Fractal Softworks Forum

Please login or register.

Login with username, password and session length

Author Topic: CR: (yet another solution proposal)  (Read 3687 times)

xenoargh

  • Admiral
  • *****
  • Posts: 5078
  • naively breaking things!
    • View Profile
CR: (yet another solution proposal)
« on: February 11, 2019, 02:52:42 PM »

OK... so, last time I talked about the issues with CR, I built a little test-mod that:

1.  Forced everybody to stay in a circular arena.
2.  Allowed for fast Retreats.

It worked just fine, but I have another, simpler idea.

How about, "if you don't hold an Objective for more than 1 minute, you start losing CR"?

I'll be building a test-mod for this one; pretty sure it solves for all of the Things CR Was Supposed To Fix (but largely killed Fun doing so):

1.  No corner / wall-kiting in a Solo Frigate / OP Destroyer / etc., because you'll lose to CR death.
2.  Objectives would have a very serious reason to exist, rather than being either totally ignorable minor bonuses or huge benefits to combat per se.
3.  Holding every single objective, on the other hand, would not be necessary, so a certain amount of flexibility would exist.

This has the advantages, however, of:

1.  You can have nice toys (killer player-frigates, etc.) without being on a timer-leash so short it's like being handed the last bite of a candy bar.
2.  You don't feel like you need to put Hardened Subsystems on every single ship at the high end, where endurance starts mattering a lot because the enemy has a lot more Stuff than you do.
3.  Coupled with the (very, very welcomed, imo) fix to "get your missiles back by Retreating", etc., this would appear to resolve the CR issue for good, imo, if it works.

Thoughts?
Logged
Please check out my SS projects :)
Xeno's Mod Pack

Blothorn

  • Commander
  • ***
  • Posts: 116
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #1 on: February 11, 2019, 03:26:50 PM »

Fast ships could still avoid combat indefinitely as long as the enemy didn't have enough ships to camp all objectives (and a sensible, not-designed-to-be-exploitative AI might do so naturally--retreat from enemy, see they are at risk of losing all objectives and go to cap one, repeat). That could be solved, but I am having trouble offhand thinking of rules that are both simple and reliable.

Conversely, this would shut off some player styles--a hit-and-run fleet could easily find itself able to pick off ships rapidly enough to avoid a PPT loss but unable to hold points against a large AI fleet.
Logged

xenoargh

  • Admiral
  • *****
  • Posts: 5078
  • naively breaking things!
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #2 on: February 11, 2019, 03:35:04 PM »

There isn't a solution that's simple and solves all possible play-style issues, honestly. 

My last solution worked (from a technical POV) at solving the worst abuses- corner-camping and using the walls- but it felt a bit wonky otherwise. 

Holding objectives with a fast ship and camping indefinitely is boring, but you can't win so long as the enemy holds multiple points, so, at some point, you have to get the enemy's objectives cleared, or kill them.  I don't mind that being the choice.
Logged
Please check out my SS projects :)
Xeno's Mod Pack

Blothorn

  • Commander
  • ***
  • Posts: 116
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #3 on: February 11, 2019, 03:48:52 PM »

Sure, the *player* can't abuse rotating points in a fast ships because it stalls the battle and players get impatient--but the AI is not constrained by patience, and I would like to be very sure that the AI can't or won't do it, even emergently. Even rare battles where I have to retreat superior forces because I can't catch the AI would be very frustrating.
Logged

xenoargh

  • Admiral
  • *****
  • Posts: 5078
  • naively breaking things!
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #4 on: February 11, 2019, 04:05:43 PM »

Ah, there's that.  That's a solvable problem, but not without creating the opportunity for players to abuse it; the easiest way would simply to tell the AI that it's not allowed to simply run away if it's the only ship left.
Logged
Please check out my SS projects :)
Xeno's Mod Pack

Ranakastrasz

  • Admiral
  • *****
  • Posts: 702
  • Prince Corwin of Amber
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #5 on: February 12, 2019, 07:02:33 AM »

Ah, there's that.  That's a solvable problem, but not without creating the opportunity for players to abuse it; the easiest way would simply to tell the AI that it's not allowed to simply run away if it's the only ship left.

Captain! We're the last one left, We must retreat and inform the commander!.
Never! We must go out in a blaze of glory! ATTACK ATTACK ATTACK!
----
In other words, yes, but you would have to make the AI chose to die instead of live.

----
TLDR:
Not really sure how to fix, but some thoughts defining the problem, and how other games seem to resolve it.

----
The solution we are looking for is one in which you make the middle case infeasible. Attack or retreat, but not kite until the other side dies of old age. Fight, pull back to regenerate, and fight again, or retreat outright if badly outmatched. Don't just kite and stick around.

CR tries to do this, but didn't really work. It caps battles, more or less, but didn't solve the problem and added new problems.
----
I seem to recall a Warcraft III map called "Warlocks" in which there were several methods to ensure an unstable equilibrium would snowball into a winner eventualy.
In the game, each player could throw projectiles that did damage and knockback, and they were surrounded by a sea of lava/death. If you touched the lava, you lost health quickly.
Each time you took damage, you also accumulated knockback amplification. I don't recall if that made you take more knockback in the future, or your target, but I believe it was the first. This made it so that as the battle went on, you were more likely to be knocked into the lava after taking a hit.
Lastly, the lava rose slowly, shrinking the battlefield, forcing players closer together, and ensuring the match ended.

These mechanics ensured the match ended brutally, and avoiding combat only helped so long. It also didn't favor one player over another, at least not much.
---
Archean Order went and ramped up fighter speed, and scaled speed by ship size such that avoidance of combat was very difficult. Added regenerating missiles and better PD, such that missiles and fighters were emphasized over direct weapons. It created it's own problems admittedly, but did away with CR for the most part without breaking things. Almost all ships had fighter wings, and fighters tended to be attacking something at all times, so it tended to force attacks or retreats.
---

Lets see..

Shrinking battlefield. (CR decay ramps up as you are further from fighting/objectives)
Battle avoidance CR (CR decays for kiting more than fighting. Not sure how that would work)

Neither of those feel right, but it might be worth looking at.

Again, what we need is a way to force retreat or combat, but not avoidance.

Why do ships avoid combat? They don't think they can favourably engage, but this may improve with positioning or if an ally assists.
Why do ships engage in combat? They are trying to destroy enemy ships. They are trying to force enemy ships into bad positions, or attack/defend a position.
Why do ships retreat? They don't think they can favourably engage, and don't expect that to improve. Tends to occur at low CR or badly damaged hull, in which durability/stats are crippled.

How can we change these pressures, such that the second and third options occur more often, and the first only as direct combat support (Flanking)

Add solid objectives? Those objectives on the map certainly help in some cases. However, speed and range help you win a kiting war, but we don't WANT a kiting war.
Control all beacons to win? doesn't fit with campaign mechanics.

Ramp up boosts over time? Range/speed increases by 5% every 30 seconds. If control point captured, power resets, but grows twice as fast until reaching old point.
This sounds interesting. You capture and hold a point, requires aggression and defensive play, but not backing down. If you get pushed away from a point, you fall behind, so you MUST hold it.
As a battle drags on, the effect amplifies, until double+ range or speed are granted, such that you must retreat.
This doesn't really feel fair though, but does give something to fight over as a priority.

Objectives that force a win/lose feel wrong, since the objective seems to be to either force a retreat, surrender(not a mechanic yet), or destroy all enemy ships (or a combination thereof)

Its all just, How do we force an engagement, without it feeling forced, such that kiting is not a viable strategy? Other games do it, by time constraints in various ways, or protecting something irreplaceable. If you have to defend a control point or lose, you have to defend that control point or lose. Kiting won't save you, because you end up moving away from it and letting them capture it.
If you have a time constraint, wasting time causes you to lose, but if the opponent has the same constraint (like CR) it doesn't do much.
---

In fiction, why do battles occur? Generally, to steal or destroy something.
Ships, loot, locations.
Steal the ship or the loot.
Capture the location.
Destroy the ship.
Destroy vital intellegence or components.
Bombard the planet from orbit.
Destroy the space station.

Why do these battles not involve kite-fests? Generally the aggressor outmasses the defenders, and the defenders can't retreat for some reason.
Starwars - Interdicter suppressing FTL. Gotta destroy interdict or get distance first. Ambush involves being surrounded. Some ships escape to FTL some destroyed. Mostly a retreat scenerio
Halo - Covenent Fleet attempting to glass planet, civilians still present. Heavily outnumbered, but can't just leave because civilians, critical infrastructure and resources. Usually a loss and painful retreat.
Stargate - Supergate. Stop them at the bottleneck, before they spread out and wreck everything. Enemy too powerful, forces a retreat after blockade smashed.

In the first, There wasn't anywhere to run, and once escaping the location they just retreat.
In the second, the planet is a critical defense point, and you don't want to retreat, but it is forced after massive losses.
In the third, you are defending a bridge, and retreat lets the enemy pass unopposed.

In Starsector, Why are battles not forced? Well, Combat occurs because I want ships/bounty/loot. I find a target, chase them down, and engage. This is in space, the only objectives are the actual ships, being destroyed. Forcing a surrender is impossible, that isn't how the game works (yet). Only destroying a ship lets you get something. There is nothing special about this area of space, so the enemy retreating isn't unreasonable.

You try to encapsulate the enemy fleet to avoid them escaping, or lock them down with emp missiles or similar. They try to retreat, some will engage in fighting. Some will decline both fighting and retreat, instead kiting and trying to find an opening. The kiters are the last to not retreat or die in combat, and continue doing so. Why don't they attack? they will die. Why don't they retreat? Their AI thinks they might still have a chance, enough to not retreat, but not enough to engage, and you cannot force them to engage or retreat with what you have.

They are stuck between two extremes. Retreat or attack, both are things they don't want, so they wait, and kite. If you can't push them into one of those, you get this situation.
---
What did CR Do?
Added ship upkeep.
Caused prolonged fighting to decay ships, biased against fighting.
Caused kiting to have an upper limit.
---
Archean Order's fighters and missiles and speed.
zero flux is 5% flux in this mod. Many weapons produce no flux. Shield damage (Critical for kiters) slows ships down, and flux decay while shields are up is crippled. (but not gone. 70% or so reduced)
Fighters have long range, are faster than the fastest frigates (short of hyperion, which still has low CR) and hit shields forcing flux slowdown. Thus, slowing speedsters down such that others can catch up and force an engagement. Kiting tends to not work long at all, and the enemy dies or retreats.
Long range missiles have the same general effect as fighters, but not to the same extent.

Causes other issues, but eh, it seemed to help a lot with the issue
« Last Edit: February 12, 2019, 07:44:42 AM by Ranakastrasz »
Logged
I think is easy for Simba and Mufasa sing the Circle of Life when they're on the top of the food chain, I bet the zebras hate that song.

Cigarettes are a lot like hamsters. Perfectly harmless, until you put one in your mouth and light it on fire

Megas

  • Admiral
  • *****
  • Posts: 12159
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #6 on: February 12, 2019, 07:25:04 AM »

The reason why Electronic Warfare 1 and Coordinated Maneuvers 1 is so good is because player can completely ignore objectives (which can die in a fire).  With 500 map size and gigantic endgame enemy fleets much bigger than mine, points get captured anyway by accident.  In a significant battle where both sides are roughly a match, both sides hold the line, which shifts a bit, almost indefinitely until CR runs out.  Since enemy has the numbers advantage, that is bad for the player, and the playership needs to do something to break the line.

Kiting is a real problem early in the game with Apogee start.  Sometimes the enemy simply backpedals to the wall until my Apogee can corner and kill it.  AI in 0.9 is almost as cowardly as in 0.8.

Kiting in endgame battles is also a problem with smaller map sizes.  With 200 map size, player can deploy or two or three big ships, and the enemy just trickles their ships in and drags out the fight.
Logged

Plantissue

  • Admiral
  • *****
  • Posts: 1231
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #7 on: February 12, 2019, 01:06:25 PM »

An interesting idea, but what happens if both sides holds at least 1 objective each?
Logged

Ranakastrasz

  • Admiral
  • *****
  • Posts: 702
  • Prince Corwin of Amber
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #8 on: February 12, 2019, 02:33:24 PM »

If both sides hold an objective, it means both sides have the force to do so still. If a ship is kitinf, thet cant hold an objective, so it works fine.
Logged
I think is easy for Simba and Mufasa sing the Circle of Life when they're on the top of the food chain, I bet the zebras hate that song.

Cigarettes are a lot like hamsters. Perfectly harmless, until you put one in your mouth and light it on fire

xenoargh

  • Admiral
  • *****
  • Posts: 5078
  • naively breaking things!
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #9 on: February 12, 2019, 02:51:10 PM »

@Ranakastrasz:

Interesting about the Warcraft III mod (sadly, my knowledge of all of those is pretty limited, other than DoTA).  Sounds like they invented PUBG early, lol.  So, I guess that would be one option; keep shrinking the battle-space.  It feels kind of artificial, though.

The issues here are:

1.  Endless-kiting scenarios are boring and potentially abusive.
2.  Some ships inevitably get to kite.
3.  Stalemates are totally boring.

So, we don't want there to be major incentive to kite, and we don't want stalemates.  There needs to be a winner.

In a situation where there are only two ships left, and neither wants to engage the other, we've got a stalemate.  The AI either needs to charge in and die, or Retreat.  Either way, there's resolution that isn't just an un-fun CR drainage situation.  I think that's fair enough, and it'd be pretty easy to code into the Admiral AI: some rule that says, "if only two ships remain and it's been 3 minutes, make a choice" oughtta be sufficient.
Logged
Please check out my SS projects :)
Xeno's Mod Pack

Ranakastrasz

  • Admiral
  • *****
  • Posts: 702
  • Prince Corwin of Amber
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #10 on: February 12, 2019, 07:02:52 PM »



While you probably cant see the knockback aml, the rest is pretty clear. Stalemates just cant be maintained in that environmemt.

But yea, shrinking battlespace is just weird and doesnt work here.

Logged
I think is easy for Simba and Mufasa sing the Circle of Life when they're on the top of the food chain, I bet the zebras hate that song.

Cigarettes are a lot like hamsters. Perfectly harmless, until you put one in your mouth and light it on fire

Megas

  • Admiral
  • *****
  • Posts: 12159
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #11 on: February 14, 2019, 01:13:46 PM »

In a situation where there are only two ships left, and neither wants to engage the other, we've got a stalemate.  The AI either needs to charge in and die, or Retreat.  Either way, there's resolution that isn't just an un-fun CR drainage situation.
This is one thing I do not like about Starsector AI.  It acts like the Black Knight from Monty Python.  It can be hopelessly crippled due to no CR (because it refuses to retreat when CR decays), but it will happily drift around taunting the other side to finish the job.  If both sides have no CR, and ships cannot move, then all player can do is exit the game.
Logged

Plantissue

  • Admiral
  • *****
  • Posts: 1231
    • View Profile
Re: CR: (yet another solution proposal)
« Reply #12 on: February 14, 2019, 01:44:46 PM »

If both sides hold an objective, it means both sides have the force to do so still. If a ship is kitinf, thet cant hold an objective, so it works fine.
Your response only presuming that both sides only have 1 ship each.

What I am asking in any case is that if both sides hold one objective each, CR decay wouldn't activate as both holds an objective, so neither has a direct incentive to attack anyways.

Other games solve the problem of both sides preferably wanting to defend by simply allowing both sides to withdraw safely if both wants to withdraw or designating an attacker and defender and combat can only be initiated by the attacker, who will have to attack at a disadvantage, or let the defender go freely. As it is the combat system in starsector seems to be set in stone as it hasn't changed since the start and changing it would require a bit of thought.
Logged