The Failed Promise of the VCS/2600 Trak-Ball Controller

By Scott Stilphen


The trackball controller was truly a wonder when it first arrived in the local arcade.  I first saw it on Atari's 1978 seminal Football arcade game at Roller King in Kingston, PA sometime in early 1979.  Released in October of 1978, the game seemed almost magical as to how it worked.  How could simply rolling a ball around control something on the TV screen?  There was however one slight oversight to Atari's design that affected those with small hands, for the spacing between the ball and the metal panel was wide enough that it would occasionally pinch your fingers when rolling it.  It certainly wasn't painful enough to keep me from playing it, though, and the game soon became a favorite of mine.  So where did the idea for the Trak-Ball controller come from?

The History of the Trackball

The trackball device can trace its roots back to the military.  A British engineer by the name of Ralph Benjamin created the first trackball (called the "roller ball") in 1946 as part of the Comprehensive Display System (CDS) radar plotting system.  It was patented in 1947 but only one prototype was built; production versions of the CDS used joysticks.  In 1952 the Royal Canadian Navy built a similar system called the Digital Automated Tracking and Radar system (DATAR), which actually used a 5-pin bowling ball (LINK).

The Royal Canadian Navy's DATAR trackball device.  Photo credit Computer History Museum.

By the 1960s, similar devices were used with other radar tracking systems in other countries, including the U.S., as well as in other non-military applications.  Trackballs were used to control giant antenna dishes, such as the Honeysuckle Creek Tracking station in Australia that was used to help with NASA's Apollo missions; years later a trackball device was tested on NASA's STS-43 shuttle flight (LINK).

Paul Mullen moves the Honeysuckle Creek (HSK) Tracking station antenna with Servo console's track ball (late 1960s).  Photo credit

A "slew ball" being used at an air traffic control station (1960s).  Photo credit IFATCA Journal of Air Traffic Control.

Contrary to popular belief, Atari's Football was not the 1st arcade game to use trackball controllers; Taito apparently released a game called Soccer Deluxe (or Soccer DX) the year before.  Midway released the coin-op Shuffleboard in 1978 as well which used a trackball controller they referred to as the "big roller ball" in their flyer.

Atari's 4.5" Trak-Ball (TM) controller was designed by Jerry Lichac and used in their Football, Basketball, Baseball, Soccer, and Missile Command arcade games.

The success of Football begat other Atari trackball sports games - Basketball, Baseball, and Soccer, as well as a 4-player version of Football.  The most-successful game to utilize the large 4.5" Trak-Ball was Missile Command.  The cabaret and cocktail table versions of Missile Command used a smaller, white 2.5" Trak-Ball controller (called a Mini-Trak-Ball in the manual), which was next used in all 3 cabinet configurations for Centipede.  1982 saw a black version of the 2.5" Trak-Ball used in Millipede, Liberator, and Quantum.  1983's Crystal Castles used a red translucent backlit ball in both the upright and cocktail table cabinets that was a real eye-catcher.


Crystal Castles would be the last Atari arcade game to use a trackball, before the company was broken up in 1984 and parts of it were sold off.  Atari's coin-op division, under the Atari Games name, would continue to use the controller, starting with Marble Madness in 1984.  The game used a pair of translucent balls - one blue and one red.  It would be another 6 years before Atari released another game with a trackball, 1990's Shuuz, which turned out to be its last.  A sequel to Marble Madness, Marble Madness II Marble Man, was prototyped, but it used joysticks.  Other companies used the controller, such as Sente with its Snake Pit and Mini Golf games, and Capcom with its popular Capcom Bowling kit.  Variants of the trackball were used in other games, such as Atari's The Adventures of Major Havoc and Bally/Midway's Kick-Man (which looks like it uses a trackball but it has a pin running through it, only allowing for left or right movement), but despite Atari's efforts to broaden its use beyond sports games, the only games that have continued to implement it are bowling and golf games, most notably Global VR's Golden Tee series.

Marketing Hype That Failed to Deliver

Back in the trackball's heyday of the early 1980s, it was obvious that translating popular arcade games like Missile Command and Centipede to home systems became an issue due to being limited to the joystick controller.  The designers did their best to adapt to it, but the joystick could never offer the same experience the arcade Trak-Ball did, so Atari proposed a solution in the form of a home version of their Trak-Ball controller.  This became one of their "Pro-Line" Advanced Controllers and was announced in the January/February 1983 issue of Atari Age as due out by that summer:

First mention of the home Trak-Ball controllers in the January/February 1983 issue of Atari Age (pg. 11).

First advertisement for the home VCS/2600 and 5200 Trak-Ball controllers in the May/June 1983 issue of Atari Age (pg. 28).

The May/June 1983 issue offered us the first advertisement for both the VCS/2600 and 5200 Trak-Ball controllers, along with their respective prices ($45 and $75), as well as an article on how they worked.  Of particular interest is the last part of that article, where it explained the difference between digital and analog controls, and in particular, between the VCS/2600 and 5200 controllers.  The author (Atari Age editor Steven Morgenstern, I assume) went on to state, "The reason 2600 and 5200 Trak-Ball controllers are different is that the Atari 2600 will only accept digital control."  Hold up.  Full stop right there.  No, the 2600 is certainly not limited to only using digital controls.  It uses both digital and analog, as both joysticks and paddles were included with the system the first few years it was sold.  Without the latter, Breakout would be about as much fun as juggling bricks.  Since Atari decided the 5200 should have analog joysticks when the vast majority of arcade games used digital joysticks, the $75 investment in a Trak-Ball controller was worth every penny for those games that took advantage of it (a better investment would have been Wico's digital joystick, but I digress).  Atari dubbed 12 titles worthy of being "Trak-Ball Compatible":

Jungle Hunt
Missile Command
Pole Position
RealSports Basketball
RealSports Football
RealSports Soccer
RealSports Tennis
Space Invaders
Super Breakout

Of those, only 4 were truly worthy of it - 1 of which, Basketball, was never released.  Inexplicably, the console's original pack-in game, Super Breakout, wasn't even listed as being Trak-Ball compatible, and that game is basically a spruced up version of the 400/800 cartridge which came out in 1979 (which uses paddles)!  The other sports titles (Baseball, Football, Soccer, and Tennis) were not bestowed with the label either, even though as I mentioned earlier, 3 of those games were early Atari coin-op games that used Trak-Balls.  Whether or not games like Space Invaders and Galaxian incorporated full analog movement, I don't recall, but I don't see how that would improve the gameplay if they did.  Gorf by CBS Electronics also supported full analog movement, even though the arcade version was strictly digital.

Although the Trak-Ball wasn't originally advertised as being for use with Atari's home computers, the VCS/2600 model (CX80) was later redesigned to better match the XL line:

Of the 5200 games mentioned, the only 400/800 counterpart that benefits from a Trak-Ball would be Missile Command, which was a bit of a mystery to me, as the game includes an undocumented keyboard command (CONTROL+T) that allows for full Trak-Ball compatibility, some 2 years before Atari released the controller!  I asked Rob Zdybel, who programmed 400/800 Missile Command, how was it he incorporated support for them that far in advance:

"We wanted to play with trackballs. and so we did.  We made our own trackballs.  I made the code cuts for trackball/three bases and distributed the initial versions (those things do tend to replicate, though) - all for the price of a very nice custom-built trackball presented to me by the 800 hardware technicians."

The only other 400/800 game known to fully support it is Synapse Software's Slime (press T before playing).  Final Legacy only partially supports it, on the Sea-to-Air screen.  Crystal Castles, Millipede, and Pole Position would have been a natural to use it, but they don't. 

The original "Pro-Line" version of the controller has a white ball, round fire buttons, and an all-black case.  There's a hole and markings on the left side of the case for a Trak-Ball/joystick mode switch, but the pcb design doesn't support it (and the hole is covered up).  The 2nd version of the controller has a white base (and no "Pro-Line" text on top) and the pcb has the mode switch.  Both versions are model CX22; the 3rd XL-style version is model CX80 and has a black ball and triangular fire buttons.  The mode switch is now on the back and is marked "JS" (for joystick) and "TB" (for Trak-Ball).  The case was designed by Tom Palecki.  Without getting too technical about the differences between the versions (the details of which are best saved for another article), the most-compatible version is the 2nd one, which is by far the more common one to find.  This version offers simulated Trak-Ball movement with joystick games (when the mode switch is set to "TB").  Since Atari never released a VCS game that actually supported a true Trak-Ball (i.e. analog) mode, this version simulates analog movement by pulsing the digital signals.  That's as close as you'll ever get to having true Trak-Ball control on the VCS.


The original VCS/2600 version (CX22) was labeled "Pro-Line Trak-Ball", with a later release dropping "Pro-Line" name.
A 3rd version (CX80) was redesigned to match the XL computer line.

Considering all the effort that was put into making and marketing the Trak-Ball controllers, it's shameful how little they were fully supported, with only 1 game for their home computers, and NO games for the VCS.  It was a sham, for all owners ended up with was a $45 controller that operated the same as a joystick, yet you had Atari telling people the controller "takes a good game (like Phoenix, Missile Command, Galaxian, and Space Invaders) and makes it awesome".  Well, it would have been awesome, had programmers ever bothered to support it.  But since it wasn't company policy to require them to do so, they didn't.

If the fact that Atari itself didn't even support its own trackball controller, consider there were 5 other companies who released or planned to release trackball controllers of their own!  There was Accu Company's "Accuball":

Marjac's "Joyball Trackball":

Sega Sports Pad offers simulated analog control similar to the Atari CX22:

TG Products' "Track Ball Controller":

Wico's "Trackball", which offers inferior simulated analog control to that of Atari's CX22:

.. and Zircon International's "Track-Ball":

Both Roklan Corporation (with their Unroller Controller) and Spectravideo (with their QuickShot IX Deluxe Joyball Controller) released controllers that look like trackballs, but are merely plastic-domed joypads.  In retrospect, they worked better as a joystick than a trackball did:


Years later someone hacked the following VCS games to take full advantage of Atari's Trak-Ball (other company's trackball controllers might require modification):

Marble Craze
Missile Command

Nexar, The Challenge of
SpaceMaster X-7
Star Wars - The Arcade Game

Flaws in the Crystal

The one game that's currently still missing from the above list has been deemed too much of an effort to hack for trackball support.  That game is Crystal Castles, and it's one arcade port that's had a love/hate relationship with fans of the coin-op game before it was even released, starting with the creator of Crystal Castles himself - Franz Lanzinger.  The game started out based on a 3-D version of Asteroids (tentatively titled Toporoids).  Although it was never a smash hit (less than 5,000 were made), it was highly original and influenced at least one variant (Lode Runner's Rescue by Synapse Software).  The game was advertised in the September/October 1983 issue of Atari Age, and home versions were already being planned by the next issue for all Atari game and home computer systems:

News blurb from the November 1983/February 1984 issue of
Atari Age (pg. 12).

The only system to see a home version released that year (1984) was the VCS/2600.  A prototype version for the home computers was leaked out after the market crash, and a completed version was finally released for the XEGS in 1988.  A 5200 version was apparently never worked on.

Steve Morganstern was in full hype mode for the VCS version.  Whether he hyped it because it was his job (shame on you for expecting otherwise, from a company-owned publication), because this was the April Fool's issue (and wasn't meant to be taken seriously), or because it ended up being the last issue of Atari Age (and knowing he wasn't going to be around to answer letters from angry subscribers), we'll never know.  But when he's telling everyone the VCS version features "flicker-free" animation, you know the fix is in, because even being a 16K, SARA-enhanced cartridge doesn't mean graphics can't flicker (and trust me, they do).  The blurb for the game includes a real zinger, proclaiming if you have a Trak-Ball controller, "the game becomes even better!":

Steve Morganstern's editorial from the March/April 1984 issue of
Atari Age (pg. 5).

Ad for Crystal Castles from the March/April 1984 issue of
Atari Age (pg. 7).  Club price was $31.95.
By the end of the year, stores were selling new copies for less than $5 each.

April Fool's joke ad from the March/April 1984 issue of
Atari Age (pg. 25).
Every copy of Crystal Castles might as well have included one, since the game doesn't fully support the Trak-Ball.

Which brings us to Franz's issues with the VCS version.  One only needs to look at both versions side-by-side to understand why he was so frustrated and angry about it:


Peter Niday (who programmed the VCS version) was a very talented designer, but asking nearly 10-year-old hardware to keep up with the latest technology was simply asking the impossible no matter who was programming it, especially when Atari's management didn't learn from rushing E.T.'s development and instead kept adhering to baseless completion deadlines for game development.

Thanks to Jed Margolin's VAX email archive, here's Franz's uncut comments about the VCS version of his game, and his failed effort to keep it from being released, as well as Chris Downend's response:

From: KIM::FXL 30-JAN-1984 02:01
Subj: The first annual Jeff Boscole Memorial letter

This letter is dedicated to Jeff Boscole, someone who wasn't afraid of sounding obscure, to speak his mind, to be strange, to be brilliant, to play games, and to use MAIL to its fullest. I don't remember when he left, but it was quite a few months ago.

To anyone who cares, but especially to game designers with more clout than FXL, and to any and all people in power at Atari (not just coin-op):

Recently I have spent an inordinate amount of time trying to instigate improvements in royalties, designer credits, and game testing procedures. I have had little success. I hereby apologize for all of the negative feelings and anger that I am emanating because of this. After all, things are pretty good here, and certainly better than at many companies. I do not however apologize or regret my negative feelings about the recent release of the Crystal Castles 2600 Cartridge. (In case you don't know, the cartridge was released without the approval of the coin-op design team, or anybody else in coin-op as far as I know).

This is pure theft!

And I do not even know who to blame for this!! It isn't the programmer, who is about as mad (or worse) as I am about this situation. He was given a unmakeable release deadline (4 days instead of 3 weeks from when he was told). The game is much worse because of this (according to the programmer Peter Niday). He had no choice in the matter. Yet another unfinished, hurried, poorly tested game from Atari. Won't we ever learn?

Games under license from other companies get reviewed by representatives of that company (Williams and Namco specifically). But games developed in-house are treated like they are in the public domain, while the original design team of in-house games is treated like dirt.

This is not an isolated incident either. Atarisoft, as a matter of policy, takes Atari Coin-op games, lets outside companies "convert" them for home computers (like Commodore 64, Vic-20, Apple 2, TI-99 and IBM-PC), and then produces them, all without the creative input or advice of the original design teams (just talk to Ed Logg about Centipede, or ? about Battlezone). Atarisoft does not ask anyone over here at coin-op for approval for the final version, but they do show the final version of the game to someone in the legal department. On the more positive side, there is a chance that Atarisoft will contribute to the Engineering Product Bonus Plan in a manner similar to 2600, 5200 and 800 products. Wouldn't it be nice to have that guaranteed and in writing? And shouldn't there be designer credits on Atarisoft products?

It's ironic that my name is on the packaging of the 2600 Crystal Castles cart, a product which I only saw an early version of. Yet when I told people that the message ("programmed by Franz Lanzinger") appeared in level 10 in the coin-op version I was told to take it out, or I loose an amount of bonus to be determined. Boy did that make me mad !!! I complained vocally, but only to be promised that a designer credit policy would be worked on. This policy is still "being worked on" eight months later. Now really. It's not that hard to do, just look at movies, books, not to mention Stern, Mylstar, Simutrek, Sente, even 2600 carts. If there were a policy right now, credits could be in in time for the Crystal Castles kits. As it is, I am still mad about the whole thing. Imagine Steven Spielberg directing a film, but not getting credit. How would he feel? Are we cogs in a machine? I am not a number !!! This isn't 1984!! (well OK, maybe it is).

While I'm at it I would like to get one more thing off my chest (right on!!). You may know that the current "coin-op engineering product bonus plan" (shouldn't it really be called a royalty plan ?) is out of date. The most recent legally binding document (if it is legally binding) is dated March 26, 1982, and it expired at the end of 1983. It is my understanding by reading that memo that the bonus plan is still in effect, but it can now be " extended, enhanced, discontinued or otherwise modified to meet management objectives ". In other words, Atari has the legal right to screw us any time they want. Personally, I would feel much more secure, happy, and motivated to work hard, if there were an updated royalty plan without a gaping loophole like that.  After all, there are plenty of precedents for people getting screwed here.

I am tired of fighting a brick wall. So I will resign myself to the facts of life at Atari. These facts seem to be that change is virtually impossible when suggested by a single employee, but mind-bogglingly fast if management wants it. And I will continue to feel bitter now and then (like right now for instance).

How do you feel about all this? How do you feel about 40% 30% 30% (the "golden handcuffs")? How do you feel about 1% under 10M, 2% over 10M? How do you feel about designer credits? How do you feel about the delays in actual payment of royalties? (I still don't have a cent for Crystal Castles, and it has been seven months since it started to earn millions for Atari).

What can you, anyone who cares, do to make me, Joe Piscopo (oops, make that Franz Lanzinger) feel less bitter? Well, misery loves company. Please tell me, better yet, tell your favorite manager, supervisor, or even CEO, how you feel about these issues. It may not change a thing, but maybe your powers of persuasion will succeed where mine failed.

Until next year, (when I will write the second annual Jeff Boscole memorial letter)
(the X stands for "eX trouble maker")

P.S. please send your answers to @SYS$MAIL:JUNK, or to someone in a position to take action, best would be both.

P.P.S. If there are any inaccuracies, please let me know. The facts are to the best of my recollection, some of it is hearsay.


From: KIM::DOWNEND 1-FEB-1984 17:22

to: Franz
from: Chris Downend
Subject: Response to Boscole Memorial Letter

First of all, rest assured the issues you mention ARE being worked on - they are constant topics at meetings I attend.  Everybody seems to be involved in making decisions these days. This ensures all viewpoints are heard but with a horrible speed penalty.

The solution is to keep plugging away and enlist support as you have! By the way Franz, you are certinly a valued employee with clout - one thing Atari undeniably values is people that can produce successful products and you have certainly done that with Crystal Castles.

I have unique perspective on the situation since I have programmed games at the "bottom" and at the same time I have seen the decision-making process at the "top" - I can empathize with both sides.

One word of caution though, I note that you suggest dialogue thru VAX and the "Junk" heading which routes the text to everybody on the VAX including employees who do not share in the Product Bonus (royalty) and they just might not appreciate hearing about our lofty concerns about credits and amounts of bonus since they get neither. May I suggest the @sys$mail:engineer.uaf heading or a similar restricted audience. Management must be sensitive to the feelings and desires of many diverse groups inside Engineering and this complicates and lengthens the decision-making process. A snap decision to address the issue bothering one party may upset another party - management has to consider the whole picture sometimes.

On Royalties: Yep, Atari can screw us anytime they want.  I do not think they would for fear of a lot of people leaving. The Company has to protect itself. Please realize that thru much of 1983, Atari paid bonuses even though Coin-op was not making money - we were operating in the RED and still paying bonuses ! Now that's commitment. Of course that cannot go on for too long or else the whole Company goes down. That's the reason Atari has escape valves built in to Bonus plans - it's not really too screw the employee, but instead to protect the well-being of the Company.  That's the price you pay for the luxury of a steady salary and a ready-made work environment including PEOPLE and technical support. Personally, I have not been screwed, and in fact I have found that Atari has handsomely rewarded hard work and a willingness to support the Company. Maybe my expectations are lower than those who feel screwed - or maybe they valued themselves more than they were really worth.  Management does care and Changes are in the works, they just take a long time especially when the players keep changing( J. Ray becomes Director, then Calfee leaves, then Farrand leaves etc. - you have to keep re-educating the new players).  Changes ARE underway (no promises, but people want to fix these things if possible):

-get rid of golden handcuffs
-generate an advance close to initial production

As for the "millions" Atari made on Crystal Castles, well lets see:

sales as of 1/13/84: 4363 uprts; 450 cocktails
sales revenue (approx): $2095*4363 + 1695*500 = $9.98M
[price was reduced in DEC(?) to $1000(?)]
cost of goods sold (fully burdened):$971*4363+971*500=$4.73M
Engineering Expense for Crystal Castles: about $1M
Engineering Expense for games that don't make it: unknown
Sales/Marketing Expense: unknown
Engineering Bonus expense:(.015*9.98M)=.15M
Pre-tax Income: 9.98-(4.73+1+.15)= $4.1M
After taxes (50%): $2.05M

So the Company retains earnings of a couple of million to get thru the many dry spells this industry faces or to buy new equipment etc.  Also note that Atari had to build about a 1000 games to break even on the Engineering costs. Thus, profit doesn't really appear until 1000 games are built, but Atari pays bonus anyway. I agree 7 months is a bit long to wait, but Atari has not made all that much and the product was not profitable until long after the initial production started. Product Bonus was paid quarterly at one time; we should go back to that scheme.

As for the 1% or 2%, I don't see a problem there - after all, Atari doesn't start making any significant money on a product till a few thousand are sold so it makes sense to reduce bonus funding till a threshold is crossed. I would however like to see another threshold at about $50M when the percentage increases to 5%. A game that can generate that much sales is a spectacular achievement for the creators and they deserve the reward at that point.

The Quality of 2600 carts is the pits - no doubt about it.  The system was introduced in 1977 so it is SEVEN years old. I think Breakout and Space Invaders are decent renditions of the coin-op originals and those are 1976 and 1979 games respectively.  But with 14 million 2600's out there, financial issues outweigh aesthetic issues.  I don't think the public would even buy Crystal Castles on a 2600 so everybody loses - you and Atari. Again, we've got new management and they have to learn from their own mistakes. By the way, Calfee knew the 2600 Crystal Castles was lousy and tried to stop it, but he was overruled. When Marketing wanted to do the same thing with Millipede (release the cart with a bug), Steve had to go all the way to J.J. Morgan. Fortunately, Morgan agreed with Steve and the release was postponed. One thing to remember though, Coin-op profits are small potatoes compared with Comsumer profits so every decision is heavily weighted toward maximizing profit in the Consumer arena. So, anticipate feeling screwed with respect to the quality of carts - it won't change - too much money is a stake. The virtue of Coin-op is extensive creative freedom (in game design and hardware base) since original work is the lifeblood of the Industry. Coin-op also gives you bearable schedules allowing you to do a satisfying job. And to my knowledge, Coin-op has yet to sacrifice quality to get an on-time delivery. Firefox was supposed to start production 1/23/84; millions in parts are all staged ready for production, but it has not started (1/31/84) because the software is not ready.

Now for Credits: Coin-op credits are more complex than Consumer credits since more people are involved and people get their feelings hurt if they are left out and they feel they contributed just as much as so-and-so and so-and-so got their name on the game... see my point? John Ray has been working on this as well as trying to learn about being a Director and managing the Project Office. Maybe its could have happened faster, but John manages by consensus which takes even more time.  John has apparently sent his recommendation to Van Elderen/Paul/Farrand for approval prior to publishing the rules for credits on the audio-visual portion of the product.  How does seven names in video for the audio-visual portion strike you?? We do not want 8kbytes of EPROM used up putting 500 credits in the game.  By the way, Star Wars got their names in the game because they did it and did not tell anybody about it. If your ethics were equally low, Franz, you could have done it in Crystal Castles too.

What's all this mean? I don't know. I hope it helps though.

- Chris Downend

Return to main menu