next »

[J-Snake]

Let me introduce you to this topic:

I think there are creative people on this forum. Some of you might have noticed that in the snes-era certain games brought a new direction how to experience a game.
I am especially regarding Super Metroid here. The initial "fleshing-those-ideas-out" and the capabilities of the old-hardware were very limited. But the main root was set. Now what is still missing is a game that is evolving this roots and brings complementary elements to the whole system.

Since in my mind I see so much potential in a metroid-rethought I am rediculous enough to go through the adventure in designing and implementing the game.

a bit to the background of what I am doing and what I want:

I have been in the planning and technology-build-up-phase for almost 2 years now. And there still lies much time ahead before the actual imlementation starts since everything has to be thought through in this massive set-up (the project name of the rethought is STARDUST-GENOM, btw). So in the meantime there will be several side-projects released (like the upcoming TrapThem) which also might share some technology and mechanics with Stardust Genom itself. For instance the physic-system in TrapThem will be used for a certain material-type that has the tendency to connect to each other. Here is an explanation how it works if you are interested:




What is very important to me next to other aspects is a competitive sportive-layer. It is something most games do not take very seriously. It means the game has to provide consistent and clean mechanics. It is important that the player can clearly grow and express himself through the game. For example in my designs there is no cheating around in quantities to reach a certain level of experience/skills... like in countless rpgs. If you want to achieve something respectable or beneficial you will accordingly have to overcome certain challenges. For example it can require you to master certain skills/tactics to beat a boss or to traverse certain environment.

I think the add of a sportive combat-layer will do the rethought good. Another thing that connects personal depth right into the game is a concept I call "ingame content creation".
In particular for Stardust Genom it means that there is a gun/weapon-arm which allows you to write your own weapon-programs. My concept for that is called "Smart-Ray". It allows you to program the flow of tiny rockets with can contain a heat-rope:



I will let the potential for its game-design-possibilities is left to your imagination;)

Another important aspect is the right mix of advanced and primitive enemy A.I.
The advanced A.I. should be used to make the game more alive and force tactics. The progress of the game will differ each time you play it. Random but intelligently designed encounters are important.

There are more aspects to my rethought but I think these are very important fragments of my vision I want to share with you. So if it captures your interest and you have own ideas what you would like to see in that game-type just don't hesitate and post all of them, no matter how rediculous they are. I would like to listen to your visions.

Remember that even bad ideas can often bring some inertia to get to good ideas.
You can only help.

[Danmark]

I have been in the planning and technology-build-up-phase for almost 2 years now. And there still lies much time ahead before the actual imlementation starts since everything has to be thought through in this massive set-up

While it's often exaggerated how broadly prototyping can apply, this raises a red flag. You don't need to think through the whole of an ambitious project ahead of time. If you do, you run the risk of making an unfun game. Right now you can only play it in your mind, and it'll never turn out being quite what you had imagined.

In particular for Stardust Genom it means that there is a gun/weapon-arm which allows you to write your own weapon-programs. My concept for that is called "Smart-Ray". It allows you to program the flow of tiny rockets with can contain a heat-rope:

Cool idea and demonstration. How about programmable allies? Programmable environment parts?

Another important aspect is the right mix of advanced and primitive enemy A.I.
The advanced A.I. should be used to make the game more alive and force tactics.

Nice! AI in action games is too often neglected.

[J-Snake]

My approach is to compose the game out from fundamental components/ideas. Some potential of those components is implemented first in several smaller applications which eventually can result as own games in side-projects. So it gives me a truly expansive view of its possibilities and limits. And I like this style of development (of course big companies unfortunately have not the luxury to afford it). As instance those components can be TrapThem-physics and everything related to them (like the game TrapThem itself with all its puzzles, but even more). The same also goes for smart-ray. There is easily even room for an innovative shmup-type using this weapon. I also have plans for a game called "Ray-Type" since Stardust Genom might have space-combat in it.

What do you mean with programmable allies and environment parts? Sounds interesting.

[Danmark]

^ It's an interesting approach to prototyping, and will undoubtedly yield useful results. To be more specific about my concerns, you might end up making a bunch of things that are cool on their own, but don't work well together, like a miasma of game mechanics. This is if you put off integrating them into Stardust Genom proper & testing them all together for too long.

What do you mean with programmable allies and environment parts? Sounds interesting.

Some of the earliest programming games had players program AI for competing combat robots. That made me think that the player could have a few small companion robots, with very low firepower.

Proper AI programming is well outside of the scope of your game, but setting (very dynamic) formations for flying vehicles could work similarly to your programmable projectiles. Beyond that, players could set simple policies (I hear Final Fantasy 12 had something like this). Another idea is to have a handful of explicit orders you can give the robots as a group, among them a "move" command, and the way they traverse the path to the location could be programmed. It'd be badass if they used programmable projectiles as well.

As for programmable environments, I just realized I probably stole the idea from Gunpoint; look yonder for inspiration. It could have tactical and puzzle significance.

[J-Snake]

 On a higher-level I carefully choose the components and think the integration of them through. So if they work on their own they will also work with the game. At least let it be my illusion for now;)

The environment system is  well set for proper working A.I. It is tile-based and the game-sections will be segmented. I will take the best advantage of all that to implement advanced A.I. (and I mean advanced) that doesn't need a nuclear power-plant to run. However there is a big counter-player. That is the dynamic environment. Potentially(if you have the needed gun-power) you can destroy all the environment and terraform it for your combat needs. A.I. in dynamic environment is very challenging( is that the reason why you are sceptic?).  
I have a good feeling for code, especially when it gets technical and edgy. I am sure it will help to overcome those type of challenges.

About those allies. I thought about all that but I think to have a better fitting idea. In short: You will discover alien-technology, on the first glance similar to the T-1000 from Terminator2. It is a material-system that is able to synthesize to any matter based on the local information. It means when it has all the current atomic-data with all its molecular-interconnections from a person it can instantly clone it and turn into this person with the initial state the person was scanned. You will need to collaborate with this material or ally (depends on how you look at it) to progress through the game. To a certain degree you can teach him something but he can also be controlled by the second player. The main advantage is that it can turn into any shape it scans or you program.
For example if you load a ball-shape program it will be a ball and can physically roll through small gaps.


Thanks for Gunpoint, I will take a look at that.

[Danmark]

^ That's another cool idea.

( is that the reason why you are sceptic?).

What I meant was that it's beyond the scope of the game to have the player program AI for allies.

[J-Snake]

Ah yeah, I think that too. My take on handling two players if you are playing alone is that the situations are designed to be solved in a sequential manner. It means you can switch between the main character and "T-1000" whenever needed.

I have some ideas to handle some simulataneous team-work if you are playing alone. But I don't want to make it too complicated.

Just playing with an idea:

you could assign the t-1000 some individual "helper-functions" by simply recording an input-sequence (like in training-mode of StreetFighter). I can imagine a number of situations where it could be helpfull, especially in a tech-toy heavy metroid game. As example for one simultaneous team-work action one character can use an air-pressure weapon to catapult the other player higher during a jump. The jumping player are you, the shooting player behaves according to the input-sequence.

How do you feel about this?

[baconman]

Personally, I would stick to one-player vs. the world; take sequencial stuff like that, and factor it into the level design with some nifty cause-effect action and a good sense of flow. Keep it simple, silly.


Flyer Enemies
-Staystills
-Pacers
-Orbiters
-Slow Homing
-Fast/Momentum-Based Homing (and susceptible to environmental damage)

Ground Enemies
-Pacers
-Bouncers
-Tracers
-Chasers
-Strategic Distance/Battlers (also takes relative vertical positioning/advantages into consideration)

Wall/Roof Enemies
-Staystills
-Tracers
-Flyacrossers
-Edge-travellers
-Mashers

Traps
-Thwomps
-Dart Traps
-Spinning Rays
-Spikes (and appearing/disappearing spikes, or even fake ones)
-Springs/Switches (identical-looking!)

Rhythm-Spawner Enemies
-Bullets
-Waves
-Divers
-Parabolic (ala CheepCheep)

Enemy Variations
-Basic Enemy
-Shooter Enemy
-Worm-body Enemy
-Defensive/Protected Enemy
-Retaliative/Kamikaze Enemy (explodes upon defeat, or something?)
-Surprise Enemy (looks like background/tiles, or a gameplay object)

Basic Shot Types (a)
-Direct Shot
-Ground-Travelling Shot
-Arc-Shot (think Castlevania's Axe, here)
-Spread Shot
-Spiral Shot
-Still Shot

xx

Basic Shot Effects (b)
-Piercing Shots
-Geometric/Bounce Shots
-Explosive Shots
-Boomerang Shots

*(notice, no mention of Charge Shots? I think they're overdone, personally.)

Platforming Gimmicks
-Elevators
-Pipe Mazes
-Pulley Lifts
-Gears
-Switch Blocks
-Player Cannons (ala Barrel Cannons)
-Balloons
-Currents (wind/water/magnets?)
-Conveyors

Puzzle Types
-Word/Missing Letter Puzzles
-Math Puzzles
-Logic Puzzles
-Pushblock Puzzles
-Sequencial Location Puzzles

Gameplay Objectives
-Speedrunning (and One-Shot Shortcutting?)
-Exploration/Collection
-Target Elimination
-Enemy Extermination
-Lock-And-Key Sequence
-One Big SuperSecret

Non-Hostile Locations
-Money Cache
-Powerup Cache
-Shop
-Item Altar
-Warp Zone
-Save/Respawn Points?
-Minigame?
-Advice/Clues

Environments/Biomes
-Fields/Surface
-Ancient Ruins
-Modern Cityscape
-Jungle/Forest
-Haunted Mansion/Castle
-Aquatic Area/Sunken Ship
-Industrial Area
-Casino/Carnival
-Renaissance Architecture
-Beach/Tropical
-Desert/Wild West

[J-Snake]

Hey, that was some exercise to pull the brainstorm-list out of the brain, wasn't it?

Yeah, just one player vs the environment is a valuable concern. 2 Players open up a lot of more possibilities and fun-challenges to master, but it can be a lot more difficult to integrate. I still want to do it, but to keep the right balance between the design/ implementation-challenge and the benefit many of the team-work-advantages will be optional. The most sections which do require team-work can either be solved in a sequential manner, or simultaneously by assisting individually programmed A.I (it just follows a recorded input-sequence (can add a lot of fun) ).

You mentioned the absence of charging weapons:

There is no way to convince me for that. Let me see what you think about that, perhaps one can add nice ideas to that:

Charging is one fundamental element of the coherent suit-energy/power-system itself. Not only that, it enables coherent and interesting game-design. It also adds great depth to the combat-system. Your health and the suit-energy is separated (similar to Halo). The suit-energy regenerates slowly. You can load the existing suit-energy with a certain power (energy-load/time) to your weaponarm. The energy you can store in the weapon-arm depends on how many energy cells you have found and dedicated to it (it is your decision to make the trade-off how many cells go to the suit and how many go to the weapon-arm). Certain materials only break when they are hit with a certain amount of energy at once( "quantized" property). To create certain type of shots it is not always possible to turn energy into any matter but for example you will also need fuel to create missiles. You will have cells to store fuel/water and other ingredients.

Now the magic and beauty of this suit-concept is that it can easily lead to a rich and coherent gaming-world-system which enables great flexibility and control over game-progression. I can theoretically make a coherent and completely destructable world while still keeping intended progression and non-linearity in it.
  

[XRA]

Take all the elements of the original Metroid and make the experience (overall pacing, balance) fully procedurally generated based on the rules of the original.

[J-Snake]

Fully procedural generation kills control in design, but partial procedural generation can be fine.

Procedural generation is a general statement. It can be applied to geometry, the distribution and action-design of enemies and so on.

[baconman]

I did have a fleeting thought about a different style of progression:

Start with Hop'n'Bop skills.
Explore a couple of areas and beat bosses to gain limited projectile powers.
Use those two to explore a couple more, and get melee weaponry.

New area or two changes the dynamic to focus more on melee-style combat; use that to explore more and beat a couple other bosses to get agility/mobility upgrades.

Combine those to make a more complex fighting system in the next areas, to acquire firearms. And then explosives, and finally vehicles (which all increase the scope of the accessible world, which is designed with consideration to them). Use explosives and vehicles to complete the objective of the game and escape.

[J-Snake]

I have thought about it for a while: Not bad sum up in a nutshell of a potential metroid game.

In fact I have thought about melee and vehicles. I am not sure about the melee, I think I want to keep it short but decent, as far as you can get with one-button combat games.

As for vehicles I thought of a cool hovering byke (like in wipeout), you can also jump from it in the air to reach some higher surface (like from yoshi in super mario world).

[vinheim3]

Follow the successes of previous Metroidvanias, which are:

* challenging puzzles
* huge starting area (think Link's Awakening) blocked off by "doors" that you need "upgrades" to pass
* each "upgrade" gives you at least 5 new areas to explore, all the way through to the last upgrade
* make sure your upgrades are actually interesting and in some way unique
* don't make your game linear towards the end. This is the problem with most Metroidvanias. The hardest path leads to the end-boss, or there are no more upgrades after the end-boss. I sometimes don't complete Metroidvanias because once everything but that final area is done, I consider it done and quit

[J-Snake]

I think some partial linearity set-up by the designer is good as long as it is constructed by the consistent rule-set of the game (no invisible walls and magical unlock-mechanisms).
I think I can achieve that by my weapon and physics-system.  

As example in my side-project (action-puzzler TrapThem) the rule-set is flexible enough to even allow on-rails-action-level-construction on its own without breaking the rules or adding temporal fake blocking elements. I strongly believe that makes up for the beauty of the game. So I am working really hard towards such systems. This is how I want to make and experience games.  

Here is an example:

[DangerMomentum]

What aspects of Metroid do you want to refine and focus on? I don't see any cohesion in the ideas posted- it seems more like a "do everything" project that will end up having a lot left on the cutting floor. One of Metroid's strengths is how streamlined it is- you rarely hit solid boundaries, and the player constantly pushes those boundaries until they find ways to break them open. This continues until the player has access to the entire map. I think too many Metroidvanias focus on "Interconnected rooms!" and don't pay any attention to what really made Metroid special- how you traverse these rooms, and how much freedom you have. Are you supposed to go up that chasm now? Nope. Can you do it with some patience and clever bomb use? Yup! There are very few areas where you just hit a brick wall. Instead, the game tends to use its difficulty curve to steer players. Is progressing in this direction too difficult? You should probably be looking somewhere else, then. I haven't really seen games labeled as Metroidvanias do this correctly often at all. Its the reason Metroid games always have huge speed run / sequence breaking audiences- the games are perfectly suited to this kind of play.

[J-Snake]

I want to refine and/or to keep quality in 5 things I have clearly in mind:

1. the combat-system needs to be more advanced

2. some sort of user-generated content creation which allows to express yourself
   individually (here Smart-Ray)

3. I want more sophisticated physical interaction with the world, using the environment
    and your abilities to traverse space and to handle combat-situations in a more
    interesting way.

4. A proper amount of randomness, some encounters you can never exactly predict, like an
   intelligent hunter is searching you in the entire complex(like SaX) based on A.I.,
   less on scripted events. I prefer it to be that way than to suit the needs for the
   speed-run audience.

5. I want to keep a decent amount of exploration in an open and consistent world.
    There should be no artificial borders. Progression is based on physical interactions,
    less on triggers.

The world in Metroid is a system of interconnected rooms. Of course it is the challenge of the designer how to wire them. Yes, skill vs detour trade-offs should be a clear concept to every good designer.

[J-Snake]

As example of point 3:
There is no requirement of Grapple-Beam to kill a boss in Super Metroid. But one can make it happen, and how! The tech-toys you use only scratch the surface of what could be done. And I am a fan of a reduced feature set (example: instead of 100 weapons you just have 6), but which allows you to exploit its potential to the fullest.

Perhaps TrapThem will introduce interested people a bit more to my thinking. The feature-set is low, but it is fully exploited. It results in great control over level-design and great variation in way of play. That is what I love and that is my sense for aesthetics.

[baconman]

Capture the elements and essence of what people like myself love from each different type of game, and create a cohesive "gaming world" using them. The "complete gamer experience" so to say. Combine this with Procedural Generation and actually solid level design and gameplay principles/mechanics (as well as variations on character abilities) in such a way that players then have dozens of interesting and uniquely structured worlds in which to experience this, and ways to go about it (since your own abilities change as well).

[DangerMomentum]

Combine this with Procedural Generation and actually solid level design

Wait, huh? I don't think procedural generation is a good idea for a Metroid like. The game relys far too heavily on the quality of the world.