Advertisement

3d Tactical RPG

Started by November 24, 2003 11:44 AM
8 comments, last by PnP Bios 21 years, 2 months ago
I am setting out to develop a 3d tactical RPG. Don''t laugh just yet. I have the experience. I have been coding in C/C++ for several years. I use OpenGL, I have designed and made several small games. This will be my first BIG GAME. I have already made a small DOS Tactical RPG, but it was crap. I already do have a small amount of experience in the area, But since this one will be 3d, I just wanted to throw out some Ideas. A level editor, preferably one styled like THPS2 for dreamcast. It was easy to use and I could quickly build fun levels. I also liked how you could rotate and drop pieces, along with editable skins. How should I handle movment? Should terrain height be an issue? Should I use a system similar to Shining Force, or use one where you can move freely, and it just counts the tiles against you? I am hoping to write an engine, then a whole bunch of data files. This first one will be mideval based (knights, horses...). I was wondering, if I should do a unique method of combat. HP vs AT seems a little overdone. Should I incorperate skills and Magic? What if I used a unique grid method. I don''t think a person would only take a couple of steps an hour. I think the grid should be supersized, that way magic and projectile weapons could be more realistic, such as wind knocking a an arrow off course. these are just a bunch of wild ideas, nobody has to answer all my proposals, but some feedback on individual points would be nice. sorry bout the length. ____________________________________________________________ The Santa Claus in the Lobby wants to kill me.
HxRender | Cornerstone SDL TutorialsCurrently picking on: Hedos, Programmer One
My main beef with all tile-based strategy games is time. I''ve seen two different ways to work out character speed and action turns: There''s the "action point" system, typified by Fallout, and the "move/act" system typified by Final Fantasy Tactics.

In the "action point" system, you have a set number of action point to expend in your turn. You can move or act with them, but when they''re gone, your turn ends. In fallout, any you save are added to your armor class, to approximate a defensive stance. In this system, I can move, shoot, or interact with the environment as long as my points last. You seldom have more than eight points, and since something like firing a pistol takes about five, the turns represent fairly small chunks of time.

The problem with this system is that it allows an incredible amount of what I would call "cheap moves". For instance, when I''m faced with an enemy that''s stronger than I am, I can hide just around a corner. At the start of my turn, I take a step out, fire and round, and return to my hiding spot. I get to take a shot at the other guy, and although I am momentarily exposed, he will NEVER get a chance to shoot at me. I can nickel and dime even the toughest adversary to death, and he''ll never touch me unless he is possessed of sufficient speed to actually run through my hail of bullets and engage me point-blank, and even then there are complications with whick I will not bore you.

With the "move/act" system, you have two distinct phases in your action turn. You get a chance to move to any point within your character''s movement radius, and you get a chance to perform any action available to your character. This system eliminates the problem I described with the action points because you must either end or begin your turn in the location in which you performed your last action, so if you''re "prairie dogging" from behind cover, you''re obliged to spend some amount of time exposed to return fire.

The shortcoming here is time. Under this ruleset, taking one step or taking five occupies the same boolean value in determining the term of your turn. If I pick up an item that''s adjacent to my character or if I run a full eight grids to retrieve one, my turn is over. The action point system would have avoided this by leaving residual APs available to the first scenario''s actor.

Both systems have strengths and weaknesses, and more than I have described in this poor summary. However, in your initial post you described your idea as "tactical". That word has been used and abused in more than one medium in recent years, but I recommend to you that you use a system of heuristics and algorithms, rather than pure commands. And of course time should be a factor, for duration effects and other processes.

Here is my suggestion. It is primitive, but it will serve better to communicate my ideas to you than a discourse on theory. Bring back the old round system, like the one seen in ancient RPGs. In Final Fantasy (the first one on the NES), you assigned each member of your team an action and a target before the action commenced. When you hit "okay", all the actions would take place according to your orders and the circumstances of the battle. If both your fighter and your black belt attacked the same beast, and the fighter killed it, the black belt''s attack would do nothing, and the game would declare it "ineffective". In all honesty, that system was more plausible than the more popular "next valid target" system. After all, if two men engage a creature, and the first kills it, the second will hardly be able to re-engage and manage a hit on a different opponent in the time between the first one''s death and the timely end of his planned attack. He''d have to start a whole new attack.

And so I recommend a system by which orders are given to characters in a "planning phase" before each automated "action phase". Tell your cleric to cast heal on your archer, who is aimed at an enemy thief who is crouching behind a log. If the thief''s action commands require him to expose himself, then your archer will take the shot. If not, your archer will hold his fire. The heal spell will happen regardless. Meanwhile, your knight is duelling an enemy gladiator mano-a-mano, and your command is to continue to fight to the best of his ability. To tip the scales, you tell your wizard to cast "flame sword" on his weapon, augmenting his attack power, and then "stone skin" to up his defensive capabilities. Your enemy, seeing his thief in danger, tells his wizard to cast "gale" in a direction perpendicular to your archer''s line of fire, and then orders his thief to dash across the open territory and bring his "back stab" skill to bear on your wizard. The gladiator continues to fight, using a special "poison blade" technique that can only be used once per battle, which will guarantee a strong hit if it is successful.

When the action phase begins, all activities are automated, and can be replayed to the player''s taste. Perhaps the gale spell ruins the archer''s aim, and the thief successfully reaches and kills the wizard right after the fire sword spell is cast. Stone skin, however, is interrupted, and so the gladiator''s wenom blade penetrates the knight''s defenses, while the knight''s flame sword deals a grievous injury to the gladiator.

So, back to a planning phase. The wizard is dead, the archer is healed, the knight is poisoned and the gladiator is wounded. Both leaders, be they human or CPU, plan their strategy for the next round.

That''s the gist of my idea. Movement will be determined by the character''s speed, and modified by terrain, gear, injury, etc. Action will probably be on an action point system, so a number of maneuvers or spells can be conducted in a reasonable amount of time.
Advertisement
thanx for the reply! I will keep this under consideration.

____________________________________________________________
The Santa Claus in the Lobby wants to kill me.
HxRender | Cornerstone SDL TutorialsCurrently picking on: Hedos, Programmer One
I forget which game i saw do this system, but it was pseudo real time. What it was was every 5 seconds or so, the game would pause long enough for you to tell your characters what to do, then when it unpaused they would do their things and the enemies would be doing their''s. So all actions that were assigned were all going on at the same time. Pause the game again, reassign new actions, and unpause it and watch the mayhem continue. I believe you could also manipulate which actions a character was performing in real time. Perhaps it was baulders gate, i''m not sure. It was a pretty effective system for realistic combat. In reality people don''t take turns slashing at each other. They both slash at each other at the same time, while parrying and dodging each others hits if they have the proper instincts for that.
In this system you could have your character hiding behind something while the game was paused, assign the action to shoot at the enemy, unpause the game and he begins shooting while he is exposed to enemy attack, then you can tell him to move back behind the wall while the game is unpaused still or pause it to rethink your strategy a little.
Just a little thought that could innovate this system, instead of pausing it completely, the game just switches to a bullet time like speed whenever you want to assign commands. This keeps it moving at a good pace and makes tactical decisions more like what real tactical decisions would be like.
"The human mind is limited only by the bounds which we impose upon ourselves." -iNfuSeD
best tactical rpg
Freedom Force

also commandos too
3D Side-Scroller game demo Project-X2 "playable"Lashkar: A 3D Game & Simulation Project demo @ lashkar.berlios.de
UFO: Aftermath has a real-time system that pauses whenever something happens which means that it never becomes a question of how fast reflexes the player has.

That system works really well in my opinion (the game suffers from slight interface problems though) and is similiar to what I would like my RPG to play like when it''s done (only I will allow more actions/moves, such as leaning around corners and going prone).

Oh, and if you''ve played Baldur''s Gate and think "yeah right, I played with a system like that and it sucked/was boring!", UFO: Aftermath''s combat system is MUCH better designed.

------------------
"Kaka e gott" - Me
Current project: An RPG with tactical, real-time combat with a realistic damage system, and randomly generated world and dialogue.
------------------"Kaka e gott" - Me
Advertisement
One fairly effective way to solve the FFT movement conundrum is to utilize the "leftover action points" system to give a certain value for ending the turn without moving the maximum amount or attacking, etc. Front Mission 3 for the PS1 does a good job of this - each character gets a certain amount of AP, which recharges at a fixed rate every turn. The amount of AP is judged by the pilot''s experience. Each square moved takes 1 AP, and actions take a variable amount of AP.

Therefore, moving less than your maximum amount of squares conserves AP, which leaves room for counterattacks (which also cost AP). Also, you might not be able to attack if you move the max range. This also allows for multiple counters if you sat there doing nothing. However, you can only move once per turn. Finally, extremely powerful weapons cost massive AP to fire, and can cause the ability to move and fire in a timely fashoin to be restricted.

Unlike FFT, the turn order is fixed. This is an important decision that must be made - if leftover AP is used to increase defense or use counterattacks, then it is too complicated and unfair to allow leftover AP to also allow turns to come faster. This would likely create a situation in which the character would sit and do nothing, counterattacking constantly while gaining their turn in seconds.

Things to ponder.

In each of us there is a force that compels us to act selfishly, no matter the consequences. It is the force that will eventually cause our downfall. It is the ID - Thanatos.
In each of us there is a force that compels us to act selfishly, no matter the consequences. It is the force that will eventually cause our downfall. It is the ID - Thanatos.
Sorry for the double post, but I forgot to say something fairly important.

If you are designing a FFT-style game, it would be a better method to break from the unit square bondage. Allowing for floating points on a cartesian plane to represent the positions of a character. This allows all elements of a tactical engagement to be used more effectively.

1) Floating point coordinates allow a circle to be cast which allows the character to move the same distance EXACTLY in every direction - discounting obstacles and terrain.

2) Same thing with attacking - not being able to attack an enemy because of grid math is frustrating. Case in point: area effects and FFT''s lancer.

3) Utilizing cover is mouch more intuitive. Instead of worrying about which of three squares around an object to stand in, a character can utilize the computer''s method of a straight line comparison between positions to use the cover of an object that is far away simply by visual judging, not - once again - grid math.

In each of us there is a force that compels us to act selfishly, no matter the consequences. It is the force that will eventually cause our downfall. It is the ID - Thanatos.
In each of us there is a force that compels us to act selfishly, no matter the consequences. It is the force that will eventually cause our downfall. It is the ID - Thanatos.
Anyone play Arc the Lad? I''ve only seen a very little bit of it, but it seems to use a fairly FFT-style combat engine, but in true 3D, with the circle cast that T H A N A T O S describes. Worth a look, anyhow.
Yes, I was going to mention Arc the Lad, but seeing as I''ve only seen commercials, I didn''t want to be wrong and referrence the wrong game. That is basically where I got the idea for circle cast (Arc the Lad commercial)

In each of us there is a force that compels us to act selfishly, no matter the consequences. It is the force that will eventually cause our downfall. It is the ID - Thanatos.
In each of us there is a force that compels us to act selfishly, no matter the consequences. It is the force that will eventually cause our downfall. It is the ID - Thanatos.

This topic is closed to new replies.

Advertisement