS cartridge ever (and Atari ended up selling approximately 5 million consoles that year).  However, Atari never fully recovered from their decision to release it and the scathing criticism that followed.

Excerpt from May 1982 issue of Electronic Games (pg. 63).

The January 1983 issue of Softline had a contest (pg. 46) where readers could nominate which games were "dogs" or the worst for 1982.  Although the magazine was primarily for computer users, the contest didn't specify it was only for computer games.  The results were posted in the March 1983 issue (pg. 22,23) and both VCS Defender and Pac-Man were named as "Milk Bone munchers more often than expected."  As for Atari computer games, Frye's Asteroids made #6 on their top-10 worst list.  In the same issue, acclaimed programmer and designer Roger Keating was asked about doing arcade-style games instead of war strategy games, and commented the success of VCS Pac-Man was "amazing for a terribly sloppy piece of programming."

In hindsight, where was all the rigorous play-testing for this that nearly every other game underwent?  And why haven't earlier prototype versions been found by now?  Over the years, Frye has contradicted himself more than once, especially when it comes to specifics (see this Atari IO thread for more information).  A thread on Atariage tells a different, but yet more believable 2nd-hand account from Frye:

I asked Tod about this when I worked with him at 3DO. He wasn't as into reminiscing about Atari as Howard Scott Warshaw was, but I did coax some tid bits out from him about this. The details that I can remember are: What he showed to the execs was only a prototype to show what could be done theoretically. But the execs felt that with a little polish, the game could ship in a matter of weeks, and they'd all make tons of money. I think it was Tod's intent to sit down and remake the game correctly, but the execs didn't want to "waste time" when they thought what he showed them was playable.

Regarding the color, Atari marketing had a strict policy on avoiding black backgrounds for any game that didn't take place in space. They felt that color helped to sell a game because color TVs were only beginning to become mainstream around that time, and they wanted to hype up the fact that the VCS could produce color. So it's doubtful that marketing would have let Tod correct the colors if he wanted to.

And lastly, the bit about the ghosts flashing: Tod knew how to implement the now-famous horizontal interrupt that was later used in Ms. Pac-Man to cut down on the flashing. But his project manager was an ass and was very patronizing to him, so he left it out and submitted a final version that did not contain the fix. He was one of the first Atari programmers to get a huge royalty check under the new bonus incentives that Atari arranged to keep from losing programmers to the competition (Activision, Imagic...) so I don't think he lost too much sleep over the lack of accuracy.

Frye admitted in Once Upon Atari he developed a technique to better manage the flickering, but decided not to use it.  His attitude was, "If nobody else has done it before, why should I?"  In the Stella at 20 documentary, he also commented that he would have had a better flicker management routine had he had enough time to finish it (?), but he also claims he had a kernel working that used the same anti-flicker management system (using vertical separation and variable flicker) that Ms. Pac-Man used, but decided not to use it because nobody else had (?).  He also mentions "The only thing that could be said for it is, it had the largest number of arbitrarily-positioned objects of any 2600 game at the time, and that was achieved entirely through flicker."  Well, Adventure has the player, 3 dragons, and a bat, and Maze Craze has 2 players and up to 5 robbers, all with very little flickering, and all with distinctly-different colors.  If anyone should have been given the challenge to convert Pac-Man to the VCS, it should have been Rick Maurer.  His solution for using vertical separation to display all the invaders w/o flickering - and not being afraid to be the first to do so - is what makes him a true pioneer.

What's essential and what isn't?  In Stella at 20, Frye talked about the tradeoff of having a 2-player variation, since it used twice as much RAM.  A 2018 article in RetroGamer shows Frye's recollections are becoming more and more revisionistic.  Frye stating having a 2-player option was essential to capturing the 'spirit' of the original... but apparently nothing else was - not the colors, not the sounds, not the maze layout, not the different bonus fruit items, not the fact that Pac-Man can face all 4 directions (and had no eye), not the intermissions... nothing.  But those were "easy-to-fix" things according to him.  "I really honestly intended my Pac-Man to be as faithful a representation as I deemed possible and necessary."  No, it's essential if we're talking about games like Combat or Football - games that require 2 players, as they don't have a 1-player option.  Given the system only has 128 bytes of RAM, most designers would have dropped the 2-player option very early in the development process, realizing what the restrictions were with translating a 16K game like Pac-Man - something GCC's programmers were quick to drop with some of their arcade ports (Dig Dug, Galaxian, Jr. Pac-Man, Ms. Pac-Man, and Phoenix).  Although nobody else has ever mentioned having a 2-player variation was a design rule, every VCS game up to then (except educational games, Adventure, Night Driver, and Stellar Track) included them, as did Defender.  Sure, it was common practice by then, but certainly not a steadfast rule to be included.  Again, at the end, it was the programmers choice.  The same RetroGamer article is full of such "Frye-isms", such as the game utilized a flicker technique (does that mean Warren Robinett can make the same claim with Adventure?), and that writing code offered him a "godlike power to make things right" (actually, that's one Robinett can claim).

In a keynote from the 2015 Portland Retro Gaming Expo, Frye states he wish he had made a black background with a blue maze (as though that would have 'fixed' everything wrong with the game), but claims Atari had a rule against black backgrounds because it would have burned the maze into the CRT (apparently this rule didn't apply to space games...).  He made a similar claim in 2017:

At the time there was a firm guideline that games which were not set in space could not use a black background.  Not that we wanted colorful games - but bright colors on black backgrounds would 'burn in' much more quickly.  And the compliant employee that I was obeyed without question.

This is absurd since Atari touted the anti-burn-in effects of the VCS from day one, plus Frye included the color cycling code routine in the game!  Besides, any game - 'space' games or not - with a black background would burn in the screen if left on long enough.  I've asked several VCS programmers - including Frye's manager in early 1981, Dennis Koble - and nobody recalled there ever being a rule or even a guideline regarding the use of black backgrounds.  So Frye's unfounded claim appears to be nothing more than another excuse for his poor design choices.  Well, someone in Atari's Marketing department knew how the game should have looked, as early screen renderings showed the game having a black background.  Look at all the other coin-op ports that were done before Pac-Man.  Even b&w arcade games like Breakout and Space Invaders duplicated the colored overlays that were used.  If the game was purely b&w, like Pong or Asteroids, then sure, use color to your advantage.  But to take a game like Pac-Man and put a colored background in it, when part of the visual appeal of most games back then was to see colors against a black background is a poor choice.  VCS Breakout, Dodge 'Em, Haunted House, Super Breakout, and Video Pinball all have black backgrounds to name a few, and they're not space games.  So much for that 'rule'.  For him to say "Nobody knew what was important" is nonsense.  Well, gamers certainly knew.  The folks at Roklan Corp knew, since their version of Pac-Man for the 400/800 computers was released a few months later and is completely faithful to the original.  "No one knew.  History was being made.  We were just finding out what the rules were.  That's what it is to be a pioneer."  Well, which is it, Frye?  Were there rules that you 'had' to follow, or weren't there?  When doing an arcade port, there's only one rule - copy the original as closely as possible in every aspect.  That includes how it looks, how it sounds, and how it plays.  "If I have one regret (and I have many), it is the color scheme.  I truly wish I had gone for the blue maze on the black background.  IIRC (and perhaps I’d – likely more so than anyone else possibly can) I was influenced by the color scheme Space Invaders used, where the ‘variations’ used color schemes different from the arcade version."  Again, Maurer's game copied the colored overlays used in the arcade game, and it had a black background.  Aside from having different invader graphics, it looked and played just like the original.

Early catalog screenshot (LEFT); hacked version by Jason Parlee to match the early screenshot (CENTER); screenshot from box (RIGHT).

A video from the same PRGE shows Frye commenting on a new VCS version of Pac-Man and exclaiming he never understood why his version got so much criticism for having a different maze layout, and for having the tunnel exits on the top and bottom, instead of on the sides like the arcade version.  And that's the essence of why his games (especially his arcade conversions) really aren't anything special.  Todd wasn't a gamer, he was a programmer.  To him, making games was simply product for the company to sell, or a project to be completed, like making a deck.  You get some boards, you put some posts up, and you nail all the boards together.  To him, it probably didn't matter if it looked aesthetically pleasing or matched the generally-accepted design.  Pac-Man was a maze game with dots and tunnels and 4 enemies, so he made a maze game with dots and tunnels and 4 enemies.  In his mind, it was "Mission Accomplished".  Making the game look or sound even remotely close to the arcade version simply wasn't a priority of his, and yet... that was the first thing everybody noticed before they even played it.  And of course once they played it, they realized it had even less in common with the arcade game.  Nobody expected it to be as good as the latest homebrew version (link) but there's been several hacks and homebrews in the last 15+ years to prove a better version could have absolutely been done with only 4K (and they all have flickering to some degree, but it doesn't matter because the gameplay is there), so there's really no excuse for why it's so bad other than he was the wrong person for the job.

Ebivision's 1999 4K homebrew (LEFT); Dennis Debro's 2007 4K/8K homebrew (CENTER); Daniel "Dintar816"'s 2018 4K/8K homebrew (RIGHT).
This is what people expected in 1982.

Frye was hardly the only one to blame for the VCS Pac-Man fiasco.  In the summer of 1981, Pac-Man was the hottest game in the world.  Atari had fortuitously secured the rights to the game before it even existed, thanks to Joe Robbins.  Someone at Atari should have demanded they needed the by Christmas of 1981, and pulled out all the stops to make it happen (using an 8K ROM, having an artist and sound engineer, etc).  Instead, the game was never assigned to anybody, and it was left up to the programmers to decide who and when it was done?  Where were Frye's co-workers and supervisor and upper management to tell him he needed to make a better effort?  Where were the playtesters to tell Frye his game played terribly?  Where was Kassar to personally oversee what was the company's most valuable property in its history?  Everyone dropped the ball.  That game is a perfect example of how nobody was really in charge at Atari, and no better example of how detached Kassar was from the business at hand.

Frye is a very good technical programmer, but not one for making games that were interesting or having a lot of replay value.  He said it himself - he's much more interested in coding 2600 games than playing them, and it shows in all of his work.  The fact is, VCS Pac-Man is atrocious.  Most of his VCS games were either never released (Aquaventure, Save Mary) or finished (Ballblazer, Shooting Arcade, SwordQuest AirWorld, Xevious).  Even his Atari 400/800 Asteroids is clunky.  The one SwordQuest game he did finish and release (FireWorld) was a disaster, along with the whole contest.  The coloring schemes with most of these were just an eyesore, and all the design choices were his and his alone.  Sorry, but to spend 6 months (and 60-80 hours per week, as he claims, which works out to 1,440 - 1,920 hours) on a 4K game and have it look or sound nothing like the game it's based on is still just as unforgivable, even all these years later.  No, he was no pioneer with Pac-Man (for one thing, true pioneers don't use their project for a big-name title as a bargaining chip and threaten to leave before its completion, or refuse to implement flicker-management code out of spite for being put on probation) other than he was the first to do a truly horrible coin-op conversion.  He can continue to come up with all the excuses he wants, but the final result is, it's a terrible game, and no amount of drugs or gradient of rose-tinted glasses will ever change that.  It's not even a case of apples and oranges; we were promised an apple, and handed a lemon labeled as an apple.  To quote him one final time from that RetroGamer article: "You know what I say?  Fuck the press."  Know what *I* say, Tod?  Fuck you.  And while you're at it, you owe me $37.

Return to main menu