Chokepoint Balance

[keith.lamothe]

Keith, given that you are more worried about checkpoints being very much required in Fallen Spire and high level game play

Actually I'm not worried about it being required in FS   FS, past a certain point, makes sense that anything outside your "front line" is going to die the next time the enemy seriously tries to kill it.

[Toranth]

Looking forward to Diazo's impending thread on this.  A question I've been ponder is, do we really need 10 difficulty levels for a minor faction?  Would 5 work better?

Maybe, but basically the upper 5 of the current scale are for when you want to intentionally unbalance your game   Which you can then adjust back into balance; i.e. taking dyson-sphere-10 but going against 2 Heroic AIs or something like that.  No idea if that would be balanced or not, but I imagine it can be entertaining.

Just for history, a Dyson-10 is completely shut down by a Mk II Laser Guardpost.  Those new guardposts are mean.

Well you cover your entire Homeworld from head to toe in Gravity Turrets.  Make sure your Command Station is covered in plenty of forcefields.  Then partially build (to 90%) a bunch of Matter Converters, in case they take a more...roundabout path and knock everything else first.  Unlock MKI and II Heavy Beam Cannons, and build them all in front of your Home Command Station Shields for an unbelievable amount of piercing damage.  Keep you entire force on your HW, on the closest wormhole to your base.  If you have a Champion, you can shield the Comm Station 5-6 times for an insane amount of health, which also gives your shields time to repair.

For the sake of argument, how is this different from a chokepoint?  You still lose if your single fortress world falls; it just takes a minute or two to die if your non-HW fortress world is destroyed.

P.S. I can certainly say whipping boy is an extremely prominent strategy for big co-op games. I'm not sure how viable a no/low chokepoint strategy is when you have eight players and giving everybody a second planet's worth of income is an additional 120 AIP. The additional resource points certainly help, but I'd like to see some shared resource gain from new command stations in co-op games if possible, for more of a resource curve rather than stepladder.

This is part of what I see as the problem.
Even in a normal single-player game, going from 1 HW to HW + 3 protective worlds is a jump from 10 to 70 AIP.  For that 7x increase in AIP, wave sizes are now 8.5x as large (before ingress-count modifications).  In addition, that 8.5x wave is directed against only one of your three defensive worlds.  Unless the AI is *required* to spread attacks evenly among all possible systems at all times (including Exowaves and Threatfleet), it will ALWAYS be better to reduce the number of defensive worlds.  This trivialy reduces to "The best defense is defending only one system", and the chokepointing problem.


To refresh some memories, here are some previous threads we've hit this topic.
Defense Supply Mechanic
Wave Size and Timing based on Points of Ingress
Is AIP too inhibitive?

[TechSY730]

P.S. I can certainly say whipping boy is an extremely prominent strategy for big co-op games. I'm not sure how viable a no/low chokepoint strategy is when you have eight players and giving everybody a second planet's worth of income is an additional 120 AIP. The additional resource points certainly help, but I'd like to see some shared resource gain from new command stations in co-op games if possible, for more of a resource curve rather than stepladder.

This is part of what I see as the problem.
Even in a normal single-player game, going from 1 HW to HW + 3 protective worlds is a jump from 10 to 70 AIP.  For that 7x increase in AIP, wave sizes are now 8.5x as large (before ingress-count modifications).  In addition, that 8.5x wave is directed against only one of your three defensive worlds.
[...]
Is AIP too inhibitive?

While I do think the AIP costs for taking out a planet could use some tweaking, I don't think by much. The game has been balanced around 25 AIP (or was it 30?) per planet for a long, long time. Changing this without at least considering everything else could really hurt balance (at least in the short term)

What I will take issue with is the exponential scaling of the AIP for the higher difficulties. Even with the tiny exponent, it still grows too fast IMO. I would rather see a polynomial increase. Increase the base strength (at AIP 10) if needed.
Also, it would be nice to see a smaller slope for the linear growth that the lower difficulties use, again, increase base strength if needed.

I think these would help a good bit with AIP preventing player from getting too "skittish" about what defensive techniques to try.

Unless the AI is *required* to spread attacks evenly among all possible systems at all times (including Exowaves and Threatfleet), it will ALWAYS be better to reduce the number of defensive worlds.  This trivialy reduces to "The best defense is defending only one system", and the chokepointing problem.

I'm not sure if you can say trivially. Right now, the scaling with the number of inpoints for waves acutally decreases max wave size as inpoint goes up. This makes 2 and 3 chokepoints actually viable now (note, I said viable, not necessarily balanced).
Note, exos aren't doing something like this, which is actually one of the points we have been going over in this topic.

Yea, multi chokepoint and no-chokepoint may still need more help, but I don't think they are bad enough compared to single chokepoint setups to write them off "trivially".

[keith.lamothe]

For the sake of argument, how is this different from a chokepoint?  You still lose if your single fortress world falls; it just takes a minute or two to die if your non-HW fortress world is destroyed.

Because if your home command station is actually within the field of battle, it can be killed by a force which would not have been able to actually defeat your forces in a straight-up battle.  Raid Starships come to mind.

[Histidine]

I think I'd be fine with +1 AIP on CS death if we got proper countermeasures to cloaked units. On my last completed game (8/8 with FS 5) I wouldn't be surprised if I lost 100 command stations over the 25 hours I played to embedded Zenith Reprocessors, even though I got annoyed enough to unlock Decloakers (cap of 4  )halfway through.

[Hearteater]

Hmm, I wonder if there is a stat to track that.  I'd be curious how many CS I lose in a game.

[keith.lamothe]

Hmm, I wonder if there is a stat to track that.  I'd be curious how many CS I lose in a game.

Sure, just add up the loss amounts for each CS type.  Though perhaps that's a lot to dig through.

[Hearteater]

Just checked the only win saves I have hanging around:

On 8/8, I lost 14 Command Stations over the course of 25 hours (I did an AAR on this...Neinzul game)
On 7/7, I lost 18 Command Stations over the course of 20 hours (this was my first win ever, I kept it for nostalgia)

[_K_]

Uh, wait wait wait, how exactly does this "AIP on CS death" thing help against chokepointing? If anything, it works in its favour.

Chokepointing is more like an "all or nothing" approach, where your whipping boy either holds, making you lose nothing, or it dies, probably resulting in massive damage.
No CS die if it holds well enough.

For a distributed defence approach, chances of failure for any single system are actually considerable. Its just that the AI will just have to go bang it's head at the next layer after beating one specific system.
CS death is a common occurence there.

I can easily lose 30+ CS over the course of a rather long, but succesful game. It usually means i just keep losing some frontier systems a bunch of times, with an occasional severe blow that wipes out half my territory.

[Hearteater]

The conversation was a bit rambling, and spanned a few threads.  +AIP on CS, which I assure you, will never happen, wasn't about choke points.  The gist of it is: +AIP on CS makes your actions in the game more meaningful.  Especially losing ground.  Since the choke point issue came up because defending Advanced Factories is hard, some talk of making Advanced Factories invulnerable/repairable came up.  That just makes losing ground completely meaningless.  Unless you actually get hit with +AIP whenever you lost a CS.  Tada.

[keith.lamothe]

+AIP on CS, which I assure you, will never happen

I was actually seriously considering adding an (optional) AI Plot to that effect.  +1 AIP and a temporary but sharp "blood in the water" wave boost or something like that.  Similar reactions when it kills an AdvFact or it detects you really got hammered by that last attack (no +AIP on that one, just a temporary "if I give a little more push, can I win?")

Or would that not help?

[Hearteater]

I would personally love it.  I've been experimenting with some pretty abusable defensive strategies that involve constantly sacrificing Command Stations instead of building a choke point.  I don't have it all ironed out, but I'm basically winning via attrition and I have it almost completely automated so far.  Once I've got my build down I'm going to try it on 9/9 or maybe go nuts on 10/10 and see if I can break things.

But I would absolutely turn that option on all the time.  If I could set a range from +1 to +10 AIP/CS that would be perfect.

[eRe4s3r]

What's stopping us from scrapping the CS to prevent that +AIP ? Or do you want AIP on scrapping too? (Am I the only one who sometimes shifts station types around?)

And who keeps factories the whole game? I go there, build to max and scrap the CS ... when you do that, AI usually ignores you and the factory ^^ Also I usually only go for Advanced research.. value of new basic unit unlocks is FAR beyond a few MK4 ships. The larger the fleet, the further it gets.

Besides that, I am all for more situational responses by AI ^^

[_K_]

+AIP on CS, which I assure you, will never happen

I was actually seriously considering adding an (optional) AI Plot to that effect.  +1 AIP and a temporary but sharp "blood in the water" wave boost or something like that.  Similar reactions when it kills an AdvFact or it detects you really got hammered by that last attack (no +AIP on that one, just a temporary "if I give a little more push, can I win?")

Or would that not help?

I assume it could work if it was a single event with a cooldown, so that it gets a boost for wrecking one border system, giving it a chance to capitalise on that, but if multiple systems go down, it doesnt just hammer the player even harder. We want the player to be able to recover after losing half his territory, right?

The +1 AIP still feels wrong, but with that "blood in the water" thing, how about letting the AI free some of its defences in systems 1-2 hops away from the destroyed system? Since the system is dead, those newly made threat ships would probably go follow the attack up, and as i understand, thats is your intention.

But anyway, there's already the punishment of having to spend time and resourcess on rebuilding the defences on the lost world. If it takes way too little time, maybe its a problem with the rebuilding speed, not the AI's inability to follow up its attacks? Are you sure you are not trying to invent a complex solution when you can just tweak resource costs instead?

Am I the only one who sometimes shifts station types around?

You can replace it by bringing a new constructor, no need to destroy the CS.

[keith.lamothe]

But I would absolutely turn that option on all the time.  If I could set a range from +1 to +10 AIP/CS that would be perfect.

Don't you mean +2/+4/+6/+8/+10?

What's stopping us from scrapping the CS to prevent that +AIP ? Or do you want AIP on scrapping too? (Am I the only one who sometimes shifts station types around?)

That's a good point, actually.  It would basically need to trigger whenever a CS died for any reason at all, because otherwise it's easily sidestepped by scrapping (or moving it via another colony ship).

Probably some people would still use the option, but many would have serious eye-rolling-syndrome if it hit you with AIP for changing the type of command station on one of your planets.  But there's not really any reliable way that lets them do that without triggering the condition without also letting them cheese out the whole thing.

Hmm.