Fleetball woes; strategic choice in AIW2

[etheric42]

I think all these changes are fine (well maybe not distributing damage to control groups, that mechanic (control groups) isn't engaged enough to put such a primary benefit on it yet and it encourages weird micro where you put specific tank ships only on that ship's control group and not your glass cannons).  But I do want to say one thing:

I would summarize the above as "that's already there", except that it's more varied than just having one "anti-big-stuff" mechanic. It's kind of a combination of the examples you gave later.

I don't agree.  These are similar mechanics, but have different implications in play.  The idea of the anti-cap ships are these are specialized weapons that don't do much (if at all) to smaller ships.  Badger mentioned Snipers were making things a pain already because you couldn't piecemeal a planet and they would be a pain even if the Ark wasn't there (due to targeting missile frigates, or just doing "normal" damage instead of 4x damage).  And even if there aren't snipers (or whatever weapon type counters that particular flagship's defenses) there's still a large quantity of ships doing 1x damage that can take it out (you said "generally individual units in AIWC or AIW2 just die in heavy combat unless they're (i) covered by tons of shields or (ii) less likely to be targeted" the latter I think is kind of a cheat and the former has problems because it takes the role of the shield starship).  Then there is the problem of monitoring multiple weapon types if you have more than one cap ship in an engagement.  Keeping capships basically immune to normal weapons and then having a specific class of weapon (that doesn't do much to normal ships) (with only bombers crossing the divide, but they are short ranged and fragile enough to be intercepted) helps make that divide very clear to the player.

But what's there is okay, just throwing out/clarifying my ideas.

[keith.lamothe]

One thing I think may help is to increase the difference between "weapon shooting at things it is good against" and "weapon shooting at everything else". Right now the bonus is only +300%, so basically everything can still kill everything, it just takes longer.

If that was changed to +900%, I think there'd be more of the desired relationships, like plasma being great against capital groups with their inevitable shields, but not very threatening to other stuff, and snipers being deadly to structural stuff but much easier to endure for other units.

I'd have to readjust one of the global constants and one of the balance formulas to make this kind of tuning work (so that non-bonus damage becomes less lethal rather than bonus damage becoming more lethal), but it would be a quick change.

[x4000]

Sounds like a good experiment to me.  And in general sounds like it might be a big win.

[etheric42]

I agree, sounds like a good experiment.  What I'd want to look out for in testing is what this does to optimal micro.  Does it substantially increase the desire for micro to ensure good matchups?  What can be done on the automation layer to deal with that?

[TheVampire100]

One thing I think may help is to increase the difference between "weapon shooting at things it is good against" and "weapon shooting at everything else". Right now the bonus is only +300%, so basically everything can still kill everything, it just takes longer.

If that was changed to +900%, I think there'd be more of the desired relationships, like plasma being great against capital groups with their inevitable shields, but not very threatening to other stuff, and snipers being deadly to structural stuff but much easier to endure for other units.

I'd have to readjust one of the global constants and one of the balance formulas to make this kind of tuning work (so that non-bonus damage becomes less lethal rather than bonus damage becoming more lethal), but it would be a quick change.

This sounds liek the best solution to me and would emphasize on the "rock-paper-scissors" mechainc, that is a core element of AI War.
Wghen I play AI War, I want to use the ships that are the most effective aganst my enemies and not simply "put everything up because it does not matter that much anyway".

Edit: While we are at it, I want to suggest to increase the speed of fighters. I find them a little underhwelming at the moment compared to other ships. I used the hell out of fighters in classic because of the fast movement speed compared to other fleet ships but I barely notice a difference here, especially since they get shot down really fast, have a close combat range and many planets have either engine damage or gravity generators (or both).
So more speed would help their role as "first front line fighters", so they can actually close the gap and fulfill that role instead of being swatted like flys. More engine health for them might also help, hard to slow down but little hp.

[keith.lamothe]

I agree, sounds like a good experiment.  What I'd want to look out for in testing is what this does to optimal micro.  Does it substantially increase the desire for micro to ensure good matchups?  What can be done on the automation layer to deal with that?

The autotargeting, as in AIWC, is extremely robust in ensuring that ships target the things they can do the most damage to (unless there's an overriding concern). It's still possible to have suboptimal targeting where there are better targets just barely out of range that theoretically it's better to hold fire and wait until you can hit those, but that's an edge case (no pun intended).

While we are at it, I want to suggest to increase the speed of fighters.

Yea, we need more speed variance in general, but the UI doesn't currently offer a decent way to have mixed-speed groups work. AIWC's group-move would help, though something "smarter" would be better (so stuff farther from the destination goes full speed until it's "caught up", etc).

But yes, I'd like to make fighters faster and either bombers or missile corvettes (probably the latter) slower, etc, while still maintaining a fair degree of coarseness in ship speeds so we don't have everything spreading out because they've all got slightly different speeds.

[etheric42]

The autotargeting, as in AIWC, is extremely robust in ensuring that ships target the things they can do the most damage to (unless there's an overriding concern). It's still possible to have suboptimal targeting where there are better targets just barely out of range that theoretically it's better to hold fire and wait until you can hit those, but that's an edge case (no pun intended).

I agree about the auto-targeting.  The micro-removing parts of AIWC were what drew me to it originally.  But the Micro I was referring to was positioning, making sure your strengths are concentrated against their weaknesses but not exposed to their counterstrengths.  Some of it is good: seeing a planet with a lot of X, sending Y;  Seeing a cluster of ships X, sending Y.  Some of it is bad: during a murderball engagement, moving all your bombers to the right flank so they stop shooting things they aren't strong against, then moving them back once more ships show up on the left.  If there was some way to make them "stay in the general formation" but drift laterally to engage preferred targets.  (Also, anything to automate or de-emphasize getting a concave.)

[BadgerBadger]

We should probably implement a "Less efficient targeter" for lower difficulty levels.

Edit: https://bugtracker.arcengames.com/view.php?id=19464 tracks this issue now

[x4000]

We should probably implement a "Less efficient targeter" for lower difficulty levels.

That's a really good point.  Used only by the AI, of course.

[keith.lamothe]

We should probably implement a "Less efficient targeter" for lower difficulty levels.

That's a really good point.  Used only by the AI, of course.

public class DumBotTargetSorter ...

BlunderBotTargetSorter would be great, though it would need to modify the filter too

[BadgerBadger]

So the list of things to do from this conversation to my mind are as follows:

• Give the player one type of turrets for free at the game start
• Emphasize Rock/Paper/Scissors by making bonus damage a lot higher
• Buff flagship health/nerf Ark health
• Have a "The AI doesn't like it when you have your Ark on a non-human planet" mechanism. Maybe like Deep Strike from AIWC (or like hacking), the AI will start spawning ships on a nearby planet on a suicide mission toward the Ark; This could be an option you can disable (like AIWC) to make the game easier.
• Tweak AI Budget to spend fewer resources on Special Forces and more on Waves/Reinforcements/other things
• Tweak unit speeds (eventually)
• Implement an inefficient target for easier AI difficulties

Incidentally, to go with the unit speeds discussion, lets imagine that I have Fighters/Bombers/Missile Corvettes and a single shield starship in an early battle. I have all my fleetships hiding under the Shields. I Right-Click on a target to say "Kill this". Once the missile corvettes are in range of the target they stop moving and start shooting. The fighters/bombers/shield starship keep going. This means that snipers immediately kill all my missile corvettes. This winds up being frustrating.

Note that if you aren't bringing your Ark onto planets, Gravity Turrets become brutal. They either need to be nerfed, not placed be default except on Nasty planets, only available to the Gravity Master AI Type, or something like that.
Edit: Also note that without your Ark, you really don't have mobile Tachyon Emitter capabilities unless you luck into Sentinel Frigates or something in an ARS....

[x4000]

All of those todo list items sound good to me, and most of those sound like they aren't that time-consuming to implement, either.  Not that I'm the one who has to do it, so I can say that freely.

Regarding the gravity turrets, I think maybe just making those have a 10% chance on nasty planets and a 25% chance in general on gravity master planets might do it?