FINISHING MAGIC GIRL (2024)

The history of theMagic Girlpinball had always been one of repeated failure.

What began as a heady mix of excitement and wonder turned, over a period of a decade, into a cursed project fated never to reach production.

Yet, what appeared to be an abject and irredeemable flop proved to be an irresistible opportunity to a small team based in the north of the Netherlands.

This is the story of what they achieved.

Magic Girlfirst came into the public consciousness in July 2011 when former pinball designer, John Popadiuk Jr., announced his company’s intention to build a small run of a high-end, full-featured pinballs.

John had been designing pinball games at Williams during the ’90s with a run of popular titles such asWorld Cup Soccer,Theatre of Magic,Tales of the Arabian NightsandCirqus Voltaire. His final title for the company,Star Wars Episode 1, was alsothelast title for the company, as Williams closed their pinball division during its production run at the end of 1999.

With the pinball business appearing to be limping into oblivion, pinball design jobs were few and far between. After setting up his own company, Zidware, John created a number of pinball design multi-media products, including books and apps.

These turned out to be very niche products and his interest in them soon waned, but greater success came in 2006 when John teamed up with Roger Shiffer – the inventor of the Furby, for those who recall what they were and how popular they briefly became – and his company, Zizzle.

Roger wanted to produce a mass-market, low-price family pinball machine, using a minimum of parts and a simplistic design. The fruit of this union was a pair of three-quarter size licensed titles, Disney’sPirates of the Caribbean: Dead Man’s Chestand Marvel’sHeroes vs Villains. Both used the same playfield design and mechanisms but each had its individual artwork, sounds and playfield toys.

Although not targeting fans or collectors of arcade pinball machines, the Zizzle titles sold well through retail outlets and went on to spawn another pair of Disney and Marvel games,Pirates of the Caribbean: At World’s EndandMarvel Super Heroes.

Two more licensed titles were in development when Zizzle shut down production leaving John once again without an outlet for his creativity.

His solution? Design and build his own pinball machine.

Magic Girl was announced by John on 21st July, 2011 and described by him as “a 1990s inspired custom pinball machine full of flipper flash, digital fun, innovative play and magnetic creative wonder!”

John didn’t have a manufacturing facility he could utilise, so he planned to build the games in his own design studio on an industrial park in Streamwood, Illinois. Consequently, there would only be 13 Magic Girl machines made and the price of each would be $15,995.

At the time, that figure was more than double the price of most popular new titles, creating an online outcry and pushing the game far beyond the reach of most collectors and operators.

Nevertheless, there were still plenty of buyers with deep pockets willing to pay the asking price and accept the somewhat bizarre terms in order to own one of these prized machines. Enough buyers, in fact, that the number of machines to be built was increased from 13 to 20 just a couple of weeks later.

To purchase one of these curiosities, a buyer needed to pay an immediate $500 deposit and sign a non-disclosure agreement (NDA) to ensure they didn’t reveal any details of the designs or mechanisms planned for the game.

Until this point buyers had no idea what they had agreed to purchase other than John’s vague description of, “…gorgeous hand-drawn artwork, updated cabinet design, new innovations in playfield mechanics, modular expandable hardware, special social connectivity ALL created just for this full-size pinball game hand-built by the designer in Chicago.”

It was an ambitious project, but as John said at the time, “Overall making pinball is very hard and this game will be as challenging. As no one else will have me, then I will go this route to keep pinball vibrant and innovative as best I know how.“

Making pinball is indeed “very hard” and, as deadlines flew past, the pre-order money started to dry up. Although he continued working on prototype designs and testing new ideas in his studio, no definitive design had yet appeared, let alone any schedule to actually manufacture the games.

In what has subsequently become a red flag for a project running out of cash, a second title was then announced, resulting in a temporary easing of the finances as more pre-order money flowed in. Initially called Ben Heck’s Zombie Adventureland, a falling out between John and Ben led to this second game being rebranded into Retro Atomic Zombie Adventureland (RAZA) instead.

Another long-term project of John’s, a game based on Lewis Carroll’s Alice’s Adventures in Wonderland was also launched with another round of pre-orders coming in.

With three titles in development but none close to reaching completion, unrest amongst buyers increased. This heightened further at Pinball Expo in 2014 when RAZA buyers were expecting to see a completed game design revealed, but John turned up instead with just two empty prototype cabinets.

After another year of experimentation and development, in 2015 John announced the initial funding for all three titles had dried up and he had no means to continue working on the games. Without further help, none of the games would ever be produced.

This looked as though it was the end of the road for Magic Girl, RAZA and Alice in Wonderland.

Magic Girl was the most developed of the three, so a rescue package was attempted to try to salvage that project with the hope of selling enough additional units to eventually fund the completion and build of RAZA and Alice in Wonderland.

However, when the most-developed prototype version of Magic Girl was sent to a potential financier in June 2015 it was clear the game was still a long way from being production-ready.

There were playfield mechanisms missing, while the playing surface was a flimsy overlay which didn’t even lie flat.

With significant design work still required, the proposed deal collapsed and all three titles really did appeared doomed to the dumpster of history – unfinished and uneconomic to complete.

To his credit, with legal action started against him for failure to deliver, John never gave up on getting his game designs built.

In 2016 he struck a deal with a new pinball design and manufacturing company, American Pinball. In return for designing their launch title based on the illusionist and escapologist Harry Houdini, American Pinball would manufacture a small run of Magic Girl machines.

It was a curious arrangement. Twenty-five Magic Girl games would be manufactured on the American Pinball production line but released under the Zidware name, with John having his own dedicated area in the American Pinball factory from which to oversee production.

American Pinball didn’t want to talk about the game at the time – Pinball News visited the factory to see production but we weren’t permitted to publish any pictures then or even talk about the work taking place. Even now the company appears sensitive about discussing the subject.

But if the manufacturing arrangement was odd, what was actually shipped to customers was totally bizarre.

With little or no further development work possible before construction began, the game was built according to the incomplete design as it then stood, using whichever parts were available.

As a result, many of the mechanisms on the playfield either didn’t work or were completely missing since they hadn’t been designed yet, while below the surface there were numerous cable connectors left hanging with nothing to plug into. The game was essentially unplayable – simply a box of lights, albeit a beautiful one.

Probably the most complete parts of the game were the artwork and the software. Both were almost finished when the games were built, although the software’s operation was hamstrung by the fact that many of the switch inputs and device outputs were absent, making it behave in unexpected ways. Production testing and QA must have been difficult or non-existent considering how the game could never have worked properly.

Also, because the parts used were largely those already in stock, nearly every machine is unique in its mix of components used and the way they are installed.

However, to the best of our knowledge, twenty-two Magic Girl games were built and boxed, with eighteen of them shipped to excited, if somewhat bemused customers who had at least received something for their money.

Retro Atomic Zombie Adventureland and Alice in Wonderland buyers were left out in the cold, although the story of those games didn’t end there as all three titles would come back as planned future game designs when John joined Deeproot Pinball’s ill-fated design team.

So, the twenty-two rare and beautiful but incomplete and unplayable machines made at American Pinball plus a couple of similar prototype games built at John’s Zidware studio were all that existed after ten years of efforts to design and manufacture Magic Girl.

With Deeproot Pinball’s eventual failure, that appeared to finally be the end of the Magic Girl story.

That was until one of those American Pinball-made machines ended up in the collection of a Dutch pinball fan near Rotterdam.

One day, pinball restorer and collector, Rogier Voorderhaak, was visiting the collector and discovered Magic Girl for the first time. He saw how beautiful it looked, heard the history of the game, fell in love with it and immediately contacted his friend Max Rogmans who owns a large, pristine collection housed in his gameroom at the rear of a factory he owns in Wormer, north of Amsterdam.

At the time, Max was just starting his collection and had around a dozen machines. Rogier brought Max to Rotterdam to see the collector and hear the story of Magic Girl.

Max has previously discussed with Rogier what, as a top-end restorer, his next dream-come-true game would be on which to work. Rogier described to Pinball News how he likes to take a game which isn’t working and make it work. Indeed, it seemed the worse the condition of the game, the more he relishes the challenge of bringing it back to life.

Rogier told Max he needed to have one of these Magic Girl machines in his collection, but more than that, he saw the potential to make the game how it was always intended to be. They both liked John’s previous titles and could see how Magic Girl integrated many of the features from those earlier designs.

Fellow friend, Eric Bartels, is the owner of coin-op supply, maintenance and restoration company Pinball Professionals and works out of Max’s building in Wormer, servicing and restoring Max’s machines too.

Rogier and Max asked Eric for his opinion as, Rogier said, “if you ever reach a dead end with a pinball, Eric will be able to find many possible ways to bring it back to life.“

Eric said he only had five minutes to decide whether it was an achievable project. “There are the obvious problems but there are also many you can’t see“, he told us. He obviously couldn’t take the collector’s game apart, but still decided nothing was impossible. “The only limitations were your imagination and money.“

Rogier, Max and Eric discussed the idea of buying a Magic Girl machine and working together to add the missing mechanisms, getting the existing ones working correctly, connecting everything together correctly and finishing the software to drive them.

It would be a huge project, but they decided it was achievable and would result in the creation of a genuinely unique game.

After contacting Magic Girl owners through Pinside – where the owners list wasn’t entirely up-to-date, with around 35 people listing the game in their collection – Max eventually sourced a game from a collector in the US and had it shipped to the Netherlands.

Although it wasn’t exactly cheap, many of the original Magic Girl buyers had tired of having this non-playable light show in their homes and were now prepared to sell. In fact, the game Max eventually bought was still essentially ‘new-in-box’.

Max said he knew the amount of work needed to make the game playable, so if they were going to invest all this time, effort and money, why not do more than just one?

Also, they didn’t know if the game they had was typical of the other Magic Girls, or if it had some unique differences. To find out, they bought two more games so they had one each, although these additional games weren’t both new-in-box like the first one.

At this stage the trio had no idea real how the game was intended to work, but they were determined to find out whatever information they could about what John had intended.

With John now working for Deeproot Pinball, the intellectual property rights to Magic Girl had been taken over by the company with the title slated as one of their upcoming games. John, therefore, wasn’t in a position to help the team fill in the blanks on this original iteration of the theme. He also expressed his belief that the machines shipped to customers were far more complete than was clearly the case.

Likewise, the game’s software programmer was now under an agreement not to talk about any details of the game’s code or distribute any assets. He told the team that there had been a time, a couple of years earlier, when he would have been willing and able to provide assistance and even try to complete the code if anyone had asked and he could have got his hands on a machine. But, by this stage he simply wanted to move on and concentrate on his current projects instead.

However, the team were helped by the fact that most of the rules appeared to be already coded into the game.

This hadn’t previously been appreciated, since many of the switch inputs and output devices were either missing or disconnected. Once they were correctly wired into the game, features started to work and the rules and objectives became a lot clearer.

There were still plenty of mysteries to be solved though.

Some devices, such as the levitating magic box, the semi-circular ramp or the ‘Magic Flip’ upper playfield could never have worked given they way they had been built. What were the intentions behind them? Was there a plan to implement them in a different way to how the game ended up being built?

The levitating magic box feature, for instance, was supposedly intended to lift the ball off the playfield and hold it inside the device, like a vertical version of the Spirit Ring in Theatre of Magic. But the magnetic coil around the box would never have been powerful enough to reliably raise the ball such a distance without some further assistance.

Also, the upper playfield utilised magnets to flip the ball, using the same technique as the Powerfield on Twilight Zone. However, the upper playfield had a drain hole positioned directly in front of the magnets in such a way that the required ‘Magic Flip’ shot to the back of the upper playfield could never be intentionally made without the ball draining, and could only be achieved by a very lucky fluke.

Also, the ball was delivered to the upper playfield in the same way every time, making its position reasonably predictable and removing much of the desired randomness. The team made a second hole to feed the upper playfield to vary where the point where ball dropped.

Meanwhile, a weak shot to the long semi-circular ramp could easily travel only part way round, roll back and then become irretrievably stuck since there was no escape hole to release it.

It was clear some decisions would have to be made about how to implement the non-functional parts of the game and fix the most obvious deficiencies. New mechanisms and assemblies would need to be built, small changes would have to be made to some of the artwork on the playfield plastics, the missing code would need to be written with the existing bugs fixed, and some of the electronics either upgraded or redesigned and built from scratch.

When problems called for solutions, Max was happy to trust Rogier and Eric to come up with the best way to resolve them. They did set one fundamental rule though – there would be no changes made to the actual playfield itself. So, everything had to be reversible, with no new posts, holes or inserts added. They said their guiding principle was always, “what would John have intended?“

It was a huge undertaking and the team didn’t have the full set of necessary skills themselves, but they had many good contacts and could call on additional specialist help when it was needed.

So, they had now three machines, and although they all outwardly looked identical they were not all the same.

The colours used in the playfield artwork varied, insert colours differed too, there were diverse standup target colours, while different screws and different numbers of screws were used throughout, with important playfield mechanisms sometimes attached by just a single screw.

The game didn’t include an auto-launcher as John doesn’t like them – only one of his Williams/Bally games had one – but the Magic Girl wizard mode is a six-ball multiball, so adding an auto-launcher was a must.

The game’s switch matrix supports 64 switches, but there are only 40 installed. While most games don’t use the maximum possible number of switches, it was clear from the presence of multiple unused cables under the playfield that quite a few switches were intended but were never simply installed on Magic Girl. The cables were labelled with the supposed location and function of the switches, but the switches themselves were missing.

In addition, what wiring had been done often proved to be faulty, with some of the switches which were installed registering as multiple switch closures in the matrix.

It turned out the one thing that was consistent across all three machines was the way the diodes on several of the switches had been installed backwards, creating electrical short circuits.

So, they fixed those diodes and started adding the missing switches based on the cable loom and switch matrix descriptions. Gradually the game started to come alive and behave in more expected ways.

However, it became clear how, during the game’s development, John would never send a physical playfield to the software programmer – the work had to be done either virtually or over Skype. So, the programmer built his own physical version of the game based on John’s design. Meanwhile, John would secretly change how certain shots and devices worked, meaning the software no longer matched the hardware.

Max said his feeling was that around 90% of the game code was complete. The wizard mode was missing and several playfield features didn’t operate correctly. They hired a programmer to complete the code to their specifications. The programmer wasn’t familiar with pinball specifically, but could still code the missing wizard mode and fix the existing software to work with the new hardware.

From the team’s investigation they discovered that the original idea was to lock balls on ramps using the ‘Hokus Pokus’ magnets, but apparently the magnets became too hot when permanently energised and melted the ramps. So that idea was abandoned, but no alternative lock method was ever installed to replace it. As a result, it wasn’t actually possible to lock balls and start multiball.

To fix this, the team added solenoid-controlled gates under the hand plastics to lock the balls on the ramps, adding switches to sense the balls’ presence and changing the single-colour LEDs under each plastic into RGB LEDs so the colour could indicate when lock is lit. It could also change colour when holding a ball or flash tell the player when the ball is about to be released.

There was a ‘Eureka!’ moment when, having added the new mechanical lock mechanisms and the associated ball lock switches, they were able to progress far enough through the game to start multiball – something no Magic Girl owner had ever seen before.

There were still plenty more challenges ahead, and it was the upper-playfield that proved especially problematic.

Not only did the magnets used to flip the ball not work and had that exit hole directly in front of them, the timing of the magnets’ pluses was all wrong to offer any kind of player control.

In addition, the magnets were too close to the edge of the playfield to be able to fling the ball any distance, and even if you managed all that, the exit from the upper playfield was blocked because the playing surface was too long. Eric said, “You’d fix one thing, but that would lead to two new problems, and they would lead to another six problems you’d have to fix“.

He said because it had never worked, nobody was able to test it, but it could never have worked the way it was built. A new upper playfield with new magnet positions and a shorter exit lane was made to correct these issues.

The game’s ball trough used a slow, outdated method of detecting balls which took too long to report back to the controller, resulting in confusion at the CPU about where the balls were in the game. Eric said they had to make new PCBs using the faster optos used in modern Stern machines to get reliable results. Likewise, the optos used on the skill shot 180-degree ramp protruded too far and would get smashed as the ball passed.

The skill shot optos were also only attached with adhesive tape despite having mounting holes available, so a new opto board was needed here too which could be properly bolted to the skill shot assembly.

The list of fixes needed to get the game working grew longer and longer.

The driver board (supposedly made by Rottendog, with the suggestion that John never paid them for the work) was hanging under the playfield on cable ties. Eric said besides mounting it properly, he had to modify it electrically to be able to reliably handle the higher voltage now used in the game.

The switching power supply originally used was under-powered ([emailprotected]) for the game and would just shut down if too much was happening on the playfield, so it was upgraded and a capacitor board added to offer more impulse power when needed. Eric said he needed at least 10A-11A, but got a PSU with 15A which limits at 15A if too much current is needed and doesn’t shut down.

Eric said that the thin wires running to the coin door interlock switch carried all the high-power for the playfield’s solenoids. To fix this he added MOSFET control to remotely switch the high power instead.

The Potions spinner didn’t work and it was only designed to give one switch-closure per spin, so Eric designed a new one with multiple closures per spin, with the exact number produced adjustable in the software to make it easier or harder.

The ‘O’ and ‘C’ inlane inserts were hidden under rabbit plastics which prevented the player seeing if they were lit. New plastics were made with transparent dots in the artwork to allow the inserts to be seen through them.

The Lion Saw mechanism was mounted differently on each machine – when it was installed at all – and even then it couldn’t turn properly, so they redesigned it so it would fit into every game and spin freely using a motor.

To get the magic box to be able to grab the ball, they added a small ramp below the magnet which lifted to ball high enough for the magnet to take hold of it.

There is a spinning disc above the left outlane, but it would only spin in one direction, giving the player no control over it.

Fortunately, amongst the many incomplete elements of the game was a second flipper button on the left side of the cabinet. It wasn’t connected to anything, so the disc’s motor was changed to a bi-directional one and a switch added to allow the player to change the disc’s direction of spin so they could attempt to save the ball from the outlane.

When it came to the final Wizard Mode, although it hadn’t been programmed there were some existing video and audio assets in the game which could be used. The team knew the wizard mode was related to extracting the sword Excalibur from a stone, but had to create the mode’s rules based around that idea for their programmer to implement. This mode would then lead to the six-ball multiball finale.

Along with the many re-purposed devices from earlier games, Magic Girl did also contain some interesting new ideas for established mechanisms.

The pop bumpers, for example, didn’t have the familiar plastic ‘spoon’-style leaf switch below the playfield to detect movement of the rod attached to the bumper when it is in contact with the ball. Instead, two opto pairs mounted at 90-degrees to each other were used either side of the rod to detect when it moved. This frictionless method should be more reliable and would also not require any adjustment over the years.

As far as the team are concerned the project is finally done, since everything now works as they believe it was intended to.

The team can’t even begin to estimate the number of hours spent finishing the game – 1,000? 2,000?
They had to make numerous parts. The original pop bumpers had no lamps installed in them, so they had to add those. Some of the magnets were wound on cardboard formers, so these had to be remade using a plastic former. The general illumination was permanently on, but they added circuitry and code in the software to switch it off at appropriate points in the game.

Now, having done all the work, Eric said he “can’t stand the machine any more.” He said at the beginning there were masses of bugs in the game – “there were only bugs, it was terrible” – but gradually they worked through them until there now are just a couple remaining.

One is, when a ball is locked for multiball, the next ball is auto-launched without the opportunity for a skill shot attempt.

The other relates to what happens if a player fully completes the final wizard mode. The team don’t expect very many players will progress far enough through the game to ever experience that one, but if it does become a problem they will get it fixed.

They keep being asked if some kind of kit will be available to make the other Magic Girl machines complete.

Although Eric said he thought they had enough custom parts to convert around twenty machines, it is patently obvious there is no ‘one-size-fits-all’ solution to taking a Magic Girl and transforming it into a fully-playable game. Each machine is different and the huge amount of work needed to address the numerous problems in each game simply can’t be made into a kit.

However, the team do say that if a Magic Girl owner really wanted the work done to their game and could ship the machine to their workshop in Wormer, they would be willing to work their own ‘magic’ on it.

Although they did pledge not to make any non-reversable changes to the game, they couldn’t help include a little reference to themselves in the form of their own initials which now appear in the default high score tables on the display.

Having finished their work the team had planned to unveil a fully-working Magic Girl game at a show in the US, but the Covid pandemic meant they weren’t able to travel to the country. Instead, Eric has placed his game at the Dutch Pinball Museum in Rotterdam since the end of last year.

Rogier said that so far only a small number of people have played a complete Magic Girl – mostly at the Museum – so they haven’t had a great deal of feedback about the game, the rules and how reliable the new parts are.

In the end though he says, “Magic Girl is Magic Girl.” The game is what it is, and while there will always be ways it could be improved, they don’t want to change the nature of the game.

And, said Eric, “It’s a really good game. A tough game, but a good game“.

Huge thanks to Max, Eric and Rogier for their hospitality in Wormer, together with the opportunity to learn about all their work and to finally play the finished game.

Update: A ‘finished’ Magic Girl is scheduled to make an appearance at the 2022 Pinball Expo held in Schaumburg, Illinois on October 19th-22nd.

If you get the opportunity to play it, either at Pinball Expo or elsewhere, you might like to read up on the rules first. So, here they are:

Magic Girl Game Rules: