Xexyz / 亀の恩返し ウラシマ伝説 redesign project

Public release: Feb 1, 2009.

Update: Feb 2, 2009 - New base building 'screenshot' in the game design section.
Noodled on the pinup image in the Foreword .

Update: Feb 3, 2009 - Wrote a text wall about weapon diversity at the bottom.
Updated Index with more anchors.


This NES/Famicom Adventure/Shooter game could have been a classic if it had been polished more. It's a bit whimsical (in particular the Japanese version), with fairies, robots, mecha lobster jet vehicles, etc. The first level is quite fun as it offers some exploration, shops, mini games, quirky encounters, etc. I also like the drop ships which sprinkle enemies over the player. Such a thing contributes to fleshing out the universe (explaining military logistics of the enemies).

The structure of the game becomes monotone and tedious because it repeats with not much new content being added. The constant barrage of superficially new enemy designs makes it hard to get attached to anything. The simplistic behaviour patterns doesn't help, it makes everything feel flimsy and mass-less. The same can be said about the controls for the player. That said, the designs for both the enemies and settings are pretty good.

My idea is to take the most memorable designs and make a consistent flora and fauna. I think there should be a few more common and diverse enemies. Star Wars has the Storm Troopers and SMB has the Koopa Troopas. They act as anchors, and feel less like temporary cardboard targets because they are capable to fit into different roles and different environments.

Some redesigns of mine. The helmet on the original design seems to have been inspired by the Solnoids(?) from the animé 'Gall Force', or perhaps Bubblegum Crisis. It was released a few years before... 'Kame no Ongaeshi - Urashima Densetsu' (a.k.a. Xexyz). I chose to band the armour because it's called the 'Brigandine'. Well, a Brigandine has the armour scale/bands on the inside of the cloth, but whatever.

Just a few pencil sketches. I might take some liberties when it comes to scale. Even though some of the original sprites are kind of strange. I think it's possible to make a fun design of almost anything as long as the silhouette is striking, and it mostly is with old sprites because at that scale it has to be.

Many of the enemies are based on some kind of animal type, like bugs, fishes, reptiles, etc. I think all of the enemies are robotic since their faction is called The Machine Beast Empire. As mentioned in the foreword, I'd like to pick a few of my favourite units and flesh them out a bit.



I'd classify the game as 'adventure shooter'. Its setting features Sci-fi with robots and guns meets Fantasy with fable animals and magic. The story loans parts from Japanese mythology (Urashima Tarō) and culture (stuff you find in the ocean).

Title and release

I found this information about the Japanese famicom version.

The NES version was released in March, 1990 and was stripped of the Turtle story (and thus the game intro). It was also given a new title, Xexyz.

Player avatar upgrades

Japanese story

This is the story from the intro of the game. I'm not sure what the manual says. The hero of the story is a guy named Ken Urashima, and he's a descendant of the legendary Urashima Tarou. While standing on a polluted beach, pondering... the pollution, he spots a crab tormenting a turtle (both rather large). Ken, being somewhat of a wizard, nukes the crab. The grateful Turtle tells Ken to ride on his back, because apparently Ken is the right man for some task which lies ahead.

So they fly through some kind of warp space and end up in the Dragon Kingdom. The turtle explains that they have been invaded by the 'Machine Beast Empire' and are in dire need of help. So, the turtle gives him the 'Brigandine' armour, and Ken sets out to save the princess 'Otohime'.

On his journey he meets a bunch of talking dogs, frogs, a ghost king, fairies, a collection of princesses, silly girls, Blue Landers (except they're probably something else), and other strange stuff. It ends with Ken saving all the princesses, piloting a giant turtle ship and destroying a Space Fortress. For his trouble he gets a present, a so called 'tamatebako'. If he opens it he returns to the beach (minus the pollution) and a girl looking just like Otohime is promptly asking him to marry her. If he doesn't open the box he gets to marry the real Otohime and becomes the king of the Dragon Kingdom. If he throws it away he's returned to the beach (minus the pollution) and then he's turned into a little ball. Bad end.

English Instruction booklet story

In the year 2777, the earth existed in a totally different atmosphere. Due to nuclear wars and natural disasters, the population decreased dramatically and only five small islands existed on the surface of the earth. Those five islands formed a nation and for hundreds of years people lived peacefully.

The new nation is called XEXYZ. At XEXYZ, aside from human beings, there exist fairies and spirits who lived under the ruling of King Xeu Star. But one day, XEXYZ was attacked by the mechanic fortress, GORUZA, from outer space. XEXYZ did not have many weapons to fight back and was forced to surrender. GORUZA built mechanic castles on each island, captured a queen from each island and confined them. King Xeu Star was murdered and his only daughter, Maria Star, has been missing since then. Battle-soldier Apollo could not endure the cruel treatment of GORUZA any longer. So he put on his space battle suit and challenged the mechanic beasts' empire!

Story setup

This story is a mix between the Famicom and NES version, and some ideas of my own.

There are two relevant factions, the Island Kingdom and the Beast Empire. They inhabit two separate, but linked Universes.

In Universe 1 we have the Island Kingdom (IK) faction. They are mostly fable animals and robots, and they only inhabit a single water covered planet.

In Universe 2, the Machine Beast Empire (MBE) are involved in a large scale galactic war (Legend of the Galactic Heroes scale). They are confined to a section of their galaxy, so they're not overwhelmingly powerful. If the MBE were really powerful, taking out the IK wouldn't be a problem. Since they have exhausted the resources of their galactic domains, they are in trouble. They find their way to the dimension of the IK, hoping to find more land and resources to conquer. However, there is a problem. The warp space leading to IK is riddled with some kind of mines looking like spiky asteroids. Only small ships can get through.

Still, the MBE is successfully attacking the IK. All islands but one has fallen. This last island is where the resistance forms. It is defended by turrets (some look like statues) and soldiers. There's a large central castle which acts as a lounge.

On the surrounding conquered islands, the MBE are busy building a shipyard, mines and other facilities. Their primary concern is to purge the warp corridor of mines though. If they succeed, they can take thousands of battleships through it.

The IK is concerned, and decides to infiltrate the facilities on the MBE occupied islands. Unfortunately, their attempt to contact a certain hero in another Universe (3) fails. They sent out a Turtle agent, but it got assassinated by a spy (a Crab) sent out by the MBE. Now they have to manage on their own.


10 000y BP (Before Present) Universe 1, planet Aquarius (later known as the Island Kingdom) - An ancient civilization flourishes, and suddenly disappears. They leave behind ancient underground ruins filled with traps, artifacts, robots and cryogenically frozen Princesses, as well as fairy-like information units. It is speculated that the ancients dug too deep and were annihilated by something even more ancient which had been buried.

500y BP Universe 2, planet Turtoss - The technologically adept Turtles are fleeing from superior enemy forces. They construct a device which generates a Warp Corridor into Universe 1 and the planet Aquarius. To stop the enemy from following them, they mined the warp corridor and built a defensive weapon which was capable of dealing with anything coming through the portal (perhaps by taking advantage of the special physics inside the Warp Corridor.).

The Turtles brought along other aliens such as the Frogs (greedy physicians), Dogs (tinkerers) and Robots (traders). However, on planet Aquarius there's a big conflict resulting in cultural memory loss. They regress to more primitive levels, but manage to get along. The Robots lose their memory because their AI-OS is poorly made and they have to reformat every few years or so. In Universe 2 the remaining Turtles are annihilated.

250y BP After several dark centuries, the unfortunate immigrants on planets Aquarius finds the hologram of the Dragon King. Although he's become a bit strange due to memory degradation, he still manages to guide the immigrants to a brighter future. The Island Kingdom is formed. All hail the Dragon King!

1y BP In Universe 2, the Machine Beast Empire is now being cornered by the same enemy who attacked the Turtles half a millennia ago. The MBE find the portal which was created by the Turtles, and promptly invades planet Aquarius in Universe 1. Although the Warp corridor has been mined and the MBE can't bring any large forces through, The IK'ers are still forced to fall back to their main island where they make a defensive stronghold.

6m BP The MBE finds one Princess on each of the islands which they occupy. They persuade the princesses to join their cause. Soon after the Turtles find Otohime on their remaining island. The Princesses don't remember much because they have been frozen for so long. A number of maidens (some of which become the 'bathtub tribe'), fairies (robotic information units) are also found. Smaller MBE forces pour through the Warp corridor.

0 The Dragon King starts to realize who the Princesses are, and encourages the IK'ers send Otohime on a mission to the occupied islands. Game start.

In-game story

This is the structure repeated for each island in the original game. I'm not sure whether the islands are underwater, floating in the air, or just regular islands.

I like the idea of a central hub which you return to with your loot. This could be the center island. I would have defenses against the MBE because it's constantly under siege. On the surrounding islands the MBE are working to circumvent the defences by building war machines and doing research. Each island has of a number of regular enemies, an evil Princess (+ Cyborg Beast), and a large monster boss.

The player would have to use a Cyborg Beast to fly between the islands. Perhaps a finite fuel or ammo supply could be used to force the player to explore on foot even when in open terrain. After the player has found a way into the castle and defeated the Princess (and made her dance?), another boss fight with a large mechanical monster awaits. It's trying to break through by opening a new warp corridor.

Eventually only one warp corridor remains, and the MBE focuses all of their resources on it and eventually breaks through. Meanwhile the IK'ers has found a way to deal with whatever could be coming through the warp corridor...

A number of sea star shaped keys can be found scattered around the IK. They are needed to activate an ancient weapon which is designed to fry anything inside the warp corridors. Some keys could be guarded by golems / statues which were made by the Turtles long ago. This setup was meant to prevent the weapon from falling into the wrong hands, but of course now the present day Turtles and IK'ers have no idea that they built the golems in the first place.

End-game story

The IK'ers eventually rediscover their secret defensive weapon and readies to deploy it just as the warp corridor is cleared of mines by the MBE.

However, the IK'ers now discover that they had come from Universe 2 and are not native to Aquarius. When they realize that their ancestors might have been in the same position as the MBE, they start to feel sympathy for the MBE. It also turns out that there people in the MBE who are dissatisfied with their commander. After the evil MBE commander is assassinated, the IK decides to let the MBE through. The MBE agrees to withdraw their killer robots.

It is also discovered that the last time several factions from Universe 2 had tried to get along in universe 1 (half a millennia ago), things went badly. Now it remains to be seen if the MBE can be trusted.

Meanwhile, the battle against the enemy in Universe 2 is going badly. Plans to send transports through the warp corridor are made. Then, a MBE science officer finds a hidden document in the ancient Turtle ruins in the MBE dimension. The MBE officer is unable to read it, so he brings it to the IK. It turns out that half a millennia ago the Turtles left behind on planet Turtoss in Universe 2 had found a way to defeat the enemy, but they had perished just before finishing.

Sequel: The IK and MBE gets along, but the enemy from Universe 2 has realized that the IK'ers has a weapon against them, so they shut down the Warp corridor and uses a spy to activate the thing which wiped out the ancients on planet Aquarius.

The Princesses

When researching Hudson Soft I stumbled upon 'Galaxy Fraulein Yuna'. It has some ideas which could be put into use here. Basically Yuna (the heroine) has this ability to convert evil characters into good characters. Perhaps the various Princesses start out evil and has to be won over. Gameplay-wise that could work a bit like defeating Megaman bosses, but instead of blowing them into smithereens you persuade them to join you with an act of compassion.

Looking at the sprites from the game, I can see the many types of robotic animals. Each Princess could have a type of animal as a design theme for her grunts, as well as for the Cyborg Beast.

The princesses appearing in the game had certain sea food themes which could be built upon.

Princess Wakame: A form of seaweed which looks like ribbons (dark green or brown?).
Association: Ribbons, Dry or soaked, duality, inflates when soaked (temper?), beauty treatments, fat burning.
Character type: Perhaps a beauty obsessed person who works out a lot, but loses temper fast and turns into a different person. Always advocating her lifestyle as superior.
Home island: Ruins.
Cyborg Beast: Some sort of Catfish.

Princess Marimo: A green algae ball (moss ball).
Association: Goes with the flow, unharmed.
Character type: The slow and calm person. Shiori from Galaxy Fraulein Yuna managed to pull this off very nicely without being boring. In fact, she was the highlight of the series.
Home island: Forest.
Cyborg Beast: Unicorn Dolphin.

Princess Sango: Coral.
Association: Hard but soft, and brittle in a way.
Character type: The tough and cold as ice person. With shameful secret, like a strange fetish. E.g. Natsuki from Mai Hime.
Home island: Crystal or glass.
Cyborg Beast: Dragon.

Princess Sazae: A Turban shell mollusk.
Association: Retracts into shell, defensive.
Character type:
Home island: Lake.
Cyborg Beast: Crayfish or Lobster.

Princess Shinju: Pearl.
Association: Rich. Persistent or patient.
Character type: The ambushing and planning person with a plan that never succeeds, triggering outbursts. E.g. B-ko from Project A-Ko. Tron Bonne from Megaman Legends is another example. These characters usually have a lot of resources at their disposal.
Home island: Volcano.
Cyborg Beast: Turtle.

Otohime XVIII: The 18th princess of the royal dragon dynasty.
Character type:
Home island: Ruins.
Cyborg Beast: I think she has a large ship, but perhaps she could use the mobile suits.

Game design

Xexyz was quite varied with the vehicle segments, exploration, power ups, dialogs, etc. It wanted to be a lot of things, but fell short because none of the segments had much depth or polish. That said, I think it presented a lot of fun components to play with. To me the universe is always more interesting than the actual gameplay. The universe can be reincarnated in one of the following game types:

2.5D Side view, Action, jumping, flying and shooting. - The advantage with this game type is that the jump 'game verb' can be quite elaborate. We are working with two dimensions, but the Y dimension has the gravity physics, making it different from the X dimension.

2.5D Top view, Action, 360 degree shooting, exploration. - In this case the X and Y dimension has the same physics. Jumping doesn't work well in top view games, so I'd concentrate on shooting and exploration if using this game type.

3D Third person, exploration, tactics. - The advantage here is presentation, I think. 3D is descriptive and you can really flesh out your designs. It's really tempting to do so, and it looks great. However, I'm not a fan of this perspective because of the inability to align your shots or jumps with other objects in the playfield. I'm not a fan of Z-lock, and platform jumping can never be as controlled as in a side view game. There are other things to games than elaborate shooting and jumping though. In a TBS RPG game your character is in control of locomotion and attacks. In an exploration game a rich 3D environment can offer a lot.

2.5D Top view, X-Com. - A lot of the designs from Xexyz could be adapted to fit into an X-Com type of game. There are power suits, ships, upgrades, aliens.

Build and manage bases.

Manage a team of soldiers and vehicles.

Intercept and take down enemy troop ships with your fighter ships.

Then engage the survivors at randomly generated crash sites.

Capture enemies and artifacts. Bring home these new technologies and do reverse engineer them. The [E] blocks could work as funds. Perhaps if the player does well in protecting the main island, the other IK'ers will divert more resources to the defense program (a bit like funding in X-Com).

Explore randomly generated ancient ruin sites. Bring home new recruits (the bath tubs / noodle cups might be cryo pods).

As you explore new territory, gain competence and develop new scanning equipment, vital story missions begins to show up on the map. E.g.: Raiding the bases on the surrounding islands, fighting bosses, sabotaging construction sites, intercepting transports or surviving a siege. Certain story missions only become available once other story missions have been completed (since they're based on earlier developments).

Time could also be a factor in unlocking new missions, and perhaps the MBE gets more powerful over time, although I'm not a fan of time limits (and certainly not rubber-banding).

Yes, an X-Com style game could be very complex and hard to get into, but it doesn't have to be that way. I think Accessibility (Casual?) vs. Complexity (Hardcore?) is a 'false choice' fallacy. A game can for example have certain accessible default options (such as auto equipping characters), as well as more detailed... optional options.

I'm really not a fan of games which try to be accessible and 'fun'. Games designed around 'fun' often jumps right into the closest local optimum, which tends to be pretty shallow and exhaustible for a Strategy / Simulation fan like me. I think an X-Com style game could have a replay value which lasts for decades, unlike games which offer a mere '40 hours of gameplay'. I think there's a neglected market (me! me!) which needs to be satisfied too.

The environment needs to be varied if there's to be any replayability. I've seen a few games which failed because of of boring monotone environments. First, the geometrical structure of the environment needs to be readable, and I think it's worth to make the environment a bit blocky to facilitate this. Then, the environment needs to be sensible, i.e. ... make sense in a figurative, architectural way. The environment should be varied, which means there needs to be a certain amount of detail and new content for each map. Finally, the environment should promote tactics.

One thing which makes X-com work is that the environments provide points of reference and scale. I'm talking about the use of real life objects like trees, buildings, civilians, and even gas stations. We know how these objects are likely to interact with our... explosive devices and it makes us giddy with anticipation. Therefore I think it's important to include the following environments: City/Town/Farm, Base (player or enemy), Ruins (overground and underground), Wild forest/nature, Wasteland.

The available tech should also be varied so the characters can be given some individuality. Variation comes in two flavors. There can be many weapon variants with minor stat differences, and/or there can be variation in terms of radically different behaviour. Another thing which can be done to promote replay value is to have certain sets of enemies which aren't always used in the game. Some enemies carry unique technologies (which the player will capture and reverse engineer). This way each game will give the player a limited amount of technologies. On top of this resources can be so limited so it'll be difficult or impossible to fully tech up everything, forcing the player to specialize (on apples or oranges).

Yes, it could be argued that it's wasteful to only using parts of the content for each full game, but, that's only true if the game has no replay value. If the game can be replayed 10 times, then that is actually the full game, so the content can be stretched out to cover 10 replays. This will give the player "Wow, I haven't seen this before!" - moments, and that will keep him coming for more even if the content is actually exhausted (which the player won't know unless he uses a game guide).

The end game technologies could be pretty radical in how they alter the game. I think it would be a good idea to only give the player one or two of these: Mind Control, Mind Bomb, Super Soldier Training Program, Force Fields, Super Gun, Super Sensors/Senses, Teleportation.

Regarding difficulty settings, I find it annoying when the easy or hard settings doesn't offer the resistance advertised. Sometimes easy can still be difficult for a beginner, and hard can be very easy for a veteran. Another thing which often bothers me with difficulty settings is when the game changes enemy stats or even how the game works. It makes the universe feel... less tangible, and more like a game which adapts to me. So, my solution there would be to remove certain enemies rather than changing them. This solution has two benefits: First, the player will be rewarded with more content on a higher difficulty. Second, since enemies bring technology which can be reverse engineered, players will also be able to find better tech at a higher difficulty. This technology which will change the tactical situation, thus making the game different without actually changing stats or physics.

For example, on a low difficulty the player may play fewer mission, run into more basic enemies and also have more basic, plain shooting weapons which are simpler to use. On a higher difficulty, the player will be able to develop more advanced quirky weapons. The player will have more options and the complexity will be higher. This should not be a problem since veteran players are already familiar with the basic stuff and won't be overwhelmed. This gives the game scalability. Perhaps, on the easier difficulties there are even AI functions for base building, battles, equipping, etc. With the lower complexity it's more feasible to make AI work for the player side. Once you start throwing in weapons which need to be timed or used in conjunction with other gadgets, things get tricky on the AI side.

Code design

I already have some of the classes worked out, but before I go into those, let me go into my philosophy.

I believe in designing a game around emergence. First you make a set of physics, functions and methods, (doesn't have to be real physics, just a consistent set of rules which apply to all objects within the game world). Then you see which game verbs that you can easily get from that. Then, things like weapons would be constructed from materials and physics within the world.

Some might say. "What if the weapons end up unbalanced? Isn't it faster and easier to just hardcode everything?" Well, yes, if there only are a few weapons, perhaps. But I believe that a good system is self balancing, and if you're dealing with a lot of weapons then it would be a pain in the but too balance all of those against each other. It would also be impossible to have generated weapons. In real life, some weapons are poorly made, and people don't use those. If bad weapons were generated inside a game, then those wouldn't be used either. It's not like they cost anything to develop in the first place. I also like the idea of 'buffers' with less common material, it makes the universe seem less finite.

So, the class structure which I have worked out basically works like this: There are a number of technologies. An item in the game is made up by a number of technologies. In the case of a Plasma Rifle, it might be made from a 'Duralium plating', 'Plasma coils', 'Crystal processors', etc. These technologies would be referenced when the weapon is used (or damaged). If the weapon is examined at a lab by the player then there'd be some kind of research bonus for the technologies involved (assuming he doesn't have them already). As the player gains more technologies new weapons become available for construction (based on the technologies they are made from).

This approach kills many birds with one stone since the items always know what they are made of. It also makes it easy to add, say a new plasma rifle into it the game. The technologies could of course take scale parameters to allow for variations in power or durability.

A character would of course be equipped with a number of items, such as armour, weapons and gadgets. To save memory, there would have to be items which are archetypes and not real physical items, and then individual instances of items which can have a name, be damaged, and have a bunch of other variables.

The map cells would also be able to hold items (to allow for items being dropped). The home base could either store items physically on its map, or just keep them in some hidden list. If the former is used, then it would be possible to use the general gameplay mechanics for equipping the characters before they go on a mission. It might be too slow and cumbersome for some though.

Sponsoring diversity

Of course, it would be great if the original release of the game was varied enough to last a decade, but nowadays it's pretty easy to distribute patches and content packs. With the system described above, it would be easy to add new content such as new weapons and enemies. The new content would use functions / methods which are probably already implemented in the technology classes.

Possible drawbacks with content packs: the universe could feel less tangible and more like it's a subject to the whims of some designer. The unity could also suffer and the game could end up feeling haphazard. On the other hand, I'm not really bothered when say Games Workshop adds a whole new race to the universe. Also, 'all-over-the-place'ness can be playful and fun, it makes you feel that there are a lot of paths left to explore and anything can happen.

One advantage of staying within the confines of the game world physics is that it doesn't create problems like: "Why can't I do this or that when it's possible to do this other thing?" and "This weapon breaks the physics in order to avoid an exploit (because the physics didn't sponsor diversity or because the player path control 'needed' to be strict).".

Of course, the crux to making a good physics system is making sure it sponsors diversity and not singularities. Weapon-wise, I think this can be done by making sure that not only the weapons are varied, but also the armours, character (stats) and environments:

Footnote: A possible drawback to an "it's always possible to inflict some damage" system is that people may prefer to grind. For example, imagine that Link (from Zelda) could walk into Ganon's castle by stabbing the entrance 50 000 times. Some players may choose that route even though it takes less time to collect all the items, and even though it's less fun to perform such a monotone task. I'm guessing that the monotone solution is tempting to use because it is simple and that takes precedence.

So, it has to appear simpler and more fun to take the path which the game designer has sprinkled with all the fun. The solution could be to put pressure on the player. It can be relatively dangerous enemies or a practical time constraint which forces the player away. Force is not the only method to use though. The game could tease and lure the player away from the grind by clearly showing less difficult paths along with some shiny reward. "Are you really sure that you want to headbutt your way through this wall? Look how easy it is to just go around it! And look, there are a few nice items to pick up if you go this way too. At the least they'd make the headbutting easier for you."

- By Niklas Jansson, 2008-2009. - Xexyz is (c) Hudson Soft.
- With thanks to Mac Tuitui, Liam Newton and ShihTzu who provided translation help and reference material.