Continuity in TB Combat - bad idea? (+associated musings)

In most squad/party turn-based combat, position is handled with a grid (squares/hexes...), and actions take a specific amount of time (e.g. snap shot = 17 time units). This can be annoyingly restrictive in some circumstances - e.g. where you'd like to take a shot from between two grid squares, or would like to use 30 time units to make the best shot possible, but are only able to take a snapshot for 17.

A fairly obvious "solution" would be to allow units to be positioned continuously anywhere (walls permitting), and for units to be able to put any amount of time into a shot. [perhaps some games do this already - I haven't played many in recent years]
On the face of it, this would seem to me to be an advantage. The player would be more able to focus on the tactics inherent to the situation, rather than on gaming the arbitrary cut-offs in the system.

Is this a bad idea? If so, why?

Problems I see in some contexts are:
(1) For many actions it doesn't make much sense to have variable amounts of time/accuracy/speed. E.g. equipping a weapon or similar. Since the player might need to be sure of having say 4 units left for this purpose, he'd be forced to be annoyingly precise with movement/shooting times etc. There's always the option to let the player reserve enough units for some action (as in Xcom), but that's going to be cumbersome for more than one reserved action.

(2) You'd still probably need to impose some minimum time cost for attacks. If the player will often be making calculations based on that minimum cost, he'll again need to be undesirably finicky with movement calculations....

(3) Sliding scales for shot time / movement allow the player to be much more careful/precise in his actions. This could slow things down considerably for some players, since there's no longer a few neat options, but rather a continuum of possibilities.

(4) The interface might get in the way of things. Many non-decisions would become decisions - not all of them interesting. [although some daft problems might evaporate with the removal of 15-time-units-and-can't-pull-the-trigger or can't-take-cover-in-the-safe-area-overlapping-two-squares type situations]


I'm not too bothered about (1) and (2) in the situation I'm currently considering, since I'm pretty sure I can reasonably eliminate fixed costs entirely (if it helps). Mainly, I'm concerned with the impact on the player decision-making process. I think that sliding scales for actions, and continuous positioning would allow the player to focus more on the tactics of the situation, than on the game mechanics. I'm just not certain that'll necessarily turn out to be more interesting.

Thoughts?

still leaves a purely mathematical solution to every encounter as opposed to a tactical one. actually increases the importance of the math over everything else.

Silent Storm allows you to allocate APs towards more accuracy and even allows you to carry them over to the next turn--though this only applies to sniper shots with a scoped rifle.

mister lamat said:

still leaves a purely mathematical solution to every encounter as opposed to a tactical one.

Click to expand...

It's both - with or without continuity.

actually increases the importance of the math over everything else.

Click to expand...

Why?
If things are well balanced, the most common-sense approach to a given solution will often be best. With more arbitrary time cut-offs, the solution will more often be tailored to arbitrary game mechanics, rather than the situation alone.
Things only become mathematical when doing explicit calculations grants a significant advantage over intuitive qualitative tactical reasoning. Ideally it shouldn't often - once the player knows the ropes. [also there's no reason not to do most of the calculation automatically - e.g. if the player is moving a unit, it's quite simple to show him information on the new position as he hovers (most accurate shot he can make, cover afforded, visibility...)]

Silent Storm allows you to allocate APs towards more accuracy and even allows you to carry them over to the next turn--though this only applies to sniper shots with a scoped rifle.

Click to expand...

That seems sensible. But movement's still based on a grid right?

You could liberate the rule that everyone has to chance to act in every turn and give npcs the chance to act whenever they have action points. This would eliminate turns in the classic way and would give much more use for sequencing. Here's an example:

Heres a possible and very simplified scenario:

• 
  * Player, is using magnun pistol that takes 3 ap to fire and 4 ap for a localized shot.
  * Raider1, is using a shotgun that takes 4 ap to fire and 5 ap for a localized shot. He is 10 meters away and has a 50% chance to miss at this distance.
  * Raider2, is using a combat knife that takes 2 ap to hit and 3 ap for a localized hit. He his also 10 meter away, and can only attack at 1 meter distance.
  * Anyone needs 1 ap to run 1 meters.
  * Player has the best sequence, followed by raider2 then raider1.
  * All have 50 health.
  * Magnum takes 20hp per hit and is accurate at 10 meters or less for the player.
  * Shotgun takes 40hp per hit if less than 5 meters and 20 hp if at more than 5 meters.
  * Knife takes 10 hp per hit.

And this is how it would work. Time is measured in APs. So if an AP is 1/10 of a second then 10APs is one second. This is just a possible way to do it.

-------------------- :0AP time: --------------------

****Game is paused for the player to decide what to do.***

Everyone can act, since nobody is executing a previous action (decision time), so the best sequence determines who acts first.

Player has the best sequence so he decides to shoot raider1, which takes 3 APs.

Raider2 decides to run towards the player, and is at 9 meters.

Raider1 decides to shoot the player with his shotgun which takes 4AP.

-------------------- :1AP time: --------------------

Raider2 moves closer to the player, at 8 meters.

-------------------- :2AP time: --------------------

Raider2 moves closer to the player, at7 meters.

-------------------- :3AP time: --------------------

Player fires his magnun and hits Raider1 for 20 hp. He has now 30hp.

Raider2 moves closer to the player, at 6 meters.

-------------------- :4AP time: --------------------

****Game is paused for the player to decide what to do.***

Player repeats his previous action and targets for localized damaged in the head. He needs 4AP to shoot again.

Raider1 (30HP) shoots at the player and hits. At that distance the max damage from the shotgun is 20. The player armor deflects 10hp so the damage is only 10hp.

Raider2 moves closer to the player, at 5 meters.

-------------------- :5AP time: --------------------

Raider2 moves closer to the player, at 4 meters.

Raider1 (30HP) decides to shoot again.

-------------------- :6AP time: --------------------

Player1 (40HP) fires his magnun and blows raider1 head off.

Raider2 moves closer to the player, at 3 meters.

-------------------- :7AP time: --------------------

****Game is paused for the player to decide what to do.***

Player decides to do a localized shot in the legs of raider2.

Raider2 moves closer to the player, at 2 meters.

-------------------- :8AP time: --------------------

Raider2 moves closer to the player, at 1 meters.

-------------------- :9AP time: --------------------

Raider2 prepares an attack with his knife that requires 2AP.

-------------------- :10AP time: --------------------

Raider2 moves closer to the player, at 2 meters.

-------------------- :11AP time: --------------------

Player (40HP) criticaly hits raider2 in the leg. Raider2 can only walk slowly and needs 2AP to walk a meter.

Raider2 looses his chance to act because he is being shot.

-------------------- :12AP time: --------------------

****Game is paused for the player to decide what to do.***

The player decides to step away from raider1 and shoot him from a safe distance.

The victory of the player is assured.


This is continuous but you can loose track of who is supposed to act and when. With TB you have a queue of people (ToEE has this) and you know exactly who is going to act and in what order.

That seems sensible. But movement's still based on a grid right?

Click to expand...

Yeah, but in Silent Storm you don't seem to have any problem with wanting to shoot between two grids. Or at least I can't think of any. Your characters are smart enough to shoot around the wall if it's partially blocking your view. Though a poor shot can send a bullet into the wall right in front of you. The (nearly) fully destructible terrain also helps alleviate this.

If you can't tell already, I'm a bit of a Silent Storm fanboy. I love the engine and wish more games would take advantage of it.

OldSkoolKamikaze said:

Yeah, but in Silent Storm you don't seem to have any problem with wanting to shoot between two grids. Or at least I can't think of any...

Click to expand...

Sure - perhaps there's no great disadvantage if it's done well. I'm just not sure I see much advantage either. Presuming that precise calculations of action points aren't an issue, I'm not sure what a grid buys you. [of course that's a large presumption - in most such games you frequently do need to know that you'll have at least X amount of points once you're at Y]

If you can't tell already, I'm a bit of a Silent Storm fanboy. I love the engine and wish more games would take advantage of it.

Click to expand...

I should probably play it - and Hammer & Sickle for that matter.


@elander_
That's not really what I meant with "continuity", but since you bring it up...
I'm not sure I like that approach. It's probably more realistic than a standard TB system in terms of time (since TB sequences events that ought to be happening simultaneously). That's not necessarily a good thing though. Make the turns shorter, and you end up with real time, only without the twitch skills, and with time to think. That's better than real time for tactics, but I don't know that it's an improvement over TB.

Perhaps it's not sensible to say that one system is "better" here - it's just very different, and not necessarily an improvement. I guess that the system you describe does go along with my stated aim of making the tactics about the game world situation, rather than arbitrary mechanics. However, I wonder whether that's necessarily an asset. Mechanically it's RTwP - only you're told where to pause.
You'd get interrupted by enemy action a whole lot more (surely realistic, but desirable?), and you'd be unaware of the result of one unit's actions while ordering the next (again, realistic, but not necessarily desirable). Things would have a more natural flow, but you'd be making decisions on the basis of more limited information - in parallel, rather than sequence.

I don't recall playing with such a system before. Does anyone have thoughts on how they compare in practice?

On a related note, what are thoughts on interrupts / opportunity fire etc? Is it preferable to get a full interrupt on the enemy's turn, to take any action desired - or should AI controlled opportunity fire be the only out-of-turn action? Should such fire/action be targetable anywhere, or only at the currently acting enemy?

E.g. in Xcom, opportunity fire only targets the current soldier - which usually makes sense. However, this leads to tactics like shooting a hole in a wall from distance / at an angle with one soldier, then getting others to look through the gap and observe/shoot enemies. It also only targets soldiers which are in sight range of the enemy being attacked - meaning that it's often preferable to see the enemy with one guy, then shoot with another from distance (perhaps after knocking down a wall to clear the sight lines with a third).

This is all pretty daft/nonsensical, of course - and also only possible with a fairly traditional TB approach. But is that daftness a problem to be solved, or part of the charm of the system?

I've written something remotely similar here (skip to paragraphs 10 & 11, or 2 and 3 from the bottom of the OP ).

I'd like more interrupts as well, but there should be something that whether an interrupt becomes an AI controlled reaction or the control is given to player should depend on, like stats governing stress/panic levels. Character turns around the wall and is faced with an opponent. Who will react first? The one that reacts first; will (s)he be able to retain calmness and react in a meaningful and also an efficient way? Does (s)he flinch and shoot elsewhere, or the target right between the eyes? Will his/her reaction modifiers affect/limit the options available? I turn around the wall, see an alien, I get scared but I'm reflexive, so whatever I'm about to do, I'll do something first. "Oh wait, I'm scared so I can't think of punching it in the face and jump out of the window" or, "oh wait I can't think right, maybe it really isn't a good idea to hit it in the face and I should just take off, but alas, I hit the goddamn mofo".

Do you think that kind of reaction determination is too much? I'm aiming for that kind of thing, adding misjudgements into the mix.

As for grid, I like grids because they make the presentation and the abstraction easier. Easier for designer and easier for gamer, but they don't always have to be fixed. I think more of relative grids. Relative to the object and the subject. Relative to the character selected that's about to make an action and the aim / target of action, eg. an enemy to attack, a distance to walk/run etc., something to pick. Size or length of grids may vary by very little amounts in accordance with the actual in-game distance to the subject. Everything else going on on the screen other then the object and the subject could be calculated behind the scenes with precise complex values if necessary, but when a direct encounter comes up, values are rounded or simplified to make it easier for the player. So if I'm to walk up to this bandit and hit him, the grids are adjusted accordingly between me and him. This way, player-controlled alternative movement options (if available) may become more precise and meaningful, including changing movement direction. For instance, you can more precisely and reasonably penalize action points when making a steep angle to change direction. Just a simple example off the top of my head, of course.

And for actions, I'm leaning on to seperation of action and movement (where there are more player-controlled options than just walking to a point), on condition that they affect each other. If a character is trying to attack something or someone, or take out something from his bag while running, either his AP or all non-movement actions are ought to be negatively modified. If shooting while stationary (taking cover or not) takes 5 AP and you have 12 AP, it may either take 5+X AP or you have X less APs overall when not stationary.

ToEE used radius movement instead of grid and JA2 let you add 1 AP at a time to shots.

Curses, it's been 8 months since I finally bought ToEE and I still haven't played it yet.

denizsi said:

...Do you think that kind of reaction determination is too much? I'm aiming for that kind of thing, adding misjudgements into the mix.

Click to expand...

I think it's reasonable - but preferably in combination with some player command to be ready for [insert circumstance here]. (e.g. as Section8 describes here)
That way you might instruct your guy to go through the door and be ready to open fire to his right. If he sees someone where he expects, he'd get a high (or automatic?) chance for a fully controlled interrupt. If someone comes at him from his left, he'd stand a much lower than usual chance.

You could still allow some chance to have a bad reaction even when prepared (e.g. perhaps if three guys are to the right at point-blank range), but in general I think the player should have the chance to predict+prepare, and get reasonable results where he's right. [since in this case the character isn't really surprised at all, and that's what you're modelling - response when taken by surprise]

I think more of relative grids. Relative to the object and the subject.

Click to expand...

An interesting thought. I hadn't considered that. I guess it makes good sense for clarity if nothing else - you can see precisely where your guy will end up after a move.

And for actions, I'm leaning on to seperation of action and movement...on condition that they affect each other.

Click to expand...

I agree with that - if you're going to carefully balance the relationship between the two (or more ).

Human Shield said:

ToEE used radius movement instead of grid and JA2 let you add 1 AP at a time to shots.

Click to expand...

I see my lack of culture is showing.
But how well did these features work? I know both JA2 and ToEE are well-liked for their combat systems, but did you find these particular features assets? Was there any particular downside?


On a more general note:
Are there thoughts on the importance of tense exploration vs tactical firefights?
How big can a squad/party get before the tactical side suffers? (I'd be inclined to say that more than 10 is probably too much)
Is it more satisfying/interesting to dispatch 20 light, fairly dumb opponents, or 7 armoured clever opponents who'll get injured, make tactical withdrawals, counter-attacks etc. etc. (maybe I'd prefer the latter, but then UFO:EU was the former - and a lot of fun)

Anything else that springs to mind as a TB/PB combat must-have?

Are there thoughts on the importance of tense exploration vs tactical firefights?

Click to expand...

I'm more partial to tactical firefights, but tense exploration helps change things up.

How big can a squad/party get before the tactical side suffers? (I'd be inclined to say that more than 10 is probably too much)

Click to expand...

I'd say even 10 is pushing it, but it depends on the size of the map. Also, I'd be a liar if I didn't say that I enjoy the occasional large battle where I can use more units than usual. Some of my favorite battles in the Final Fantasy Tactics and Tactics Ogre games were ones where you were allowed a large team to storm a castle or such. Though I think something like that should be somewhat rare.

Is it more satisfying/interesting to dispatch 20 light, fairly dumb opponents, or 7 armoured clever opponents who'll get injured, make tactical withdrawals, counter-attacks etc. etc.

Click to expand...

I really want to say the latter, but a part of me deep down loves dispatching simple cannon fodder. I'd lean more towards a large number of light, dumb opponents with intelligent, clever opponents more rare and in places where it'd make sense.

Here I am, and here is my compulsory Brigade E5 reference!
Well, I could stop here, but I must add that you nearly 'word-by-word' repeated thoughts of that game's designer, so congrats .
If you didn't try it, you certainly must. Too bad the graphics and most of the quests kinda suck (not stellar in Russian, completely borked with translation last I heard), but that's perhaps the best implementation of RTwP ever.

Can still be rather fiddly when you have a lot of mercs fighting a lot of enemies (especially if you are oblivious to alt-click and Z key), but otherwise simulator-level of realism. It has everything you ask... and btw, the AI part does NOT suck. Don't get fooled by early bandits - the are supposed to be dumb.

Since someone already mentioned Silent Storm, I'll throw in an example from the afterfall (a game that meeds some pimping anyway) combat system, which adresses some of the things you mentioned (http://www.afterfall.pl/index/en/44)

Action Points and Movement Points
Two variables govern the number and quality of actions the character can perform:
Action Points used to perform manual actions, attacks using hands and handheld weapons, as well as other activities requiring arms or making them temporarilty busy.
Movement Points used primarily during movement. Each step made by the character decreases their availible level, the same applies to activities involving legs (e.g. a kick) or requiring the character to stand still (e.g. lockpicking).
Thanks to such a solution we will avoid problematic situations where the Gamer had to choose – either to inflict maximum damage or to move to a strategically better position. In Afterfall, these two options are smoothly combined.

The MINIMUM / OPTIMUM Rule
Most of the actions that can be performed in Afterfall during combat does not have a predefined constant cost in Action / Movement Points. The cost of an activity with an efficiency lesser than 100% is calculated using the MINIMUM / OPTIMUM rule. Generally, an activity has got a defined range of costs. The player, on his/her own decides how many points to spend – this has is a decisive influence on action resolution mechanics (the lower the chosen cost, the lower the effectiveness).

Stretching Actions
Sometimes during combat the player wants the character to perform an action, but there are not enough availible Action/Movement Points. This is not a problem in Afterfall, the system automatically stretches the action’s performance on two or more turns, until the appropriate amount of Action/Movement Points will be spent.

Click to expand...

galsiah said:

(3) Sliding scales for shot time / movement allow the player to be much more careful/precise in his actions. This could slow things down considerably for some players, since there's no longer a few neat options, but rather a continuum of possibilities.

Click to expand...

Sliding scales would be all over my dream turn based game of the future. I don't really see them as a huge slow down, there could always be defaults. But sometimes (I found this a lot in silent storm) that broken up options just aren't enough. Especially for things like the length of a burst, quick burst and full auto are uaually on the two extreme ends, and don't cut it for every situation.

The other option that would solve a lot of those things would be to give more full control of a charater for the players turn (at least in some aspects), realtime style. So, you want to fire longer? You hold down the mouse longer. You want to run while firing? Do it as if you are playing a realtime game, instead of worrying about switching all the options for a given situation. Of course, this sort of thing could cause plenty of problems on its own, but it's an interesting concept.

Anyway, I want to get back to this and read/reply to the whole thread, but my brain is completely fried now, so I come back later.

The player would be more able to focus on the tactics inherent to the situation, rather than on gaming the arbitrary cut-offs in the system.

Click to expand...

This is a suggestion to the problem I *think* you are describing: "Continuity in turn base" - continuity from turn to turn in a classical sequential TB.

lets say for example that one of the player controlled combatant in a certain situation have only 5 AP left, however the player wishes for that combatant to make an highly aim shot (9 AP). Now the player only has one option - to find a way to use the remaining AP in a productive way whether it is to fire a less aimed shot i.e. squeezing out of the sequence as much as possible OR to better situate the player controlled combatant for the next sequence i.e. movement / prone change (the other option is to discard the remaining AP and wait for the next sequence which is of course not a desired situation).

The only solution that I can see to this problem is to offer a mechanism so the player can (in the example I described above) move the remaining 5 AP to the subsequent sequence, so that in the following sequence the player controlled combatant now have an addition of 5 AP for the originally intended aim shot that the player wanted to perform, now the aim shot will first consume the 5 AP from the previous sequence and 4 AP from the current sequence.

This should of course come with the following restriction – if the player changes the action that he originally intended to some kind of other action, then he automatically forfeit any AP from the previous sequence(!), or in other word, the player can move AP from sequence to sequence only if it is tied to a specific action!

I also want to add that in general imho there is very little you can do/change in classical sequential TB (X-COM/JA) without changing it to some kind of phase based / CTB (aka FT – AP regeneration)/ WeGo / other. To rephrase it, if you are planning to offer in a game some kind of advanced TB, I would advise you to consider from the outset one of the non classical sequential TB.

On a personal note – I am very interested in the "WeGo" mechanics; I personally believe it can be a very good fit to squad tactical games.

galsiah said:

but in general I think the player should have the chance to predict+prepare, and get reasonable results where he's right. [since in this case the character isn't really surprised at all, and that's what you're modelling - response when taken by surprise]

Click to expand...

This reminded me of a medieval martial arts system I've been designing that takes prediction and preparation into account. To give some basics without going into detail, there two primary elements: stance and action. These two are interchangable: actions you can perform while positioned in a certain stance, stances you can enter after performing actions and actions following actions are limited / predetermined.

Let's say these are the stances and actions in the game: S1,S2,S3; A1,A2,..A9; and the table below shows the interchangable nature of stances and actions:

Current Stance : Performable action
S1: A1, A2, A3, A4
S2: A3, A5, A7, A8
S3: A4, A6, A7, A9

Last action : action available following last action
A1: A3, A8, A9
A2: A6, A7, A3
etc.

Last action : stance available from last action
A4, A5, A9 : S1
A2, A3, A8 : S2
etc.

If character's level of knowledge and experience allows, you can see the actions the opponent may perform with probabilities based on stats again. Going a step further, characters may attempt to dominate and control the flow of the encounter by prediction.

Say, you predict that, based on the probabilities, your opponent who's currently in S1 , will most likely perform A2. You tag A2 as prediction and if your prediction is met, you gain certain advantages, eg. efficiency, moral, a chance to demoralize the opponent. Based on your stats, you can make further levels / depths of prediction, eg. you predict that he will perform A2, you look at the options available to the opponent actions after A2 and predict that, as well based on your own action. Better (not just "more" - it takes taking risks, the prediction here being an example, to try new or different things in combat to get "better") experienced you are, further you can predict, and even make branching predictions. Each successfull prediction has a chance of gradually demoralizing the opponent and allowing you to control the flow of the combat, which is what I call dominating (he will do A2, I will response with A3 and force him to do an A9, at which point I'll do A7 and enter S3, etc. etc.)

Branching prediction I mentioned above amounts to being prepared for more than one actions. He may perform A2 but he may also perform A3. A less experienced character may focus on one choice, but a better experienced character can focus on both actions. Of course, should your predictions fail, you give certain opportunities to the opponent, receiving some penalties yourself.

There are also other aspects like footwork, faking actions, stalling, opportunity attacks etc. that adds a lot, but not necessary for the general explanation here. Interestingly, I really haven't thought of making even a simplified prediction pattern for systems not based on medieval martial arts. Perhaps because I don't know much about them in reality, as opposed to all the reading I do on medieval martial arts.

On a more general note:
Are there thoughts on the importance of tense exploration vs tactical firefights?
How big can a squad/party get before the tactical side suffers? (I'd be inclined to say that more than 10 is probably too much)
Is it more satisfying/interesting to dispatch 20 light, fairly dumb opponents, or 7 armoured clever opponents who'll get injured, make tactical withdrawals, counter-attacks etc. etc. (maybe I'd prefer the latter, but then UFO:EU was the former - and a lot of fun)

Click to expand...

All I can think of atm is that vertical spaces (floors) should be used a lot more when reasonable. They help a lot by limiting the perimeter of the action while expanding the exploration.

Anything else that springs to mind as a TB/PB combat must-have?

Click to expand...

Initiating and ending movement options manually. Starting to run, continuing to run, changing direction while running and stopping running for instance, should cost different MPs, with appropriate modifiers for APs (it's gotta be hard to perform any action while changing direction when running, trying to preserve your balance).

Wouldn't a variation of the ToEE engine be the answer?

It switches to a timed combat mode when necessary and uses a 'time bar' to judge movement and actions.

Might be worth a look if you haven't seen it already Galsiah. If you can come up with an engine which removes the DnD ruleset and improves on what is present in ToEE, from a TB combat perspective, you'd have a winner.

Assuming you're going with a "My turn, Your turn" style (I personally prefer simultaneity):

I like the idea of low base costs for actions, which have a correspondingly lower chance of success. Then you add points which bring the success rate to 'average' levels.

Eg
Shooting a rifle

Normally has 70% accuracy and takes 7 AP (= 3.5 seconds)

In this new version you can shoot a rifle with 1 AP but it has a very low (16%) accuracy. Every AP added up to fourth adds 10%, next three add 7% each, next two adds 5%, next adds 3%.

So you have:

1AP - 16% (0.5 seconds, surprised reaction fire)
2AP - 26%
3AP - 36%
4AP - 46% (2.0 seconds, rushed aimed shot)
5AP - 56%
6AP - 63%
7AP - 70% (3.5 seconds, standard aimed shot)
8AP - 77%
9AP - 82%
10AP - 87%
11 AP - 90% (5.5 seconds, carefully aimed shot)
12+ - no improvement

Or something like that.

Same thing for other actions like reloading...

Base cost of loading one bullet into a revolver = 1 AP, base chance of fumble (drop bullet) at 20%
+1 AP reduces fumble by 12%, +2 AP reduces by +17%, +3 AP by etc

So quickly loading a revolver (during a fight) gives you a pretty high chance of fumbling and wasting time and spilling your bullets all over the ground. If you manage to find temporary cover and can spend one second for each bullet, you won't be able to shoot for longer but it will be more efficient. The best would be to have a spare and have your revolvers loaded before fighting (as in real life)... or get a clip-loading pistol

Some people say its too much math, but I don't think so. You wouldn't necessarily have to calculate anything, if designed properly you'd just make decisions as you would in real-life, and the mechanics adapt which is really what RPGs should be about. Better of course to have simultaneous turns.

As for 'fun', I personally think realism is fun, and the idea of having to think about things like how fast you can reload your weapon is strategic and does adds to the experience.

Re: Continuity in TB Combat - bad idea? (+associated musings

galsiah said:

In most squad/party turn-based combat, position is handled with a grid (squares/hexes...), and actions take a specific amount of time (e.g. snap shot = 17 time units). This can be annoyingly restrictive in some circumstances - e.g. where you'd like to take a shot from between two grid squares, or would like to use 30 time units to make the best shot possible, but are only able to take a snapshot for 17.

A fairly obvious "solution" would be to allow units to be positioned continuously anywhere (walls permitting), and for units to be able to put any amount of time into a shot. [perhaps some games do this already - I haven't played many in recent years]
On the face of it, this would seem to me to be an advantage. The player would be more able to focus on the tactics inherent to the situation, rather than on gaming the arbitrary cut-offs in the system.

Is this a bad idea? If so, why?

Click to expand...

Well, I don't really see it as inherently good or bad, just different. Because gaming the arbitrary cut-offs in the system can be enjoyable in itself. In the end, there's not necessarily any gain from adding more simulationist elements. It just makes the game more of a simulation.

In some ways it can be of benefit to have simple mathematical costs for actions to better inform decision-making, and as such, the line between success and failure can become thinner. The less informed a player is, the more guessing is required, and you have to either be forgiving of that, or anticipate a large degree of trial and error. Are either desirable? Probably not.

One thing I like about chess is the dynamic of "threatened" squares. For any given move there are way too many choices to be properly considered and compared. Once you take "suicidal" moves (almost) out of the equation, you are left with a much more finite set of possibilities, which aids both immediate decision making, and forethought/anticipation.

However in the specific case of AP cost for shooting, I think that removing or reworking one of the more irritating artifacts of TB abstraction is probably for the better. A fairly simple alternative for me would be to allow the player to have their firearm brought to bear as they walk. They move slower, but if they have to react, they can do so at a lessened cost.

And as a quick aside, one of the true beauties of turn-based is that it's fairly trivial to replace individual systems with more elaborate ones without interface concerns.

(3) Sliding scales for shot time / movement allow the player to be much more careful/precise in his actions. This could slow things down considerably for some players, since there's no longer a few neat options, but rather a continuum of possibilities.

Click to expand...

You can still achieve similar results without a sliding scale. Silent Storm's solution of having snap shots, aimed shots, and then a careful shot that would utilise your entire remaining AP pool (plus similar mechanics for sustained bursts) was elegant enough, and I can't remember ever feeling as though I was wasting APs by idling.

(4) The interface might get in the way of things. Many non-decisions would become decisions - not all of them interesting. [although some daft problems might evaporate with the removal of 15-time-units-and-can't-pull-the-trigger or can't-take-cover-in-the-safe-area-overlapping-two-squares type situations]

Click to expand...

Would this be overcome by a standard interface and the option to finetune actions when required? Assuming that players infrequently want to take cover on a square overlap, then you can present the standard interface unless the player requests otherwise. Likewise, a "desperation shot" option that burns all remaining APs at a fairly steep accuracy falloff would allow some forgiveness for anyone who either accidentally over-runs their reserve mark, or needs just half a step to get a line of sight.

There are countless solutions, really, and I think the best ones revolve around giving the player a very limited set of additional actions in case of rather than far more complex actions and sliding scales being present all the time and hindering "everyday" play.

I'll dig into the continued discussion sometime soonish.

I increasingly think I'm a simulationist. I've found my main problem with traditional turn-based is initiative stuff (long turns + you move/I move). Like the example Galsiah gave, where you shoot a hole in wall with the last APs of one guy and use the first APs of the second to shoot the enemy revealed by the hole.

Are there any CRPG combat systems with simultaneity, which are not real-time?

I think that kind of approach is the solution, it just needs to be designed and balanced right.

I do not think it's important for players to fully understand the mechanics or to be able to calculate exact odds. A good system should let the player to make decisions based on feel, and the system should match realistic expectations. And with CRPGs complexity of mechanics simply isn't an issue anyway... I think many designers have gone with the approach that mechanics can be understood and calculated, and detail has been sacrificed for that reason.

An example of simultaneous turn-based: Each 'round' is 0.5 seconds or something of that order. Actions which take longer than 0.5 seconds are spread out over several rounds... if you're interrupted you may lose your action or have a lower chance of success. The more time you dedicate the better the action quality (the bonus per added round and caps depending on the action).

[In general I'm leaning towards simultaneity - though I'm a long way from anything concrete]

denizsi said:

All I can think of atm is that vertical spaces (floors) should be used a lot more when reasonable. They help a lot by limiting the perimeter of the action while expanding the exploration.

Click to expand...

I agree with that. It's a little more fiddly to get everything working smoothly with a third dimension in the mix, but probably worth it.

Section8 said:

Well, I don't really see it as inherently good or bad, just different. Because gaming the arbitrary cut-offs in the system can be enjoyable in itself. In the end, there's not necessarily any gain from adding more simulationist elements. It just makes the game more of a simulation.

Click to expand...

Sure - but I'd prefer to have the player game the (perhaps unrealistic) game world, rather than game the numbers (where possible).
I'd rather emphasize reasoning between two very different tactical courses (whether or not realistic), than between percentage assessments of auto-shots and snapshots. That's not to say that decisions between shot times/accuracies shouldn't be important - I just think the calculations should be done automatically. Where the player has all the information to make some relatively complex numerical calculation, he shouldn't need to do the calculating himself.

In some ways it can be of benefit to have simple mathematical costs for actions to better inform decision-making, and as such, the line between success and failure can become thinner. The less informed a player is, the more guessing is required...

Click to expand...

Sure - I'm not advocating concealment of information in this context. I'm advocating a versatile sliding scale action set, along with all the relevant calculation available at the touch of a button (or at the touch of no buttons). Hopefully a sliding scale would often make numeric calculation less important - since the player can take the shot he wants to, rather than judging between two less-than-ideal options.

Simple costs/odds for individual actions don't necessarily make for simple deductions/conclusions in any case. For example, the odds of missing entirely with a 20% 3-shot auto-shot, then a 40% snapshot are: 0.8^3 * 0.6 = 0.3072. That's important information in Xcom, but most players aren't going to bother working it out - even though they have all the requisite information. With or without sliding scales, calculations can become cumbersome (so I'd prefer to de-emphasize them where possible).

I'm also not sure that having clear cut-offs is necessarily an aid to planning/prediction.
For example, take this Xcom situation:
You see a snakeman enter one side of a moderately-sized room.
You have a soldier standing at another door to the same room.

Q: If you go in, can the snakeman react and shoot you?
A: Without arbitrary shot times, almost certainly (perhaps inaccurately). With arbitrary shot times, maybe.

Q: Can the snakeman open the door and shoot you if you wait by the door?
A: Without arbitrary shot times, almost certainly (perhaps inaccurately). With arbitrary shot times, maybe.

Planning is simpler there without clear cut-offs, because then there are only qualitative differences: in both cases, you're almost certainly going to be shot at - the accuracy is all that's unknown. With the cut-offs you have very different situations, and your plan starts to take a "what are the odds that snakeman had 15 TU left?", or "can he open the door and still have 15 TU?" form - with very different results dependent on your best guess.

That's not necessarily a bad thing - it could be argued that the situation is more tense that way. However, I don't buy the idea that simplified action sets necessarily aid planning/prediction. Knowing that "X is about 0.5" is much more helpful in planning than "There's about a 50% chance X is 0, and about a 50% chance it's 1".

I do agree that it's preferable to keep the interface simple for run-of-the-mill play, but I think that a greater set of options can often simplify decision making, rather than complicate it.


[N.B. the particular situation I'm considering right now (if any) is a somewhat daft tangent to this discussion - e.g. there are no guns/bows.... There's nothing wrong with using familiar examples of course, but I'm more interested in ideas that can be applied across all potential actions, in any context, than in stuff which only works for reloading a shotgun at dusk, in high wind, while jogging]

For the realism wing and the simulationists there; I'm asking your opinion on the medieval martial arts system I vaguely described above.

The game (not just combat though, this also applies to some roguish activities like sneaking, pickpocketing and lockpicking) plays turn-based but turns play out in in a simultaneous way. Think of Teudogar, but less simple (or more complex). Turns are relative to the actions you intend to perform and to certain checks regarding interrupts. That said, turn-wise, the world revolves around the player. Whatever action it is you intend to perform, a turn is either a momentary action or a series of preplanned / predicted actions. Other characters also make all the rolls and checks the player does, but there is no syncronisation to "even out the turns". Little (or big) differences between when actions start and end are translated into opportunities, reaction times etc.

Take my prediction example above: you predict actions and reactions. If any of them fails to succeed (note that outcomes aren't black and white as fail/succeed. There are also efficiency levels, eg. the turn doesn't progress to your advantage as you've predicted nor does it to your disadvantage; various other shades of gray) you make rolls to determine if you can retain control and are able to make new instant decisions (where your stress levels and ability to think healthily is also accounted for). Whenever you'd have only one option to advance the encounter, it goes on automatically until an opportunity arises.

All isn't predictions either. You can still quee actions without specific expectations, but there would have to be a limit to that as well, but I'm not sure how I feel about the using MP/AP points. I want it to be as context dependant as possible, removing the need for point calculations (not that they're a bad thing; I'm just aiming for a different thing specifically for this medieval martial arts system). Prediction, queeing or none, you need to rely on perception checks heavily, describing the momentary position of the opponents / targets, what they may be doing next, to be able to make healthy (in your or your character's opinion) actions.

Sure, this doesn't unconditionally shorten the amount of time you may spend on playing your turns and in fact it may even prolong it in theory, but the assumption is that situations where you take 'queed' or predicted / preplanned actions, or automatic one-choice reactions would occur more.

What do you think? Too vague to make anything out of it?

galsiah said:

[In general I'm leaning towards simultaneity - though I'm a long way from anything concrete]
...
[N.B. the particular situation I'm considering right now (if any) is a somewhat daft tangent to this discussion - e.g. there are no guns/bows.... There's nothing wrong with using familiar examples of course, but I'm more interested in ideas that can be applied across all potential actions, in any context, than in stuff which only works for reloading a shotgun at dusk, in high wind, while jogging]

Click to expand...

I think you are making a mistake by trying to generalize the problem, doing so will hinder your ability understand the problem and to offer a solution, you need to go in the exact other way. The more precise you are in describing the problem the easier it is to find the answer you are looking for. And in case you are leaning toward simultaneity it becomes even more important to accurately describing the problem in order to design a solution.

For example –

A fairly obvious "solution" would be to allow units to be positioned continuously anywhere (walls permitting), and for units to be able to put any amount of time into a shot.

Click to expand...

These are mutually exclusive, in order for the character to better his aim while moving he needs to first stop moving, and then, if he want to, change to a prone position.

This is assuming the target is at a great distance away, and this distinguish is important because aiming a rifle to a target 50m away is completely different then aiming a rifle at a target 15m away, aiming for the purpose of burst fire and aiming for single round is different types of aiming, aiming a single handed weapon is a completely different thing then aiming a rifle altogether (although even with single handed weapons you cannot better your aim while on the move).

Hip aim, shoulder aim, and crosshair aim are used for different situation, and dictated ability to act and react (ability to acquiring new targets, ability to move, ability to aim, ability to perceive the immediate surroundings, etc) and if you are going to go the simulation route, you need to understand these differences.

And in case you do understand these differences, I am completely at lost of what you are trying to accomplish here.

@denizsi
That doesn't sound too bad. Perhaps one way to keep things simpler much of the time would be to generate implicit predictions from player actions if he doesn't make any explicitly.
For example, if the player opts to perform a high-front-kick, he's implicitly saying "I predict that my opponent will perform some action to which a high-front-kick is a good counter." He shouldn't need to say that twice, so an automated system could raise his expectation levels for opponent actions in proportion to the effectiveness of a high-front-kick against them (perhaps taking into account the abilities of each fighter).
Of course, if the player has an idea of the specific action he's expecting, he could predict that explicitly. [and this should probably lower the implicitly generated expectations of other actions, in proportion to the strength of the player's prediction: the player has explained his reasoning for the high-front-kick (with a certain confidence), so there's need to spread the odds evenly over all possible moves where a high-front-kick is a good response.]

Hopefully this would make things simpler and smoother - since if the player is simply thinking "A high-front-kick is best here", the game will automatically calculate and make the best possible predictions as soon as he clicks "high-front-kick".

The possible trouble I see with all this is that it turns into something of a thinly veiled statistics exercise. I guess it could be interesting where the player knows little about his opponent's abilities - so a series of fights becomes a learning/adaptation experience.

Where the player knows the abilities of his opponent well (and vice versa), I think it'd be a good idea to introduce a mental element to an opponent's state - and to have each opponent's tactics/tendencies vary differently with his state (mental, physical, positional...). A large part of the strategy with a well understood, well matched opponent would then be to manufacture a situation in which his moves were more clearly predictable. How that were achieved would depend on the opponent - since each would have various behaviour tendencies in different circumstances.

Koby said:

I think you are making a mistake by trying to generalize the problem, doing so will hinder your ability understand the problem and to offer a solution, you need to go in the exact other way. The more precise you are in describing the problem the easier it is to find the answer you are looking for.

Click to expand...

Sure - but I'm not generalizing a specific problem. My problem is specifically general, since I'm not envisaging a setting with much (or any) gritty realism. It should be coherent when I'm finished, but I'm intending to throw a good bit of magic/other-worldly-tech in there. In the specific context where these battles occur, pretty much any mechanic could fit (though of course I'd narrow down themes/actions/explanations at some point).
At the moment I have no firm decisions on specific details. I know I want things to be tactical/strategic; I know I want a wide range of possible tactics; I know I want combats to vary widely dependent on strategic conditions.... Almost anything else is uncertain.

It's just simpler to start a discussion with familiar, tried and tested examples, and to generalize from there. Clearly balancing some more unusual mechanics would be a challenge - in UFO: EU, the two most odd combat mechanics (mind-control and blaster launchers) throw the balance off the most; in XCOM: Apocalypse it's mind-control and instantaneous teleportation.
My aim would be to use a variety of interesting/different mechanics to create an unfamiliar tactical landscape - hopefully with a bit of depth, and without turning things into an absurd, unbalanced free-for-all. [though the odds of my producing anything concrete are, as ever, near zero]
Once I have some nifty mechanics, I can bend the reality of the game world to fit them. (in practice the two would probably emerge together - I'm just saying that I have absolutely no problem with radical setting changes to accommodate interesting gameplay)

And in case you are leaning toward simultaneity it becomes even more important to accurately describing the problem in order to design a solution.

Click to expand...

Sure - once I have firm ideas, I'll be sure to throw them around a bit. At the moment I'm just looking at different systems for ideas on how things might work. However, I'd rather define the "problem" as some interesting mechanics - then adapt the setting to provide a "solution". [I already have a general setting in mind - it just allows me pretty much free reign in mechanical terms]

galsiah said:

A fairly obvious "solution" would be to allow units to be positioned continuously anywhere (walls permitting), and for units to be able to put any amount of time into a shot.

Click to expand...

These are mutually exclusive, in order for the character to better his aim while moving he needs to first stop moving, and then, if he want to, change to a prone position.

Click to expand...

That's not what I mean by "positioned continuously". I mean positional continuity (moving to any point in the world), rather than discrete positions (on a square/hex grid). I don't mean "continuously" in terms of time.
My bad - I couldn't think of a clearer way to say it.

Clearly you're right when my statement is interpreted in the sense of continuous time - aiming while running would be rather silly. (though I don't think the same applies for preparing to fire - inaccurately, of course)

This is assuming the target is at a great distance away, and this distinguish is important because aiming a rifle to a target 50m away is completely different then aiming a rifle at a target 15m away, aiming for the purpose of burst fire and aiming for single round is different types of aiming, aiming a single handed weapon is a completely different thing then aiming a rifle altogether (although even with single handed weapons you cannot better your aim while on the move).
Hip aim, shoulder aim, and crosshair aim are used for different situation, and dictated ability to act and react (ability to acquiring new targets, ability to move, ability to aim, ability to perceive the immediate surroundings, etc)

Click to expand...

Interesting. I doubt I'd use these directly, but the target-acquirement/aim/movement/perception trade-off is workable in many contexts.

...and if you are going to go the simulation route, you need to understand these differences.

Click to expand...

Ok. To be clear, when/if I say "simulation", I don't mean "simulation of the real world" - I mean "simulation of some world" (vs. an abstract system). I'd call any coherent, representational system a simulation - even if it's a coherent representation (vs abstraction) of a fantasy world. I guess it's fair to say that a relatively high level of representational detail helps to get the feeling of coherence too.

One of the main reasons I wouldn't choose to do a game focused on real-world combat simulation is that I'm neither passionate, nor knowledgeable, about real-world combat. There are many people who are both of those things, and such people would be best suited to making such games. Even if I acquired the knowledge, I wouldn't have the enthusiasm.
Having said that, many concepts from real-world combat sims can be useful for any combat system - when viewed through a general/abstract lens. The presentation, the control, the tactical trade-offs, the strategic situation... will all have parallels in any combat context. It's also pretty likely in any physical world that projectiles/explosives are going to crop up in some form - and therefore probably aiming.

And in case you do understand these differences, I am completely at lost of what you are trying to accomplish here.

Click to expand...

You and me both .
Just throwing around ideas (and forgetting that "continuous" tends to refer to time in everyday speak).


On a more general note:
(1) Any thoughts on interesting ways to handle magical combat/conflict (magic missile and fireball need not apply)? [NB, there needn't be wizards/spells, not all "characters" need be physically on the battlefield, victory needn't involve killing funny little men....]

(2) Bearing in mind that I know it's a terrible idea etc. etc., any thoughts on using a physics system as a significant combat component? Emotional/implementational objections aside, what's the thinking on the unpredictability aspect? [e.g. non-physics-based: "aimed shot gives 73.2% chance of hitting for 60 damage", vs physics-based: "blowing up supporting pillar stands ??? chance of collapsing ??? onto ??? for ??? damage"]
I'm wondering if it'd be desirable to introduce such uncertainty. In one sense I like the idea - since it forces the player to strategize on a game-world rather than mathematical basis. On the other hand, I'm also aware that taking away the numbers would probably upset some people - and raise the whole player skill vs character skill issue with a "Why can't my explosives expert tell whether blowing up pillar X will bring down the building??". Would a physics system which allowed some prediction (accuracy according to character skill) be preferable?
[Yes - I'm aware that it's a terrible idea etc. etc., for a variety of reasons.]

THEORY START:

O.k. let me attempt to summaries some of my thoughts on this issue before I move on.

Well, to start, the biggest problem I see with continuity, in many regards is that while, for example, replacing tile based movement with vector based movement can bring many advantages and new freedoms, from a tactical perspective and probably a few others, as long as at the base of the calculation movement is mastered in APs, you are not really removing any kind of arbitrary cutoffs, and visa versa, If you remove AP from movement calculation, then tile based game world makes [/i]less[/i] sense, even though tile based game world still offer other advantages.

If I where asked to attempt to extrapolate from the movement example above to the general case I would say that the first thing you need to look at is the units; these are, in my opinion, the heart and soul of any abstraction system, attempting to delve deeper into the core I would say that arbitrary units are at the foundation of every good abstraction system, i.e. in order to achieve a high order of abstraction I would recommend the use of arbitrary units to represent different elements in the game world and the game mechanics.

Every time you examine these units: how do I use them, what do I with them, what do they represent and so forth, but most importantly when you examine a certain system, an abstracted system (tile base AND AP based in our example), you need to keep in mind that these abstracted, arbitrary* units have a strong connection to each other, so when you do alter one of the sub system (AP OR tile base game world) it will have a very significant effect on other sub system.

*almost by definition in an abstracted system, imho at least, but maybe this needs to be discussed further.

So, in some kind of limited scope conclusion, I would say that the arbitrary cutoffs which are inherently to arbitrary units are a price you need to pay for a highly abstracted system.

This is why I *agree* with your "leaning towards simultaneity", even though it appeared otherwise. I think simultaneity is inherently more suited to offer 'continuity' in the broader sense, however keep in mind that simultaneity and abstraction (through arbitrariness) are on the opposite side of the axis, or if we would now go from the general case to the specific case 'continuity' and 'arbitrary cutoffs' are on the opposite sides of the axis.

I do think however that going the simultaneity route will hamper you ability to abstract the system which in turn will limit your ability to offer tactical gameplay, which leads me to the other important point I think I need to make which is basically the deep connection I personally see between providing highly abstracted system to offering a deep tactical gameplay, but lets not bite more then we can chew on, we can go back to this later.

One last caveat I need to mention about the tie linking abstraction and arbitrariness as I see it is that I don’t think they are necessarily joint at the hip, just that I see a strong connecting between them, they do seemed to fit together quite well (the infamous HP case comes into mind, for better or worse). I'm sure a high level of abstraction can be achieved while maintaining minimum arbitrariness.

END THEORY

Putting my pragmatic hat on...

positioned continuously

Click to expand...

While contemplating on continuity have you asked yourself how much continuity is enough? Would offering the ability to move half a tile or a quarter of a tile would suffice? Because if it does, a simple solution would be to make the avatar occupy 2x2 or 4x4 tiles respectively.

Ok. To be clear, when/if I say "simulation", I don't mean "simulation of the real world" - I mean "simulation of some world" (vs. an abstract system). I'd call any coherent, representational system a simulation - even if it's a coherent representation (vs abstraction) of a fantasy world. I guess it's fair to say that a relatively high level of representational detail helps to get the feeling of coherence too.

Click to expand...

I meant – "if you are going to go the *realism* route..." - which is usually what I refer to by saying simulation, my bad, I miss worded that part.

One of the main reasons I wouldn't choose to do a game focused on real-world combat simulation is that I'm neither passionate, nor knowledgeable, about real-world combat. There are many people who are both of those things, and such people would be best suited to making such games. Even if I acquired the knowledge, I wouldn't have the enthusiasm.

Click to expand...

Honestly, I actually prefer that, I'm not interested in a real-world combat simulation also, too much real life, not enough game.

(1) Any thoughts on interesting ways to handle magical combat/conflict (magic missile and fireball need not apply)? [NB, there needn't be wizards/spells, not all "characters" need be physically on the battlefield, victory needn't involve killing funny little men....]

Click to expand...

Without knowing what magic is in some kind of abstract way (points of interaction/influence, crucial points/levels[as in critical mass], dynamics, scope, homogeneity, detectability, flexibility, maneuverability, etc), no, not really.

...what's the thinking on the unpredictability aspect?

Click to expand...

Highly depends on their effect on the combat/objectives, just a little bit too much will be enough to make the player go "Is this a tactical game or Russian roulette?". The link between player decisions and outcome needs to reach some kind of minimal certainty, and that minimum in most cases is quite a lot.