Nemesis Response Scaling

[keith.lamothe]

I don't like letting the AI ever setting foot on a planet I control, let alone having a planet be sacrificial.

Game balance and aesthetic concerns are separate.  If you don't want the AI to ever get on your lawn you've got an additional set of goals that the game hasn't given you, which is fine, but it may not let you meet all of them  

Besides, defense in depth would need to draw its resources form the wall they are already trying to breech.

Unless you're using mobile assets in such a way as does not cause attrition against the what the wall will actually have when they hit it.  You can do that through fighting with range/kiting/etc such that your losses are replaced by production facilities back at the wall before the enemy actually reaches it, and/or using the champs which just warp back when they die anyway. 

Concentrated firepower usually is more effective than firepower drawn out over a distance.

Usually, but if you run into a situation where the distance you're capable of concentrating 100% firepower on is not enough distance, you need to make adjustments either to your methods or your objectives

[Cinth]

And that's relevant too.  Those concerns will be addressed when the ST hack is done and I can project full mobile firepower against one of the AIs.  And then there is always the threat of FS coming in (which will when I'm ready).

Back to carriers and issues with them, this just happened in my game in response to a CPA launching. SS start with the release then carrier creation. Out of 37 carrier created, only 3 carriers have 1k units and 6 had 10 or less units.

[Toranth]

Or more defensive depth

If only we had per-planet turret caps, reduced AIP response rates, rebuildable uniques, and the AI wouldn't just stop after overrunning one planet and sit there on the wormhole for 30 minutes before becoming even more annoying threatfleet, so that defense in depth becomes practical...

[keith.lamothe]

@Cinth: there was a change in 6.024 to prevent it from moving a ship to a different planet when putting it in a carrier, which is why you're seeing that.   Though I see why that leads to issues with the 3-second-rule. Perhaps having carriers merge when they're on the same planet would be a good idea, though that could lead to some... "interesting" results

so that defense in depth becomes practical...

We're obviously talking about two different things   Possibly more than two.

[Cinth]

@Cinth: there was a change in 6.024 to prevent it from moving a ship to a different planet when putting it in a carrier, which is why you're seeing that.   Though I see why that leads to issues with the 3-second-rule. Perhaps having carriers merge when they're on the same planet would be a good idea, though that could lead to some... "interesting" results

I've seen it before (carrier spamming) when I would play against 10s (after the barracks change went in so the first CPA would have ships to pull from).  The game has been spewing partially full carriers at me for awhile though.  That with the 3 second rule would be like playing Russian Roulette with carriers, especially if I have more than one choke running.  And ftr, those weren't the only carriers made for the CPA.  I think each planet the AI has left produced at least one (I have 75 planets behind the main chokepoint).

[TechSY730]

Wait, why is the AI bothering to make carriers on a planet when that planet only has 6 units eligible to go in a carrier? I thought there was a minimum number of ships that had to be eligible for a single carrier before the AI would even be allowed to try putting them in a carrier.

And what is this 3 second rule? A ship has to be alive for 3 seconds or more before it can be placed in a carrier? After a carrier is popped on a planet, no carriers can be formed using stuff on that planet for 3 seconds after the pop? Something else?

[TechSY730]

Or more defensive depth

If only we had per-planet turret caps, reduced AIP response rates, rebuildable uniques, and the AI wouldn't just stop after overrunning one planet and sit there on the wormhole for 30 minutes before becoming even more annoying threatfleet, so that defense in depth becomes practical...

For the first three items, there have been plenty of threads about these, so I won't go into this here.

For the last one, yea. There is this obnoxious bug where freed ships sometimes get "stuck" until they get pulled into threatfleet, especially after a planet is lost. They will even sometimes sit there even if there is a totally vulnerable human planet next to it.

I'm pretty sure this isn't some sort of intentional thing to do a semi-contain (trying to prevent you from reclaiming that planet), as the ships in the behavior don't move, even when new enemies come around.


Now if only some of the ships stayed behind (but still moved around on the planet to go after enemies), that would be a cool technique. But right now, too many ships do it, and they stay in place, don't leave the planet when the strategy no longer becomes viable (either too much comes back in, thus outnumbering them, or all adjacent human planets are taken out as well).

This has been reported on mantis several times, and this bug has been around a, to be frank, embarrassing, closely approaching unacceptably long time. (like back in the 4.0 days) 

[Cinth]

Killing a carrier within 3 seconds of a death of a carrier will kick in the compression logic.

[TechSY730]

Killing a carrier within 3 seconds of a death of a carrier will kick in the compression logic.

That, seems...weird.

Why a time based criterion? Shouldn't it be some sort of a current planetary population based criterion about whether the carrier will "compress" its contents on release?

If you have 2 carriers with 200 ships each, why should popping both of them in quick succession cause any sort of "consolidation"? The engine can easily handle 400 ships. Now, if they both had 1000 ships, can you popped one of them, then popped the other without killing a bunch of stuff first, then that would make sense the second one would consolidate, as 2000 does begin to push the limits of the engine.

[keith.lamothe]

Killing a carrier within 3 seconds of a death of a carrier will kick in the compression logic.

That, seems...weird.

Why a time based criterion?

As I explained earlier, there has to be some time delay because otherwise some of the previous carrier pops won't have fully resolved (and thus factored into the unit counts) before the next uncompressed pop.  So if you pop 20 1k carriers with an FS fleet all in one second, the game would be back to trying to simulate 20k individual units on the planet.  In fact it would just drop everything that didn't fit in the 5k limit from the game entirely, but that's not a desirable outcome from my perspective

As I also said, I can try to remove the need for that 3 second rule by having each planet remember how many units are currently in the "uncarriering queue" and factor that in directly rather than being paranoid about it, but that's easier said than done in terms of getting it right (the decision to uncarrier and the actual creation of the units happen in very different parts of the code at different times, due to the MP model of command->execution)

[Cinth]

If it's just me complaining loudly, then I don't see the need to work on it right now (or at all).  I've been dealing with the "3 second rule" for awhile anyway. 

[TechSY730]

Killing a carrier within 3 seconds of a death of a carrier will kick in the compression logic.

That, seems...weird.

Why a time based criterion?

As I explained earlier, there has to be some time delay because otherwise some of the previous carrier pops won't have fully resolved (and thus factored into the unit counts) before the next uncompressed pop.  So if you pop 20 1k carriers with an FS fleet all in one second, the game would be back to trying to simulate 20k individual units on the planet.  In fact it would just drop everything that didn't fit in the 5k limit from the game entirely, but that's not a desirable outcome from my perspective

As I also said, I can try to remove the need for that 3 second rule by having each planet remember how many units are currently in the "uncarriering queue" and factor that in directly rather than being paranoid about it, but that's easier said than done in terms of getting it right (the decision to uncarrier and the actual creation of the units happen in very different parts of the code at different times, due to the MP model of command->execution)

Ah, I didn't realize there was a timed queue synchronization thing going on here. Yea, that could be tricky to refactor.

My bigger concern (involving carriers) are about why you are seeing carriers with like only 6 units in them.
Also, was the thing with the carriers in special forces, when popped, sometimes very quickly just reform back into carriers thing ever fixed?

[chemical_art]

Concentrated firepower usually is more effective than firepower drawn out over a distance.

This is one instance when it is not 

[Cinth]

Concentrated firepower usually is more effective than firepower drawn out over a distance.

This is one instance when it is not 

My choke still held so I guess it still held true in this case. 

[Faulty Logic]

So, rerailing this thread:
Anyone object to rewards being capped by max(#homes, #champs) ?