Long Range Gunnery

[Spartan-228]

So we finely have weapons that can hit targets further way then 2k, but now we have a new problem.
Wile it is nice to no longer be forced to fight at knife fighting ranges, the question on how to see your opponent let alone hit him at ranges grater than 2k becomes a reality.
Now granted missiles can be set to track an enemy ship and AI turrets don't have to see a target like we do to hit it. However player controlled long range cannons are badly handy caped. All we have to shoot at is a diamond and a very small dot in the middle if you are lucky.
Now cannon rounds are fast, but there is still quite a delay from when you fire to when you hit your target, if you are shooting at something say on the other side of the sector. This means we need something that not only gives us a better view of our target but also helps us lead it.

 

[Winterhome]

Could just get a zoom function.

Lead indicators are a bit controversial here. Some players like the idea, some don't.

 

[SapioiT]

I'm for the zoom!

 

[Tacotown99]

Considering how "advanced" the starmade universe is we should go for a lead or heading indicator but if it breaks gameplay too much a zoom would work just as well. I have no idea how to implement it though.

 

[Comr4de]

A lead indicator would be based around the core. Meaning that shooting that indicator = instacore.

Honestly just git gud. Certain weapons are designed for certain scenarios. You're not going to take down a fighter with a pulse cannon, and you're gonna have a hard time taking down a megaship with a machine cannon. Optimize your ships for their roles.

 

[Spartan-228]

Well I may have come up with a solution, but it comes with a balancing problem.
As humans we have difficulty hitting targets however Al don't, at lest as long as they are within their effective range which is decided by the server admin. Now as long as the turrets effective targeting is set up to be 99% within the turrets engagement rang, which is the default weapons range. The gun will hit and as long as it is powerful enough it will always kill its target, but here is the downside. No one wants all turrets in the game to be that good.

So I had the idea that there could be added a AI targeting enhancer block, that would increase the AI turret or ship's effectiveness at hitting targets beyond the servers default AI 99% range. Also it would be nice if AI using weapons that are enhanced with beams, be able to shoot at targets within that weapons range rather then waiting until they are within the default weapons rang.

 

[BeefBacon]

How about a lead computer? Link it to a weapon system and you get a lead marker. Maybe in turn it could increase range or projectile speed. Alternatively give beam support a lead function so it becomes more useful for Cannon+beam combos (since beam gives missiles a lock-on function it makes sense that cannons should receive a lead marker.)

There's changes in general I'd make to the combat system. Firstly I want focus fire to be removed completely. Second, I want weapons to have dispersion so they're not pin-point accurate. Not much dispersion, mind, and certain effects (like beam support) should increase accuracy. And, of course, I'd like the targeting diamond to encompass the whole ship, not just the core.

However, I doubt that the current iteration is going to be the final combat system, so it'll be interesting to see what the devs do with it.

 

[jayman38]

Well I may have come up with a solution, but it comes with a balancing problem.
As humans we have difficulty hitting targets however Al don't, at lest as long as they are within their effective range which is decided by the server admin. Now as long as the turrets effective targeting is set up to be 99% within the turrets engagement rang, which is the default weapons range. The gun will hit and as long as it is powerful enough it will always kill its target, but here is the downside. No one wants all turrets in the game to be that good.

So I had the idea that there could be added a AI targeting enhancer block, that would increase the AI turret or ship's effectiveness at hitting targets beyond the servers default AI 99% range. Also it would be nice if AI using weapons that are enhanced with beams, be able to shoot at targets within that weapons range rather then waiting until they are within the default weapons rang.

This gives me another idea. AI Self-improvement. Imagine if a server has two AI accuracy settings: starting accuracy and maximum accuracy. When AI first is implemented/built/bought/etc., it starts out with "starting accuracy". Then, when it makes the number of hits that is the same as the current accuracy, that AI's accuracy level moves up a point.

Example:
You are playing on a server with a minimum accuracy of 400 (99% accurate at 400 m) and a maximum accuracy of 800 (99% accurate at 800m). You attach an AI turret, activate it, and head into hostile territory. At first, the turret is terrible, barely scoring any hits. After 400 rare hits, it upgrades to an accuracy level of 401. Much, much later, the turret is at an accuracy level of 550 and finally gets the 550th hit at that level, and upgrades to an accuracy level of 551.

With these numbers, the AI will probably be destroyed before it has a chance to get much better, so different numbers should probably be implemented. (E.g. increase by 10 accuracy points after the number of hits equal to 1/10th of the current accuracy level.)

This mechanic simulates the AI "learning", while gradually reducing the speed at which it learns as it gets better. Actually, with the increasing accuracy, it gets faster to score hits, so the simulated learning doesn't slow down. Maybe someone can think of a good idea (different algorithm) to truly slow the "learning".

 

[Keptick]

Honestly, it's rather easy to nail ships with long range cannons even up to 8km away (provided that the enemy ship is of decent size). And you shouldn't expect to snipe a fighter from a distance anyways.

Not to mention that out of the three long range weapon combos, only one needs aiming skill to use. Practise with it or use lock-on missiles :p

I agree that a simple zoom function could be cool, though.

 

[der_scheme]

Honestly just git gud.

As much as I hate phrases like this, I have to agree here.
Seriously, AI can't hit you at that range if you make use of your ship's strafing capabilities, as they try to match your constantly changing velocities.

 

[Ithirahad]

This gives me another idea. AI Self-improvement. Imagine if a server has two AI accuracy settings: starting accuracy and maximum accuracy. When AI first is implemented/built/bought/etc., it starts out with "starting accuracy". Then, when it makes the number of hits that is the same as the current accuracy, that AI's accuracy level moves up a point.

Example:
You are playing on a server with a minimum accuracy of 400 (99% accurate at 400 m) and a maximum accuracy of 800 (99% accurate at 800m). You attach an AI turret, activate it, and head into hostile territory. At first, the turret is terrible, barely scoring any hits. After 400 rare hits, it upgrades to an accuracy level of 401. Much, much later, the turret is at an accuracy level of 550 and finally gets the 550th hit at that level, and upgrades to an accuracy level of 551.

With these numbers, the AI will probably be destroyed before it has a chance to get much better, so different numbers should probably be implemented. (E.g. increase by 10 accuracy points after the number of hits equal to 1/10th of the current accuracy level.)

This mechanic simulates the AI "learning", while gradually reducing the speed at which it learns as it gets better. Actually, with the increasing accuracy, it gets faster to score hits, so the simulated learning doesn't slow down. Maybe someone can think of a good idea (different algorithm) to truly slow the "learning".

This sounds like a good idea for AI crew manning turrets. BOBBY AI? Meh, not so much; BOBBY AI should just be kind of dumb.

 

[Tacotown99]

Am going to say once a again a simple heading indicator would be vary nice it would simply shou what direction there going in but of course this means that sceama would have to code in a compass of some sort too.this way you could tell if you have a single ship coming at you or a whole fleet.

 

[DivineEvil]

I don't consider either option a necessity. If you have a problem hitting your target using just a target marker, then you are not supposed to hit it in the first place. Considering that we now have the crosshair blink, that notify us as we hit our target, implementing the leading mechanic would change nothing but make Starmade combat even more casual.

Even so, cumulative turret precision increase worth looking into, since their only drawbacks are blind spots and separate shielding.

 

[Winterhome]

Just give us a zoom-in function and I'll be happy. That'll be great for people with "normal" screen resolutions, or anything lower than the default.

I want to be able to hit a target at maximum sniper cannon range if it's standing still. If it's moving, that's an entirely different issue - but being limited by my hardware is a bit annoying.

 

[Spartan-228]

I was doing some reading and I think what we need is a fire control computer.
It would of course use some power when active and could be counteract by radar jamming.
What this computer should do is align your ship with its target taking target speed, heading, and your weapons projectile speed into account.
Now of course it is not going to be effective if you are maneuvering and it will be difficult if you are moving at high speeds, but this is meant for sniper ships so they can lob shells at each other from across sectors.

 

[Valiant70]

This gives me another idea. AI Self-improvement. Imagine if a server has two AI accuracy settings: starting accuracy and maximum accuracy. When AI first is implemented/built/bought/etc., it starts out with "starting accuracy". Then, when it makes the number of hits that is the same as the current accuracy, that AI's accuracy level moves up a point.

Example:
You are playing on a server with a minimum accuracy of 400 (99% accurate at 400 m) and a maximum accuracy of 800 (99% accurate at 800m). You attach an AI turret, activate it, and head into hostile territory. At first, the turret is terrible, barely scoring any hits. After 400 rare hits, it upgrades to an accuracy level of 401. Much, much later, the turret is at an accuracy level of 550 and finally gets the 550th hit at that level, and upgrades to an accuracy level of 551.

With these numbers, the AI will probably be destroyed before it has a chance to get much better, so different numbers should probably be implemented. (E.g. increase by 10 accuracy points after the number of hits equal to 1/10th of the current accuracy level.)

This mechanic simulates the AI "learning", while gradually reducing the speed at which it learns as it gets better. Actually, with the increasing accuracy, it gets faster to score hits, so the simulated learning doesn't slow down. Maybe someone can think of a good idea (different algorithm) to truly slow the "learning".

This would be more appropriate for an NPC crewman, but I really like that idea. Everyone knows the "AI" in "BOBBY AI" stands for "Artificial Idiocy" anyway. AIs can learn, yes, but anything that would increase their weapon accuracy would come stock anyway. Learning would be more like getting better at dogfighting, which is a far more complicated task than aiming.

 

[madmann135]

This gives me another idea. AI Self-improvement. Imagine if a server has two AI accuracy settings: starting accuracy and maximum accuracy. When AI first is implemented/built/bought/etc., it starts out with "starting accuracy". Then, when it makes the number of hits that is the same as the current accuracy, that AI's accuracy level moves up a point.

Example:
You are playing on a server with a minimum accuracy of 400 (99% accurate at 400 m) and a maximum accuracy of 800 (99% accurate at 800m). You attach an AI turret, activate it, and head into hostile territory. At first, the turret is terrible, barely scoring any hits. After 400 rare hits, it upgrades to an accuracy level of 401. Much, much later, the turret is at an accuracy level of 550 and finally gets the 550th hit at that level, and upgrades to an accuracy level of 551.

With these numbers, the AI will probably be destroyed before it has a chance to get much better, so different numbers should probably be implemented. (E.g. increase by 10 accuracy points after the number of hits equal to 1/10th of the current accuracy level.)

This mechanic simulates the AI "learning", while gradually reducing the speed at which it learns as it gets better. Actually, with the increasing accuracy, it gets faster to score hits, so the simulated learning doesn't slow down. Maybe someone can think of a good idea (different algorithm) to truly slow the "learning".

As much as I like this idea it does have numerous faults to it.
First and foremost being how easy it is to destroy the AI and potentially hours of work that can never be recovered.

My concept is a bit different yet more forgiving to block destruction.
Simply put have the AI optimized for a specific range and any range out of that set range would suffer when it comes to accuracy.
This is a very rough idea; short range (0 < 500m, this includes anti-missile defense), Mid Range ( 250 < 1250m), and long range (750 < 2000m).

This way you would have to build ships/turrets for a specific range and as a result purpose and you would have to know what your weapons can do.