Jump to content
Durendal5150

Wild Speculation: Approaches to Combat

Recommended Posts

Let's start this one by reiterating the first part of the title. This is all wild speculation on my part. Though some of it is, I feel, extrapolation and inference. A combination of what's been said about the combat in DU, (very little.) with my sense of game design and the probable best ways to make this work.

 

Let's start with what we know and why. It's been well established that DU will have "Lock on combat." So what does this mean? There's no saying with any precision. But what it most certainly isn't is combat with active player involvement in aiming, physically modeled projectiles, or any of the other trappings of most high-end combat centric games. The why should be obvious: The difficulty of tracking all of these disparate elements is likely to overwhelm client and server both. While it might work with very large player counts, it isn't as scalable as the alternative.

 

That alternative of course, is to remove the player's aimpoint and projectile modeling from the equation entirely. Instead we simply determine have the player choose something tos hoot, initiate the attack, and after some fairly simple math, the outcome takes effect.  In most other games, the input scheme for this is rather banal.

 

  • The player clicks on, or selects from a list, their intended target.
  • The player selects the attack to initiate, usually by clicking it from a hotbar.
  • The attack is calculated.

 

This is a tried and true interface method. But it lacks engagement with the player, isn't terribly immersive, and would be awkward in DU's seemingly always-first-person gameplay. So. Provided we have to remain within this framework of "The player selects a target and attack, and everlasting else the system handles," what are some possibilities? Let's start with a potential case of on-foot combat between player avatars. (At this point, as an aside, I ask you to excuse my crude example visuals.)

 

pzmCZs5.png

(Screen gratuitously ripped from one of the tutorial vids.)

 

So in this example, we have a single potential target, and we have a black circle that represents the player's 'targeting area.' (It was translucent blue, but I am a smart man who collapsed the layers wrong.) Since the player has, we assume, his combat mode/weapon tool selected, the target zone appears. There's a potential target in that zone, so it's bracketed in red. We can assume, in the image above, that if the player clicks his LMB and fires the weapon, that this is the target he's going to be attacking.

 

So what if there are more than a single target? Here another key. (Tab perhaps?) Could be used to cycle the brackets to another. Or the RMB used to 'aim' at the target closest to the center of the cursor. It's possible we can also determine if there are voxels or terrain in the way, to what extent, and flub the hit percentage (or disengage the lock entirely) if line of sight is lost, without any extreme levels of computational overhead.

 

So, in this way, it's possible to essentially 'sneak' the lock on system the scope of the game requires it to use into something that has the surface appearance, and many of the gameplay loops and conventions, of a more typical shooter. The attacks hit chances could be modified by aiming down sights. Targets could take cover and benefit from it. Pretty much the whole nine yards. Since precision aiming isn't required as well, this also has the added effect of giving us the tactical elements of more conventional games. (Placement, cover, movement.) without requiring any snap reflexes from the participants.

 

Let's move to a space combat example using some similar assumptions!

4GVDTLe.png

(Also ripped shamelessly from the youtube channel)

 

So here we have a very similar setup. There's a target out there. He's got brackets. And we have an aimpoint of some sort. But in space, we don't need that guy to be *in* our targeting zone. We've got sensors! So we resort to a target selection system more like a traditional space sim. T for nearest target. hold it to select the guy nearest the center of your aim point. All that good stuff. Once the target is within your aimpoint, and you pull the trigger, your attack is calculated just like above. This gets us, once again, all that positioning, flying skill, and maneuvering. Just sans the need for aiming skill, or the need to calculate all the business involved with manually aimed weapons.

 

So ultimately I think the main point here is that I feel like, when NQ says that DU has a 'lock on combat system," there's still a wide amount of room to work within that frame. They could quite possibly make the system feel more action-oriented regardless with the right approach.

 

A couple closing thoughts:

  • A weapons stats coupled with a players stats could govern the size of their aim zone in AvA combat.
  • There's a bit of an edge case when the target is obstructed by another valid target, or a friendly actor. We could possibly just assume that line of sight is broken in this case, or just choose to hit the obstruction.
  • Gunners in larger ships would probably use basically the same system to operate their turrets as the pilot uses in the above example.

 

Last thing, I'd like to keep this discussion as largely as possible about the possible implementations of combat, and their repercussions for the game and gamestate. So I'm politely asking we try to avoid the topics of the nature of weapons systems, (Although mechanics is of course fine.) appeals to realism, potential stats outside of mechanical necessity, and all those sundry topics. Unless you've got a Really Good Reason, of course.

 

 

 

 

Share this post


Link to post
Share on other sites

I thought about possible combat quite a lot too.

 

I think your system may work for avatar vs avatar combat, as they'll have unified hitbox and HP-pool.

 

But for construct vs construct there no way to make it that simple. Unlike games like EVE, there is no HP-pool, but localised damage. So "where" is as important as "if" you hit.

Also constructs can be 95% hidden underground or behind some asteroids, while still having guns sticking out with clear line of sight. There gonna be huge ships, with a lot of small sub-systems, each with different field of fire.

 

So IMO while targeting will be just as you described, the firing itself will be done with ray-casting (hit-scan projectiles). You need it anyway, to determine if specific part is behind cover, so can be used to determine hit chances as well. And hit-scan aren't even close to be as expensive as physical bullets.

 

Also calculating % chance to hit completely remove ability to manually dodge. Sure it can count target speed or angular velocity, but not how unpredictable given manoeuvre is.

 

And I don't want my battle outcome to be determined by RNG. It may work for games with HP, as multiple hits get averages. Here loosing specific part may render your craft useless, while taking multiple hits somewhere else may mean nothing.

 

For AvA auto-targeting + hit-scan would be probably better option than % chance too.

 

 


Another aspect CvC I thought about is how to encourage diversity on the battlefield. Just to ensure that each highly successful fleet consist of both small and large crafts. And that smart engineering is properly rewarded. IMO ships should be in general more effective at fighting ships of similar size.

 

 

First step towards that is how projectiles work against armour (voxels). IMO it would be better if they were AP (2-5) or APHE (6-9) rather than pure HE (1).Armour.jpg

Most block-building combat games go for HE, and it dumb down the combat and engineering to "who has more DPS*armour".

With AP mechanics you get rewarded for using the right tools for the job.

Using too small guns (2, 6) leads to under-penetration, so destruction of less valuable armour (grey), rather than important internal components (blue).

Using the correct gun size (3, 7) deal the most damage. Especially with APHE.

Using too big gun makes you waste some potential on destruction of armour on the other side (4, 8) or even more if it flies right through (5, 9).

 

Low AP guns should have much better DPS/price, to make them the better choice against lightly armoured vehicles. (by price I mean weight, size, energy/ammo usage and actual crafting cost)

It also prevents building bad-looking external grate armour, dedicated to triggering explosion far away from actual armour.

 

For elements similar system could be used, just with HP-pool and armour value. So for shooting off small guns it's better to use other small guns. But not too small.

Damage to elements.jpg

 

Second step are chances to hit. Obvious solution would be random spread, but as I said I'd rather have more skill than luck involved. Turret tracking speed doesn't work against targets far away. So IMO firing delay could nicely ensure that large guns are meant to shoot large targets. 

 

How would it work? Simply each gun has firing delay - time between AI decide to fire and actual shoot being fired. Big guns (with high AP) have more, small - less. During that time, AI can't change the point in space at which it aims. The point is determined by movement prediction, so it's not as simple to avoid it as going fast enough. But changing the movement vector allows you to make them miss. If you're agile and small enough.

Picture to easier visualise:Firing delay.jpg

-Green is predicted movement vector. Orange is predicted intercept point for (too) heavy gun, red is for (proper) small gun.

-Yellow line is actual flight path. It's not curved enough to avoid small guns, but it's enough against big ones.

 

Third thing is ensuring that it's much much harder to destroy vehicle (completely blow it up) than disable it (destroy just vital components) or even temporary disable (damage components in a way that can be repaired in combat).

IMO you should never be able to do the first, unless you waste way too much firepower than the target is worth. Second only if your ship is way bigger (except small  gun-boats that have no repair capabilities due to size). And third should be the most common way to prepare for boarding.

So more of a WWI or WWII armour to firepower balance than modern combat. It's maybe less realistic for far future, but way more entertaining for a game IMO.

 

And last point is NO SHIELDS. They just dumb down localised damage into stupid HP-pool. It's like throwing away half of the whole engineering part of the game for absolutely no gain. Just because of retarded "Shields are so sci-fi! They must be there because reason!".

Share this post


Link to post
Share on other sites

@CalenLoki That's a very well written argument, but I worried you are getting into voxel damage territory. IIRC, NQ has said that they cannot afford to simulate voxels being destroyed or blasted off.

 

Now, I strongly approve of the AP/APHE approach. With (probable) hitscan weapons, each voxel should decrease penetration, based on the armor value of the voxel. This would be countered by the penetration value of the ammo. If penetration value is too high, you go through a ship with little damage (unless your target is behind them). If it's too low, your shot lodges in the armor (potential reducing the armor value? Maybe tol much calculation) Each element along that hitscan should take damage. 

 

With APHE ammo, the final element would take the brunt of the damage. Plus, some surrounding damage. 

 

You could also have weapons have special effects, like Sharpnel against avatars, micro EMP to distrupt systems, or even heat damage to overheat the reactor/engines/weapons/whatever. 

 

Ultimately, I don't see why we would need a lock on system. It'd be helpful for CvC, but AvA or AvC should be free aim. Simpler to account for and more fulfilling. 

 

Edit: I critique you about voxel damage and then wander in and suggest it! *sigh* Round riddle ship corpses are JUST so appealing. 

 

Edit 2: Lock on would help with lag compensation... 

Share this post


Link to post
Share on other sites

@CalenLoki  I'm going to have to agree with @AzureSkye's assessment here. Much of the reason NQ has said they need to use a lock-on combat system is to avoid a lot of the server overhead. It, as Azure notes in his second edit, helps with lag compensation and to hid necessary interpolation. I think everything you've written about selective damage is cool as hell,  and seeing FTD-like damage mechanics in DU would be amazing. But It really is way too much overhead for the limitations they have to work with. At least to the best of my understanding.

 

It may be possible to give *some* realm of selective targeting against certain constructs. possibly by breaking them up into zones and directing the attack against one with the most visibility by LOS or something. To avoid the 'blowing up things that are mostly underground by shooting the antennas' problem. (Good call on that one!)

 

So one thing I do want to address specifically is Calen's comment about skill over luck and Azure's suggestion that AvA should be free-aimed, kinda at the same time.

 

A big part of the system I was proposing (extrapolating, guessing at.) is that it's simple and it specifically removes player skills with high ceilings from the equation. It retains the vital 'intelligent' skills of combat. Positioning, usage of cover, all that tactical stuff. But it explicitly removes the higher ceilinged twitch skill of aiming. Both because of the limits of the server architecture, and because ultimately, despite its presentation, DU is simply not a typical action game. It can use some of their conventions and presentation, because they're familiar and comfortable to most gamers. But ultimately, the twitch-shooter crowd just isn't the core audience, and there are no NPCs. If all combat is PVP, it shouldn't be skill gated off and made the realm of CS vets who can snap-shot you in a quarter second. I think you are right in the assertion that player skill is important and should be rewarded, but it's important to consider which skills should be and to what degree, to make sure the game remains accessible through all of its loops to its core audience.

 

As a last note, I agree with Calen that shields are a questionable addition. Traditional 100% mitigative, 100% cover shields remove all economic risk from committing to battle with a lesser opponent, continuing to balloon existing advantages that larger craft and richer players will naturally have. I don't think the solution is, necessarily, "have no shields," but it's possible to implement them in ways that lessen those advantages. An example I've considered; have shields be pin-point affairs. A shield generator needs an operator, just like a weapon turret. It can be aimed at one enemy ship, and provides its protection against them for the duration of its operating cycle, at which point it has a short cooldown before being 'fired' again. This shifts a little advantage to smaller craft (Which we can assume there will be more of,) as well as reinforcing some of the multicrew gameplay NQ really wants to focus on.

Share this post


Link to post
Share on other sites

It's going to be a really simple foundation isn't it?

 

1. Position as the server sees it.

2. Calc stuff baked into the equipment being used.

3. The reticle just being a tick or cross on hitting specifically so players have to visually track their targets as the human input thing to the above basic server stuff.

 

Whoever said above, different weapons affect different parts of a ship (or equipment) I think solved the issue about local vs overall damage: you just have a specific weapon calc effect, right?

 

Idk about shields? I think larger ships having shields means smaller ships have a harder time against them but they cancel on larger ship combat or else EMP or other shield-knock out weapon?

7 hours ago, CalenLoki said:

Another aspect CvC I thought about is how to encourage diversity on the battlefield. Just to ensure that each highly successful fleet consist of both small and large crafts. And that smart engineering is properly rewarded. IMO ships should be in general more effective at fighting ships of similar size.

I don't agree with this at all. I see a dingy vs a battle-cruiser.

 

Larger are going to crush smaller. So what is the point of smaller?

 

Larger are going to be much more resource intensive to fully equip and also challenging to always crew given coordinating large groups of players simultaneously to play online.

 

So there's always going to be a niche for smaller vessels purely from frequency and perhaps economics of ship output and crew frequency. I mean a sub-crew large ship might get blown up by another large ship: It's better then not to risk the ship...

 

Some over-elaborate warship ecosystem is just going to be a nightmare to design I get the impression? Of course this imho and counter-arguments are very welcome to it.

Share this post


Link to post
Share on other sites
10 hours ago, AzureSkye said:

@CalenLoki That's a very well written argument, but I worried you are getting into voxel damage territory. IIRC, NQ has said that they cannot afford to simulate voxels being destroyed or blasted off.

I've read and watched some NQ statements about CvC combat. And it seems that there will be localised damage (both voxels and elements). Maybe I watched something outdated? What's the most recent article?

https://www.youtube.com/watch?v=efu_129hI9o around 9:45 to 13:00
https://board.dualthegame.com/index.php?/topic/841-ask-us-anything-event/&tab=comments#comment-8072

 

NQ also stated that you can aim at specific point at the ship (which is selected randomly, if the target is too far for you to select anything specific). Then damage is deal within certain sphere from that point. I bet damage will be applied from outside-in, thus it need ray-casting to check what is outside.

And that system need to check if locked target is "valid" which probably means there is clear line of sight, is within weapon reach, ect. So it needs to use ray-casting too.

 

Also don't get me wrong: I'm not advocating for FPS style aiming - it's still done via lock&fire (so AI assisted aiming). Thus it's not affected by lag or interpolation, and not reflex based.

I'm only suggesting (or wildly speculating) how hit chances could be calculated via ray-casting, rather than some arbitrary % numbers.

8 hours ago, Durendal5150 said:

So one thing I do want to address specifically is Calen's comment about skill over luck and Azure's suggestion that AvA should be free-aimed, kinda at the same time.

 

A big part of the system I was proposing (extrapolating, guessing at.) is that it's simple and it specifically removes player skills with high ceilings from the equation. It retains the vital 'intelligent' skills of combat. Positioning, usage of cover, all that tactical stuff. But it explicitly removes the higher ceilinged twitch skill of aiming. Both because of the limits of the server architecture, and because ultimately, despite its presentation, DU is simply not a typical action game. It can use some of their conventions and presentation, because they're familiar and comfortable to most gamers. But ultimately, the twitch-shooter crowd just isn't the core audience, and there are no NPCs. If all combat is PVP, it shouldn't be skill gated off and made the realm of CS vets who can snap-shot you in a quarter second. I think you are right in the assertion that player skill is important and should be rewarded, but it's important to consider which skills should be and to what degree, to make sure the game remains accessible through all of its loops to its core audience.

Only twitch skill I mentioned was evasive manoeuvring in CvC, not aiming. And it only prevents usage of oversized guns against small targets. And it's more of tactical skill to know when and which direction to turn.

10 hours ago, AzureSkye said:

Now, I strongly approve of the AP/APHE approach. With (probable) hitscan weapons, each voxel should decrease penetration, based on the armor value of the voxel. This would be countered by the penetration value of the ammo. If penetration value is too high, you go through a ship with little damage (unless your target is behind them). If it's too low, your shot lodges in the armor (potential reducing the armor value? Maybe tol much calculation) Each element along that hitscan should take damage. 

7 hours ago, Durendal5150 said:

and seeing FTD-like damage mechanics in DU would be amazing.

 

Unlike FTD and other games, DU have one huge advantage: voxels are small. 64 times smaller compared to FTD. Thus we don't need any elaborate penetration calculations, going through voxels without destroying them, partially damaging voxels, ect. We can have just 0/1 system where voxels are either alive or dead. If projectile (ray) damage potential is higher than voxel resistance, remove it. Then damage potential is reduced accordingly, and proceed to the next voxel. Repeat until damage potential is lower than voxel resistance.

Weapons with high AP would be simply those with high damage potential per shoot, while low AP would just spray small bullets.

8 hours ago, Durendal5150 said:

As a last note, I agree with Calen that shields are a questionable addition. Traditional 100% mitigative, 100% cover shields remove all economic risk from committing to battle with a lesser opponent, continuing to balloon existing advantages that larger craft and richer players will naturally have. I don't think the solution is, necessarily, "have no shields," but it's possible to implement them in ways that lessen those advantages. An example I've considered; have shields be pin-point affairs. A shield generator needs an operator, just like a weapon turret. It can be aimed at one enemy ship, and provides its protection against them for the duration of its operating cycle, at which point it has a short cooldown before being 'fired' again. This shifts a little advantage to smaller craft (Which we can assume there will be more of,) as well as reinforcing some of the multicrew gameplay NQ really wants to focus on.

My main concern regarding shields is that they protect weak-spots in armour. Things like engines, radiators, gates, weapons. With shields, all those parts are virtually as durable as anything else. Complete mitigation of damage for the winning side is also important though.

I went for "simply no shields" as that's the good point to start balancing from. Your idea may work, but IMO protection should just reduce % of damage potential, preferably for both incoming and outgoing shoots. Thus armour design and proper positioning is still important, and shield is something you use when you really need extra protection, not all the time.

3 hours ago, MookMcMook said:

I don't agree with this at all. I see a dingy vs a battle-cruiser.

 

Larger are going to crush smaller. So what is the point of smaller?

Of course large vs smaller is always going to larger.

The question is: what if one large ship worth 100kk with 20 crew members fight twenty small, single crewed ships worth 5kk each?

Not so obvious any more.

Small ones have advantage in positioning, agility and acceleration. Large one has much thicker armour, and higher repair ability. Small ships have too small caliber to kill large ship, but they can disable some external equipment. While large ship have it's main armament too clumsy to hit small targets, and secondary armament is out-shoot by small crafts, due to being more stationary.

 

Or situation where two large ships meet fleet with one large and 20 small ships. IMO mixed fleet should win, by first disabling enemy with small ships, then entering with large ship for some deep-penetration.

 

Just like tanks IRL - they are much more powerful than infantry. But for a reason they never fight alone, without infantry support.

And battle-cruisers required cover of destroyers for protection from torpedo-boats.

Share this post


Link to post
Share on other sites
8 hours ago, CalenLoki said:

The question is: what if one large ship worth 100kk with 20 crew members fight twenty small, single crewed ships worth 5kk each?

Not so obvious any more.

Gotcha. Good scenario to consider. I guess large ship plus smaller vs large ship is where the fun will be in these mixed battles then. I'm still wondering how things will scale. I mean a thousand ants vs an elephant... (!) :P

Share this post


Link to post
Share on other sites
8 hours ago, CalenLoki said:

I've read and watched some NQ statements about CvC combat. And it seems that there will be localised damage (both voxels and elements). Maybe I watched something outdated? What's the most recent article?

https://www.youtube.com/watch?v=efu_129hI9o around 9:45 to 13:00
https://board.dualthegame.com/index.php?/topic/841-ask-us-anything-event/&tab=comments#comment-8072

You, sir, are correct. I was thinking of debris and breaking apart ships. From this: https://www.dualthegame.com/en/news/2018/02/08/in-game-ama-feb-3rd-2018-transcription/

Quote

Kulkija: Will there be structural integrity?
NQ-Sophon: You will have “holes” in the voxel structure when damage is taken, but no ship dislocation, which would be cool but taxing a lot on performance.

 

 

I'm disappointed that the AvA combat is going to be a lock&fire system, however I understand the load issues when transversing servers. It just makes me very sad, because I love sniping. Hopefully, it's as transparent as they say it will be, and we can mount up miniguns against waves of infantry.

Share this post


Link to post
Share on other sites
56 minutes ago, AzureSkye said:

Hopefully, it's as transparent as they say it will be, and we can mount up miniguns against waves of infantry.

Multi target weapon ?

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...