wait a minute...are you trying to downplay twitch while keeping close to realtime?

or are you subdivding "turns" to nearly infinity?

how do you want a fight to play out?

I'd like to add to the conversation, but I'm not sure what you're aiming at.

You and I disagree on that fundamental point. I think of an action as a series of premeditated movements, and you think of each movement as a seperate, distinct decision. Consequently, where I want a maneuver to be the actualization of an intention, you want every instant of the operation to be micromanaged by the player with the full use of his cognition.

I don't think we'll ever agree, and I doubt that I can help you refine your idea further, so I'll just wish you the best of luck and stop posting in your thread.

Best of luck.

AH. Yes. You have been much clearer minded than I have been, for discovering where each of us went apart. Thank you althesame for trying to help me go through with this. It has been interesting and helpful, to some extent, for you pointed me to limitations of the concept.

As for abstractimmersion, I was trying to downplay twitch while remaining as close to realtme as possible indeed. In fact, what most players, gamers and designers are advocating here, as a change in teh MMORPG genre, is the possibility for the fights in a MMO to happen in real-time. I was only pointing out that real-time has a limit on the net, and that twitch can only go as far as player can get it, unlike the character's progression.

I reasonned that since the MMORPG genre added RPG to its name, it meant that, to some extent, it needed at least a little bit of RP in it.

To add RolePlaying to the fights, with the current model, you give the player more options to choose from in the "press hotkey and wait" scheme of things. But I wanted the player to be able to interact a little more, and even act out, with some twitch gameplay. Only, as I pointed out a little earlier, twicth is a PLAYER limitation, not a character limitation. hence, twitch and real-time are to be ruled out. In PnP tabletop roleplaying games, the pace adjusts to the decisions and dicerolling of the players. In our case, the dicerolling can be taken out, since it can be done almost instantly with computer operation. Therefore what remains is the player's decisions, with the character's limitations.

If you can't have real-time speed on an online fighting game, you have to have non-real-time speed, ie slower speed. And since you are NOT relying on twitch gameplay, because of induced player limitations, you are then relying on tactical gameplay.

I thought that having the tactical gameplay being used with slow speed could transform the twitch gameplay into a tactical nerve war.

Using the slow-mo animation meant that the active players were involved in a tight fight were the nerves slowed things down so the player could still follow the movements of his character, without BECOMING himself the character. With the differences in stats, it also means that one character is moviong slower than the other. Since two characters with different stats could be doing the same moves, it meant that the distances crossed per frame had to be different.

One of the problems that has arisen from this thread is that if a fight involved more than two protagonists, then the scheme became impractical, because you cannot handle ALL the relative speeds AND the connexion speeds at the same time.

Still want to add something to the thread, or have you given up?

Actually a turn based combat is perfectly adaptable to multiple combatants.

Consider AD&D. A combat round (approximately 1 minute) is divided into 20 segments. Every move has an attack speed, characters with decent stats get bonuses, below par stats get penalties to when they can move. Things kick off in a sequence dependent on the equipment and the stats. Works nicely.

You can allow a player to queue commands (effectively what ICC is saying), allowing a 'kata-like' style to fighting which if you put in bonuses on certain move combinations can have advantages in effectiveness, but disadvantages in predicatability (on the part of your opponent(s)). Within the same systems, you're not obliged to queue your moves and so can have a much more 'twitchy' style of play, with moves being selected as they're about to complete. Perhaps at certain levels of skill with a weapon, a longer move queue (longer kata) is allowed? Couple this with some form of automatic parry (again based on weapon skill, perhaps relative positions of weapons/blocking items) and you're left with a system that works pretty well for melee combat of all types (including longer range weapons like polearms).

The combat system for Bloodspear is similar to this, although I also limit movement to combat commands when in melee (no free movement unless you want to break and run - instead you have 'tactical movements' like advance, back off, shield charge, circle left, circle right etc.). In playtests it works quite nicely, although there is usually a slight delay on entering combat. Non - actions against a target are still animated as attack-parry pairs, so it LOOKS like a stand-offish fight is going on. To understand this concept it helps if you consider your actual attack moves as bursts of aggression in a longer fight.

A major problem (although this applies to all MMO combat, especially melee) is collision modelling within a combat. Anything in a fight (or close by) has to be checked so that characters can't walk through each other and line / trajectory of fire intersection is checked to ensure ranged combat works correctly. This presents an issue when you consider a combat in an area with non-combatants, or on a routefinding chokepoint, such as a bridge. Doing this fast is a game of aggregation and abstraction, or just simple fudging.

One approach is to constrain collision detection to combatants only, making it possible for you to grief someone with a spear or other longer range weapon by running within your own range before attacking. This allows combats to happen anywhere without blocking movement for other players, which can be frustrating if fights happen a lot, and allows ranged attackers to shoot through non-combatants, people to run through the middle of a combat just to screw up your targetting (not desirable) and all sorts of other nastiness.

Another approach is to enforce collision detection and combat control within a radius dictated by the positions of combatants. This makes picking your way through a fight possible but also makes it possible that a combat on a chokepoint will block the movement of innocent people. It does however leave people on the 'outer edge' of the combat vulnerable to the run-up-while-not-a-combatant exploit, and can result in a lot of people involved in the combat because somebody started shooting at extreme range and expanded the combat radius to cover them.

My final considered approach (and the one I use) is to have the entire melee combat as an aggro zone with an increased area (roughly twice 1 turn's movement at combat rate) range from outermost positions, calculated per turn. Ranged combatants generate a radius of twice movement around them as normal, but the combat zone is extended (per turn) by tangents from that radius to the radius of their target if the ranged attack is 'flat trajectory' (guns, crossbows, point-shooting archery). This produces a 'cigar shaped' combat zone.

Bloodspear is geared for large group combat, and provides a number of mechanisms to encourage it (as well as make it technically feasible). Group attacks are one such mechanism - a ranged attacker can elect to target a whole melee group (aggregate target) for an easier to hit roll, or a melee attacker can elect to go for any hostile in range. The precise target for ranged attack is selected at random though, and could be on either side. Melee combatants can also gain bonuses by defending their linemen (friendlies to either side) and through use of the 'formation' skills to keep defensible positions in pike and bill blocks, as well as executing group special moves (under the direction of a group leader) like the famous testudo. Ranged group attacks are not possible if the attacking ranged group is in melee (part of the aggregate target).

I have to admit that I had never imagined the "many moves combined in one single line" issue as katas. It is a good idea, and a nice vision. But it leaves me a bad taste in my mouth for some reason. Let me explain.

A kata is something you learn, as a basis for your evolving skills in combat. It is a series of moves you repeat, endlessly, to learn the perfection of movement, until you have completely mastered the series and the movements, and can train on something more difficult. Meanwhile, you can also hone your combat skills, but nowadays, and I am quite sure there were also kata specialists in ancient times, kata specialists are NOT the same as combat specialists. They specialize in movement perfection instead of combat efficiency. It doesn't mean they won't be efficient in combat, only that someone less interested in the perfect movement and a little more in the "movement that will hurt a lot more" will have an edge over them, no matter how much they have learned their katas.

Which means that either the katas are to be considered newbies weapons, which ruins the whole scheme, making it more difficult to touch a newbie if the first move connects than to touch an old timer in free combat mode. Or that the katas are a form of macro, a line of commands which, since they have been repeated endlessly, are achieved faster and with unblockable damage if the first move connects, but leaves the fighter open to counter-attack if opponent evades early in the combo, like in SoulCalibur.

But on the other hand, a combat doesn't always appear as a duel. Sometimes it is a group fight. If a kata is almost impractcal in a duel, since someone clearminded enough to forget to use katas against you can dodge your first attack and transform you in sashimi before you end your last move, it becomes a mincing machine when you're pitted against two ennemies, because one occupies you while the other one creeps behind you and launches his deadliest kata, which you cannot defend against. And this kind of behavior can be called "exploiting a feature", since there is nothing you can do against that.

So maybe katas are out of question because of unblockable series of damage. But once again, how can you use this system to withstand a grouped attack against you? I guess you can only defend a grouped attack as long as you can move faster than two ennemies at once, meaning that effectively, someone with speed 20 can make his way through three persons with speed 6 with difficulty, but can also be killed by the same grouip if it acts with purpose and surrounds him completely, putting strategy at the core of fighting gameplay.

As for katas, again, maybe the concept of combo, which would be short katas of two to four movements, can be used with benefits. And then, katas can be used with deadly purpose by high-speed combattants against low-speed newbies for griefing without much thinking of it.

And about the "tick", yes, that was what I had in thought. It works quite well with tabletop PnP RPG called Feng-Shui (which you should check if you haven't already) and I thought it could be adapted with benefits to the Online RolePlaying issues. Thanks for clarifying.

have you considered using a simulation model?

you have 3 well defined hurdles.

1)the issue of judging depth at close ranges in first person mode.

2)the issue of lag ruining the liquidity of your simulation.

3)the issue of having a complex combat resolution system but being able to manage it with an intuitive interface.

whether you use a class based system or a skills based system a players personal skills can be optimized by their class or skills choice.


a horrible twitcher could still make a great healer or a mage that doesn't use aim and cast type spells. other skills and classes would minimize twitch necessity while some classes would be primarily twitch.


stats would effect outcomes but the "to hit" mechanism would be handled by the physics system.


twitch skills could be further augmented based on level or skill, here's an example;



A physics based archery system would be a bow projector and an arrow projectile. the physics simulation would propel the arrow and your collision detection would determine whether the archer scores a hit. to properly hit a moving target you're going to have to lead it. at higher levels of the archery class or archery skill you can speed up the projectile, or you can paint a "lead reticule" to help in the target acquisition.


my point being you can still have stats be king if that is your wish, you simply manipulate the simulation based on stats, level, or ability.

In many single player FPS games the player has a weapon that has some variation in its targeting (ie it doesn't nessesarily hit where the reticle is). Some of these weapons accuracy varies with use (in tribes vaegance the chain gun has worse accuracy if you use it constantly due to over heating).

Wat if you had the level of skill of the character reflect the opposite of this. As the stat level of skill increases the aiming becomes better. That way twitch gamers will also need to upgrade their skills to get the advanatage, and it still gives the twitch gamer control over the aiming (so both skills are needed but it does not totaly rely on them).

With melee this could be reflected in the amount of damage done, or even the accuracy of the swings (it might go high or low, the swing might be slow or be delayed a bit). The ways that you can have skill reflected in the physics of such a game is quite extensive.

You could even have sub skills that reflect different physics aspects that can be effected, increasing the number of skills that can be selected to either compliment or cover for the lack of skill in the way the player plays and their skill levels.

I basically agree with above poster, I think, or at least my ideas are similar. If you want to maximize stats and minimize twitch, a system that is more forgiving for higher level players would be the way to go. This is somewhat counter-intuitive to the norm, based on player skill, but I guess I'm talking about real time.

I see 2 major flaws in your position:

1: at some point, if you wish to not divide "ticks" to infinity (here there be granularity problems) , you need to have a character COMMIT to an action.

2: You don't seem to understand, unless there's some magic formula I'm missing, that slo mo combat, even in an instanced bubble, would still be at the mercy of whatever lag the collective system deals out. You just have more time to make a mistake. Now, I'm not saying that slo-mo can't be done well (I am skeptical that "bubbling", which I've thought of before, can be done well), but that it isn't going to solve any problems of twitch, skill, or lag. Instead, it will move these problems to a finer scale, which reductively would be WoW with bullet time.

Edit: OK, my personal solution, though it be nothing like what you probably had in mind:

Keep the stategic element of combat and high levels with higher ability to hit "sweet spots" or parry based on skill. (talking about swords for simplicity)

Say a waist-level slash is activated by the player when the opponenet has just missed a sweeping slash to the legs. The high skill allows a certain margin of error (+/- ten centimeters) from the normal slash line. If the opponent's neck is within the margin, the player cuts it--chance modified by player's perception, opponent's agility, etc.

Or say a certain attack has a chance of intercepting and blocking an opponent's strike at certain coordinates when the player's blade is at certain positions, during the attack. (Way of the Intercepting Fist, anyone?)

There could even be aggressive/defensive "sliders" calculated into these equasions. Playing a higher level foe? Allow the dice roll to intercept more attacks than follow through.

If we're talking about RPG's, and if you want stats to be supreme over player skill, you have to come up with a way to let the dice do it.

I personnaly don't like the idea of having dice intevening in calculation, when it comes to simulating a combat, since it all depends on stats, skills, and ultimately, timing.

The stats are static. There is no random component there, therefore, no dice to roll.

The skills are, if not completely static, at least stable during a single fight. once again, no need to roll a dice.

The timing. Ah, well, timing is a matter of twitch. And I was precisely hoping to remove twitch from fighting with slowing the motion, so as to replace twitch with strategy. But once again, be it twitch timing, or strategic timing, I can't see a reason to use a roll.

But maybe the dice roll is only there to emulate randomness in a perfectly predictable world...

Why should a fight be unpredictable? Mostly, a fight is merely taking into account a multiplicity of variables, and using them to your advantage, instead to let your opponent use them. Position relative to that of the sun, angle of lighting, general movement space, relative heights, relative speeds and strengths, there is almost nothing there that is subject to randomness here. SO why make it unpredictable?

The only reason I can find to use randomness is if the simulation isn't pushed far enough to allow for complete predictability. You never see randomness in chess games. And I strongly believe that, at the ultimate levels of fighting skills, a fight is very much like a game of chess. You never leave an open space for your opponent that you don't want to be there, and you never make a move that you do not want. No unpredictability. Except what comes from your opponent.

So maybe, since the unpredictability comes from the lack of experience, you could slowly remove the randomness as you level, or gain experience in something. Which could give an edge to the newbie over the experienced fighter, because although an experienced fighter would certainly move his blade along THAT line, a newbie maybe would WANT to move it along that line, but wouldn't be able to master it enough to make it perfect, and then would bypass the master's block, through sheer luck? Of course, luck doesn't win every time, but it would be sweet to see a newbie slash an old timer just through luck, even once...

I have also devised a way for the stats to reflect the level of tiredness, through, well, fractions. They would be multiplied by (actual Stamina/Overall Stamina), which means that you would get worse at fighting as the fight goes longer. The simulation DOES work. But I'll grant you it demands a little experience, as a player, to get used to it and take full advantage of the system. And my friends have helped me see that there were many different strategies available, through that system, just because of the different stats builds you can achieve, and the different weapons you can use, which is precisely what I wanted to achieve. The simulation is still a little long to accomplish through the paper prototype, but it remains fun. I hope I can make a computerized one.