
Puzzle Game Creators – Developer Interview
Originally published in Game Hihyou magazine in 1998, this feature saw several key puzzle game creators reflecting on the essence of puzzle video game design and sharing the origin stories of several of the most fondly-remembered puzzles games of the late ’80s and ’90s, from arcade classics like Puyo Puyo, Puzzle Bobble, Columns and Magical Drop to the console-exclusive Panel de Pon.
Babel no Tou
Hiromu Nagashima – Namco VS Department manager; joined Namco in 1978. Developed Sky Kid and Pro Yakyuu Family Stadium as leader of the so-called “Pikkari Team”.
Babel no Tou was the first game I ever designed. Back then, I was working for Namco as a mechanical design engineer on arcade cabinets, right around the time the Famicom came out. Video games were starting to enter the home in earnest and it seemed like the potential for games was widening; I naturally wanted to make a game myself, working at Namco and all, so that’s how the project kicked off.

The reason I pitched a puzzle game was because I thought the appeal of the game would be something people could grasp from the design stage — the deterministic, rules-based world of a puzzle game is such that the fun of the game is easy to communicate on paper. At that time, challenging puzzle games like Sokoban and Lode Runner were popular, so I wanted to make a home game in that vein that was even better… one day, as I absentmindedly stared off into space, I suddenly envisioned tons of rocks clanking down on top of each other and towering off into the sky, and I thought that with the right rules, that could make for a pretty exciting game.
When it comes to puzzle games, I don’t think determining the rules is that tough — it’s ultimately a black-and-white, it’s-good-or-it-isn’t proposition, so all you have to do is come up with a solid, well-considered set of rules that everyone’s happy with. However, just making a game isn’t enough; I think a game is only truly “complete’ when the player figures out their own way to interact with the game and then “uses” it to play, so in that respect, coming up with puzzles was somewhat challenging.
The appeal of puzzle games is, in a nutshell, the elegant simplicity of the game system — Tetris is certainly a beautiful game, isn’t it? There’s also the catharsis that comes with playing the game, and overcoming some hair-pulling challenge; that’s true of all kinds of games, of course, but it’s acutely true of puzzle games, and being able to to both elicit and amplify that cathartic feeling is crucial when making a puzzle game.
Columns
Hisaki Nimiya – Sega R&D AM3 planning section manager; joined Sega in 1986. Representative works include Columns and Dennou Senki Net Merc.
Around 1988-89, I was doing in-house research on American games; at the time, Tetris had just become a big hit, and a whole new market had emerged. From a business perspective, it was decided that “we need to ride this wave”, but there was no sense in merely imitating what had already been done, so based on a few original prototypes, we ended up making Columns.1
People don’t simply discover diamonds, so to speak; it takes human skill to craft a rough gem into a jewel. My role on Columns was to sift through the rubble until I found a diamond in the rough, polish it, pair it with a band, design the base and craft the finished ring. Columns’ noteworthy attributes were the colored blocks and the rule around erasing blocks diagonally — diagonal matches are hard to identify, which leads to a lot of unexpected clears, and the chains created during those moments are also important.
I deliberated for a long time on the size of the play field: it’s 13 blocks tall and 6 blocks wide, and that’s an important detail. Falling-block puzzle games like these present you with challenge after to challenge to overcome, and the pace of the game is dictated by the size of the play field, the speed at which the pieces fall and the number of blocks per piece. In the case of Columns, one three-block piece falls at a time, and you need to put down at least two pieces (6 blocks) in order to form a line, so if you’ve placed two pieces without forming a line, you’ve made the game a little harder for yourself. The basic rule require one to match 3 same-colored blocks for a clear, so ideally, the player should be averaging one clear for every piece they put down. That’s pretty strict, huh? Tetris averages at two-and-a-half blocks per line. In that respect, only by chaining can the player stand on equal footing with the game rules.
As a creator, I work on designing games from an engineer’s perspective — that is, using the left-brain processes of analysis and logic rather than the right-brain processes of emotion and intuition. Having that hotter-than-average passion to be like, “this is what I want to make!!!”, while simultaneously being able to detach yourself from it and view it objectively as a product, is an important skill. Columns is an example of that analytical approach yielding good results. The side-effect is that I’m often criticized for being overly logical. (laughs)
Magical Drop
Takashi Kobayashi – Data East development section manager; joined Data East in 1985. Representative works include Fighters History (home version).
We’ve always been into puzzle games, and whenever someone discovered one they liked we’d all duke it out in the office during our lunch breaks and evening breaks. XBAND2 lets you battle people complete strangers in games like Tetris and Panel de Pon, so I’ve been playing in spite of the slow connection speed.
The beauty of puzzle games is their casual, pick-up-and-play nature. We wanted to make a game like that for ourselves, and after a lot of research for reference materials, we came across an intriguing game in a software compilation from a Russian company called RUSS; it was a simple game about vacuuming up blocks and spitting them back out, and there wasn’t much else to it.
This was around the time of the collapse of the Soviet Union, when the academics and military personnel over there were all scrambling to strike it rich, toiling under the shadow of Tetris. The game itself was amateurish — it wasn’t very playable and didn’t feature a combo system, and the blocks were a hodgepodge of donuts and polar bears and whatever else — but there was something about that basic system that stuck with me.

I started thinking of ways to make it more interesting, and from there came the genesis of Magical Drop. If it wasn’t for the mess over in Russia, this game might never have come to be (laughs) When we signed the contract with RUSS and began production, our version naturally evolved into a very different game to the one they’d produced; subsequently, we ended up completely acquiring the copyrights.
Hiromichi Nakamoto – Data East development section chief; joined Data East in 1985. Representative works include Metal Max.
The puzzle games we’re into are all action-centric: rather than focusing on building combos, we want to bring out that frustrating feeling of “argh, if I’d just pushed the stick to the left a little faster…!” Naturally, Magical Drop has evolved in that direction, too, and it’s become quite a vigorous game, so it might look a little intimidating at first, but once you grasp the concept of active chaining, even newcomers can quickly start making chains, and from there, you’ll be hooked. What’s fun about puzzle games, I think, is that definite sensation of honing your skills as you become better, day by day, and it’s also nice that the player can easily recognize their own screw-ups when they fail.
Panel de Pon
Hitoshi Yamagami (Nintendo) – born in 1966. Representative works include Yoshi’s Cookie, Tetris Flash (Tetris 2) and Game Boy Gallery..
The manager of Nintendo’s R&D1 division at the time, the late Gunpei Yokoi-san, directed me to “make some kind of puzzle game for the Super Famicom”, and so in April of 1994 we held an informal planning competition, but nobody came up with any interesting ideas.
Around ten days later, Yokoi-san said “I thought of something, so hear me out” and called us in for a meeting. As he went over the idea, he described something in relation to a “15-puzzle”; I had no idea what he was talking about, so I asked, what’s a 15-puzzle?, so he began drawing a picture of a 15-puzzle up on the board and suddenly, a light bulb went off in my head: the picture he drew looked like a panel sliding off the top of a stack and dropping vertically, which seemed like it could be quite fun, so I blurted out “boss, you could make a puzzle game out of quickly sliding panels into matching formations to make them disappear”, and that was the idea that became Panel de Pon.
The success of the project was largely due to the programmer being a puzzle aficionado with a good sense for puzzle games; the program he wrote early on were designed to accommodate large chains of 10 or more and techniques like lag-chain combos that we developers weren’t able to predict. I think the idea itself, the game production process and the presentation all gelled to make a really strong product.

Toshitaka Muramatsu (Intelligent Systems) – born in 196X. Participated in the design of the Famicom Fire Emblem games and Super Scope.
After being presented with the initial plans for Panel de Pon, our company was entrusted with development, but coming up with the competitive mechanics was difficult and we didn’t settle on those until quite late into development. At the time, we developers were barely capable of forming 2- or 3-chains, so we weren’t sure we could anchor a competitive system around them.
Also, the initial attack method would push up the opponent’s tiles from below, but it never really seemed to fit, and there was a part of my that was consciously trying to avoid doing anything too similar to other games. I was really struggling until the programmer suggested having those big rod-shaped objects fall from the top of the screen.
The characters are drawn in a very non-Nintendo manner, too — they were originally just temporary images. We told the female designer, “they’re eventually going to have us add Yoshi or that geezer with the mustache, so until then, go ahead and draw whatever you like”, and they ended up being adopted as-is.
Puyo Puyo
Masamitsu “Moo” Niitani – president of Compile.
After witnessing the explosive success of Tetris in 1990, we also desired to make a hit puzzle game. An employee came up with a plan for a falling-block game, and upon entrusting the concept to the Madou Monogatari4 team, Puyo Puyo came to be. The Puyo itself is a character that originated from Madou Monogatari, and it was adopted for our puzzle game at the suggestion of the staff.
It must have taken around two years to create the first Puyo Puyo game. People often comment on Puyo Puyo’s cute characters but first and foremost, we focused on the game system — we went down every path, researching every obscure game and testing every possible pattern, eliminating ideas one by one until the game that became Puyo Puyo was all that was left.
I think the game was so well-received, and the series so successful, because it established a “post-Tetris” standard for falling-block puzzle games. The reason you see so many games imitating Puyo Puyo’s competitive battle system is that it’s exceptionally well-crafted: for example, even large chains can be stuffed out by small chains that are launched before the big chain is activated, and there are all sorts of tactics you can employ on the fly by observing your opponent’s actions. I think we created something people can play for a long time without getting bored, like Go or Shogi.
When we first made this game, I thought “this game offers nothing original”, but in hindsight, I wonder if Puyo Puyo’s originality comes from broader integration of the competitive battle system and the characters into one cohesive package.
Just as with non-puzzle games, I think it’s important to be able to understand a game at a glance and immediately be able to play — that’s precisely true of Puyo Puyo, and it’s why it’s so widely played by people of all ages, from the very young to the elderly. Easy to understand, but endlessly deep: that’s Puyo Puyo.
Puzzle Bobble (Bust-A-Move)
Seiichi Nakakuki – development section #3, Taito CP Development Department; joined Taito at age 30 after switching from his former career as a designer of audio equipment. Currently active as a producer. In addition to his work on the Puzzle Bobble series, he also worked on the recent puzzle game Land Maker..
When I created Puzzle Bobble, I was picturing marbles and billiards: the game would use glittering marble-like balls, with the deliberate, angled aiming and shooting mechanic coming from billiards, and so the game began development with that image in mind.
When making a game, you’d typically write up spec documents, outline the game system and then get to work, but none of that happened for Puzzle Bobble — it felt more like the programmers and I were just casually discussing how to proceed, and the game sort of just came together.
The first version we came up with was called “Billiades”, but it really was just about hitting balls around at angles billiards-style, with no particular rules to speak of. Later on, we added the system where matching balls of the same color would make them disappear, and as we added rules little by little, Puzzle Bobble gradually came into being.
The hardest part of making the game was the schedule — we made the game in just one-and-a-half months. What’s more, the development period overlapped with Golden Week, during which time I pretty much stopped working, which I’m sure must have annoyed the hell out of the people working on the software side. Truthfully, it’s difficult to justify spending a lot of time on or assigning a lot of people to puzzle game development.

Even at the time, Puzzle Bobble seemed like an old-fashioned sort of puzzle game, as there weren’t really any stage-based puzzle games around. Honestly, I didn’t think it was going to make it to market… we sent the finished game out for a location test and the income was not where it needed to be, but even so, the game gradually started bringing in more and more money and that’s when the prevailing attitude shifted to, “this ain’t half bad!”
The “idea” for Puzzle Bobble, if you want to put it that way, is something I think anyone can could have come up with; what really matters is how well one is able to craft their idea into a game. Using Puzzle Bobble as an example, I think the attention we paid to the smaller details, like making sure the fired balls clump up neatly along an invisible grid, or choosing the next ball in the launcher based on the current situation, are what led the game to be successful.
Uo Poko
Toshiaki Tomizawa – General Manager of the Development Department at Cave. A programmer who joined the company in 1994, whose works include the Donpachi series and Panic Ringoskii (Hiroshi Sekiguchi Tokyo Friend Park attraction).
We started off making shooting games like Donpachi, but we decided to take a stab at creating something for the arcades in a new genre. I thought our agility as developers, and our idiosyncratic personality, would lend themselves to making a puzzle game, so from there, we spent the last year or so experimenting with a few different ideas.
We went through several variations in those early stages, including stuff like a “Restructuring Falling-Block Puzzle” in which you’d lop the heads off redundant employees (laughs), but Uo Poko as it currently exists began to take shape around the beginning of this year.

When you really analyze it at its core, the fundamental appeal of a video game is that almost tactile sensation you get when the game reacts to your inputs, so Uo Poko was made with an emphasis on that kind of pleasant sensory feedback. The president came up with the basic game system and input methods, but in order for players to really get hooked, you need that satisfying, tactile game feel that will make them want to play over and over again, on top of the core game loop.
I think everyone has fun memories of playing with bubbles in a bath or washtub as a child, right? I used to froth up my juice by blowing into it with a straw, and my mom would always get mad at me (laughs). I wondered if I could recreate that fun sensation of playing with bubbles, and from there, the setting for Uo Poko began to expand. Along with bubbles, there are breakable ice balls and glass balls that make a really soothing tink! as they collide or break, and because the game was bubble-themed, it was natural to set it underwater, and in that case it’d be cool to have fish swimming around, and so forth.
With that in mind, I was very particular about about the movement and sounds made by the various kinda of balls, and the movement of the fish in the background. Puzzle games certainly require you to engage the players’ minds, but figuring out how to captivate their senses is also crucial.

If you've enjoyed reading this interview and would like to be able to vote each month on what I translate, please consider supporting me on Patreon! I can't do it without your help!
Columns was based on a game created by Hewlett-Packard employee Jay Geertsen, who wrote it as a programming exercise for the X11 graphics library and later sold it to an IP lawyer; Sega of America’s Steve Hanawa allegedly discovered the game and forwarded it to Sega of Japan, who then acquired the rights.↩
An early modem & online network package, produced for Super Nintendo and Sega Genesis in North America and Super Famicom and Sega Saturn in Japan.↩