AI War 2 v0.716 Released! "Activating Mk4 Grinding Machine"

[chemical_art]

The first fundamental question we have to ask is just how common was it in AIW1 to on the fly replace losses instantly? It could be done either by putting manufacturing facilities on a border world or by investing in ships that could produce in enemy space. But I never really got the vibe it was a central pillar of tactics. Rather, it was the exception because of deliberate choices. I bring up the neinzul because that was what they were meant to do. They were very fast to both produce and move to another world but their attrition meant they could only go as far as one world which limited their tactical choices. They served a blunt purpose well but never could effectively serve anything nuanced.
If you wanted to do a deep strike, engage in fleetball tactics, or otherwise not just mindlessly fling units at the enemy you had to use ships that were more costly, time consuming, or otherwise more difficult to replace. Often the fleet you brought into the enemy world was for the most part what you had to work with, which is why camping a wormhole to ensure an easy escape was common. If you wanted to replace losses quickly you had to drag engineers along as well which means your fleetball is dragging around quite a camp following which complicates movements. The very fact it takes more steps to replenish the losses was part of the balance. As of right now though if I wanted to engage in deepstrikes I could bring a factory with me that can reasonably replace any losses I get for no research cost.

It is true that having the economy hits zero reduces the ability to replace losses but with current economics even with just 3 worlds I didn’t experience it. I could lose half my forces and my 2.4 mil economy could tank it just fine. Coming from the AIW 1 mindset of not expecting to replace my forces means I am never in a spot where the fact I can’t is threatening. Yes, I would hit negative values in my economy but it tactically makes sense to retreat long before I get to the point that the economy is the reason why I must. To elaborate I may elect to retreat because the 3x strength AI garrison is converging and overwhelming the dps my fleetball can dish out so I retreat to hit another planet that isn’t stirred up. So I stir up that other planet and can return the first planet and do it all over again until some planet is empty. If I were stubborn and tried to fight that enemy ball to the death I could suffer enough of a tactical disadvantage that my economy may finally sputter, but at least as of yet I've not hit such a point from 80 aip. Perhaps later aip will make it more of an issue.

To give an example of player starship fragility, a siege starship of mk 1 can die from three shots of an AI MK 2 guardian (without bonuses). This means a starship can die in maybe 15 seconds from a single common ai unit if targeted. Stronger starships can last longer, but if they ever get it with that 10x multiplier a single shot from a guardian will kill them Considering that is common for the AI to have infinite range it is just dice rolls on when a player starship can die. If I ever were to experience economic hardship starships would be the first to go, but I experience research hardships instead, so I use them.

[keith.lamothe]

The first fundamental question we have to ask is just how common was it in AIW1 to on the fly replace losses instantly?

A few examples come to mind:
1) Deliberately picking a neinzul bonus ship (fireflies especially, but railpods and tigers were also somewhat common)
2) Spending starting K on Mobile Space Docks
3) Spending starting K on Enclave Starships (or picking TDLs or other drone-launchers; not exactly what you mean but the style is similar: expendable units)

So to answer your question: Not ubiquitous, but fairly common. Not everyone did these things, but many did. Some did more than one at once.

To answer your broader point: yes, having mobile production like we do in AIW2 does represent a significant change from AIWC. There are several other such changes. We might disagree on how big the change is, but that's not especially important.

That leads to some intermediate questions:

1) What exactly are the differences?
2) What other parts of AIWC's design does that impact?
3) How are those downstream impacts dealt with in AIW2's design, if at all?

And the end-goal questions:

4) What do you think is good about the change?
5) What do you thing is bad about the change?
6) Are there reasonable modifications we can make to diminish the bad items (and perhaps increase the good items)?

It is true that having the economy hits zero reduces the ability to replace losses but with current economics even with just 3 worlds I didn’t experience it. I could lose half my forces and my 2.4 mil economy could tank it just fine.

Then the metal costs of operating like that in enemy territory probably need to go up. It depends on exactly what kind of situation you were in.

Coming from the AIW 1 mindset of not expecting to replace my forces means I am never in a spot where the fact I can’t is threatening. Yes, I would hit negative values in my economy but it tactically makes sense to retreat long before I get to the point that the economy is the reason why I must. To elaborate I may elect to retreat because the 3x strength AI garrison is converging and overwhelming the dps my fleetball can dish out so I retreat to hit another planet that isn’t stirred up. So I stir up that other planet and can return the first planet and do it all over again until some planet is empty. If I were stubborn and tried to fight that enemy ball to the death I could suffer enough of a tactical disadvantage that my economy may finally sputter, but at least as of yet I've not hit such a point from 80 aip. Perhaps later aip will make it more of an issue.

There are two points that I'm confused to not see mentioned in this paragraph:

1) Loss of Starships. You mention this separately, but is it not critical in this point that you can't replace starship losses in-the-field? They have to be rebuilt at a facility on one of your planets and moved (possibly by-rally) to the battlefield.

2) Loss of Flagship. Your mobile production facility is tanky, but the AI can catch it, and if you attack a mk3/mk4 planet it can kill it pretty quickly once your main fleetball has been pared down (even at full production, you can't instantly pop out a new fleet). You can't rebuild this at all: you have to bring another flagship (or your Ark) to reclaim it where it died.

Do you find those points to be unimportant?

To give an example of player starship fragility, a siege starship of mk 1 can die from three shots of an AI MK 2 guardian (without bonuses). This means a starship can die in maybe 15 seconds from a single common ai unit if targeted.

Some reference stats:

internal_name	squad_total_health	defense_type	shot_damage	subsquad_shots_per_second	squad_damage_per_second	counters_defense
LaserGuardian_Mark1	100000	Evasion	2280	0.33	759.81	Evasion
NeedlerGuardian_Mark1	100000	Armor	4320	0.25	1080	Armor
MLRSGuardian_Mark1	100000	Structure	1980	0.5	990	Structure
PlasmaGuardian_Mark1	100000	Structure	6450	0.09	585.79	Structure
StealthGuardian_Mark1	50000	Evasion	2308	0.33	769.15	Evasion
SniperGuardian_Mark1	50000	Structure	2750	0.25	687.5	Structure
MissileGuardian_Mark1	50000	Structure	8796	0.09	798.86	Evasion
SiegeStarship_Mark1	25000	Structure	12899	0.09	1171.49	Structure

Two notes:
- Guardians, on an individual basis, are balanced at a stronger level than starships. This is something that could be readily changed, it's just seemed reasonable thus far.
- A mk2 guardian will shred your mk1 squishy starship; yes. Like in AIWC Mark 2s are balanced as twice the hp and dps of Mark 1s. So if you want something to be able to take fire, you're going to want at least Mark 2. The mark differences after that are less stark.

A Laser Guardian can kill a Siege Starship of the same mark in 32 seconds. It has to close to medium range to do so, so the Siege can kite it.

A Needler Guardian can kill a Siege Starship of the same mark in 23 seconds. It has to close to short range to do so, so the Siege is likely to be able to stay away from it, or you can put other units in the way to give the guardian other things to worry about.

A Stealth Guardian can kill a Siege Starship of the same mark in 32 seconds. It also has to close to short range, though its cloak makes that easier to achieve. That said, shouldn't cloaked attack units be a natural enemy of artillery?

A Missile Guardian can kill a Siege Starship of the same mark in ~30 seconds. It can do so from long range; the same range the siege itself has. But the Siege can one-shot the Missile Guardian from that range (unless it's a big pile of other units to spread out the damage).

A MLRS guardian has a bonus against Structure and can kill a Siege Starship in 5 seconds, from medium range. Stay out of medium range of MLRS guardians. This is easier because a Siege can one-shot an MLRS guardian (again, if the damage isn't divided too much).

A Plasma guardian is basically an AI equivalent of the Siege starship, just more balanced on hp vs dps. They can one-shot one another (despite the guardian's tankiness), though if you have a lot of ships with your siege starship it is likely to diffuse the plasma damage and allow your siege to survive longer.

A Sniper guardian is going to ruin your Siege's day, from outside the Siege's range (though it's not infinite). Basically any Sniper Guardians or Sniper Turrets have to be dealt with by other units before you can bring a Siege in range.

Considering that is common for the AI to have infinite range it is just dice rolls on when a player starship can die.

That's not common anymore; in fact it never happens. When shields were removed (0.714, on March 20th) we also changed "Sniper range" from "infinite" to merely "50% longer than long".


In summary:

- Thanks very much for the feedback
- There are a variety of dials we can turn to address these concerns, and I'm happy to do so.
- But first I want to make sure we're talking about the same data.

[chemical_art]

Then the metal costs of operating like that in enemy territory probably need to go up. It depends on exactly what kind of situation you were in.

There is a lot to digest in what you said Keith that I will but I mentally perked up at this idea: This seems to a very elegant solution to my concerns. It addresses my concerns about the economy feeling unimportant, the fact that K isn't spent on mobile production, and that there isn't a need for neinzul ships for using neinzul tactics. Mobile production is still an option but it finally comes with a trade off.

Only other quick bit: I really do feel like sniper guardians do have infinite range. But maybe they don't, it is hard to say since range ratings use descriptions instead of numbers. I'll take your word for it.

[keith.lamothe]

Only other quick bit: I really do feel like sniper guardians do have infinite range. But maybe they don't, it is hard to say since range ratings use descriptions instead of numbers. I'll take your word for it.

You can hold Z while hovering a guardian to see its range.

[chemical_art]

1) What exactly are the differences?

I would guess that there was a research investment and ship trade offs in securing the means to obtain the ability to produce in the field. Neinzul units to have a true impact really needed at least 2 different types and that meant that hacking was necessary to secure that second ship and/or the player did a two homeworld game which is its own can of worms.

2) What other parts of AIWC's design does that impact?
I would say that the K trade off resulted in a weaken of the player in other ways. The K had to come at the cost of something else. A player may get less turrets, less starships, etc. Inevitably it put a sort of onus on the player to ensure they are using that mobile production because the player would be in a static game at a comparative disadvantage.

3) How are those downstream impacts dealt with in AIW2's design, if at all?
By removing the K cost a player has little reason not to use mobile production. As a result tactics shift to using that and the AI on some level is not very good at coping with it. It was OK that the AI wasn't good at coping with it in AIW 1 because the player was more stressed in other ways so it evens out but for AIW 2 there isn't that trade off as of yet. The AI can't really cope with it in a fun way either, for the logical thing an AI would do is kill off the mobile producer first. A bit like the classic rpg you kill the healer first. Actually that comparison can be fleshed out: In AIW 1 the player started with two people in a party, and going for mobile production was having a healer as one of those two. Your dps would suffer and if you are in big trouble if your healing can't keep up but for skirmishing it was ideal. Early game the AI isn't smart enough to really deal with the healer directly but later game it can slam enough dps that a player either manages game flow very well or moves on from the tactic. Later on the party could go two healers if you wanted to go really crazy and double down on it, but it was more common to have it be a secondary role. But for AIW 2 it is like you always have a healer, so you have to from the start balance the AI around that because now it isn't some potential option, it is the option. That in itself isn't a problem, but it can be a challenge to make the AI counter it in a fun way that isn't just kill the healer first.

And the end-goal questions:

4) What do you think is good about the change?
The best thing about this is that it is streamlined. I have been championing reducing curves to ease new players and it would be very hypocritical for me to suddenly say otherwise just because I could make the old system work. It is straightforward and powerful.
5) What do you thing is bad about the change?
Because it is so good and always available the AI has to find a means to counter it for it to feel fun. One thing I have not touched on is that going the mobile route in 1 was so fun was because it was different. Unless you took the research and/or ship choices to secure it then it was a game path that just wasn't available. So if the AI couldn't cope with it well it felt like a player was rewarded for using the "toolkit" of choices to create a combo that was effective. That made it not such a big deal if either in early game or even mid game it felt a bit powerful. The fact it was a game route of sorts made it feel more special and fun, a bit like how the spire campaign was fun because you could take the AI head on with needing to whittle them away which is unlike anything else.  But now that mobile production is mainstream the AI does need to cope with it better otherwise it feels tactically boring.
6) Are there reasonable modifications we can make to diminish the bad items (and perhaps increase the good items)?
Your suggestion of having production on enemy worlds cost more is I think the best solution given time constraints. The system provides a passive trade off that makes mobile production not always the best option. It can allow specialization down the road as certain units can be exempt from this rule, or perhaps the prototype starship can be upgraded to reduce this penalty at the cost of something else. It allows the option to still be strong early game but as fleetcaps increase faster then the economy it loses it's power.

It is true that having the economy hits zero reduces the ability to replace losses but with current economics even with just 3 worlds I didn’t experience it. I could lose half my forces and my 2.4 mil economy could tank it just fine.

Then the metal costs of operating like that in enemy territory probably need to go up. It depends on exactly what kind of situation you were in.
As stated before, very excited about this idea.

Coming from the AIW 1 mindset of not expecting to replace my forces means I am never in a spot where the fact I can’t is threatening. Yes, I would hit negative values in my economy but it tactically makes sense to retreat long before I get to the point that the economy is the reason why I must. To elaborate I may elect to retreat because the 3x strength AI garrison is converging and overwhelming the dps my fleetball can dish out so I retreat to hit another planet that isn’t stirred up. So I stir up that other planet and can return the first planet and do it all over again until some planet is empty. If I were stubborn and tried to fight that enemy ball to the death I could suffer enough of a tactical disadvantage that my economy may finally sputter, but at least as of yet I've not hit such a point from 80 aip. Perhaps later aip will make it more of an issue.

There are two points that I'm confused to not see mentioned in this paragraph:

1) Loss of Starships. You mention this separately, but is it not critical in this point that you can't replace starship losses in-the-field? They have to be rebuilt at a facility on one of your planets and moved (possibly by-rally) to the battlefield.

Part of this is a hold over from AIW 1, where starships compartively had smaller dps in exchange for a much larger health pool. So if I see starships dying relatively quickly my first knee jerk reaction is that they just be weak altogether. It seems in AIW 2 they are more dps focused compared to 1 so I have to remember that their deaths are more inevitable. This is some mental retooling I have to do.

2) Loss of Flagship. Your mobile production facility is tanky, but the AI can catch it, and if you attack a mk3/mk4 planet it can kill it pretty quickly once your main fleetball has been pared down (even at full production, you can't instantly pop out a new fleet). You can't rebuild this at all: you have to bring another flagship (or your Ark) to reclaim it where it died.

Do you find those points to be unimportant?

They are! Very good points to consider. I have not managed to get into mid game yet where Mk3/mk4 planets are factor. This is in part because mk II worlds are still a real slog to progress. They have to be whittled away over many offenses and it is a bit boring I must say. All my commentary thus far in terms of actual experience is only to MK I and II's so it makes perfect sense that down the road things are different.

To give an example of player starship fragility, a siege starship of mk 1 can die from three shots of an AI MK 2 guardian (without bonuses). This means a starship can die in maybe 15 seconds from a single common ai unit if targeted.

Some reference stats:

internal_name	squad_total_health	defense_type	shot_damage	subsquad_shots_per_second	squad_damage_per_second	counters_defense
LaserGuardian_Mark1	100000	Evasion	2280	0.33	759.81	Evasion
NeedlerGuardian_Mark1	100000	Armor	4320	0.25	1080	Armor
MLRSGuardian_Mark1	100000	Structure	1980	0.5	990	Structure
PlasmaGuardian_Mark1	100000	Structure	6450	0.09	585.79	Structure
StealthGuardian_Mark1	50000	Evasion	2308	0.33	769.15	Evasion
SniperGuardian_Mark1	50000	Structure	2750	0.25	687.5	Structure
MissileGuardian_Mark1	50000	Structure	8796	0.09	798.86	Evasion
SiegeStarship_Mark1	25000	Structure	12899	0.09	1171.49	Structure

Two notes:
- Guardians, on an individual basis, are balanced at a stronger level than starships. This is something that could be readily changed, it's just seemed reasonable thus far.
- A mk2 guardian will shred your mk1 squishy starship; yes. Like in AIWC Mark 2s are balanced as twice the hp and dps of Mark 1s. So if you want something to be able to take fire, you're going to want at least Mark 2. The mark differences after that are less stark.

A Laser Guardian can kill a Siege Starship of the same mark in 32 seconds. It has to close to medium range to do so, so the Siege can kite it.

A Needler Guardian can kill a Siege Starship of the same mark in 23 seconds. It has to close to short range to do so, so the Siege is likely to be able to stay away from it, or you can put other units in the way to give the guardian other things to worry about.

A Stealth Guardian can kill a Siege Starship of the same mark in 32 seconds. It also has to close to short range, though its cloak makes that easier to achieve. That said, shouldn't cloaked attack units be a natural enemy of artillery?

A Missile Guardian can kill a Siege Starship of the same mark in ~30 seconds. It can do so from long range; the same range the siege itself has. But the Siege can one-shot the Missile Guardian from that range (unless it's a big pile of other units to spread out the damage).

A MLRS guardian has a bonus against Structure and can kill a Siege Starship in 5 seconds, from medium range. Stay out of medium range of MLRS guardians. This is easier because a Siege can one-shot an MLRS guardian (again, if the damage isn't divided too much).

A Plasma guardian is basically an AI equivalent of the Siege starship, just more balanced on hp vs dps. They can one-shot one another (despite the guardian's tankiness), though if you have a lot of ships with your siege starship it is likely to diffuse the plasma damage and allow your siege to survive longer.

A Sniper guardian is going to ruin your Siege's day, from outside the Siege's range (though it's not infinite). Basically any Sniper Guardians or Sniper Turrets have to be dealt with by other units before you can bring a Siege in range.

Yes part of mental retooling I must do is just that unit production across the board is increased so losses to even star ships are not as bad. It makes sense.

Considering that is common for the AI to have infinite range it is just dice rolls on when a player starship can die.

That's not common anymore; in fact it never happens. When shields were removed (0.714, on March 20th) we also changed "Sniper range" from "infinite" to merely "50% longer than long".


In summary:

- Thanks very much for the feedback
- There are a variety of dials we can turn to address these concerns, and I'm happy to do so.
- But first I want to make sure we're talking about the same data.

[chemical_art]

I decided to throw up some data points compared to AIW 1 (7/7) in terms of speed:

Time to make a cap of fighters, bombers, and frigates: 5 minutes with me adding 10 engineers.

Total cap: 300.

Enemy ships on first MK I planet: 20. 3 turrets, this possibly included special forces.

Popped command center and massed on a second planet by 9:00. I had 280 ships.

Second MK I planet: 22 ships. Half did not aggro at the start. Was able to pop the world within 12 minutes. First wave warning at 10 minutes.

This is what I mean about the early game being a slog. I managed to do in in 12 minutes in 1 what would take me 45 minutes in 2. This was just with triangle ships: No starships, no bonus ships, no research done at all. They were two easier AI, so I cannot say for sure. My game had to stop at this point because AIW 1 got very unhappy with me alt tabbing, but it gives an idea.

[chemical_art]

Before I run off I did a second game. Same as the first except I choose two Turtles this time.

I was able to pop two MK I worlds and an MK II world in 12 minutes. The attack then stalled because an AI Eye was gumpy I had too many ships and a riot ship killed the survivor's engines. This attack thus was foiled not to due to garrison forces but just from me (intentionally) just being a brute.

[keith.lamothe]

I would guess that there was a research investment and ship trade offs in securing the means to obtain the ability to produce in the field. Neinzul units to have a true impact really needed at least 2 different types and that meant that hacking was necessary to secure that second ship and/or the player did a two homeworld game which is its own can of worms.

Both fireflies and railpods could be used by themselves to great effect, but sure.

I would say that the K trade off resulted in a weaken of the player in other ways. The K had to come at the cost of something else. A player may get less turrets, less starships, etc. Inevitably it put a sort of onus on the player to ensure they are using that mobile production because the player would be in a static game at a comparative disadvantage.

True.

I think your "healer" analogy is very helpful. Ultimately in AIWC if you figured out how to use a "healer" (mobile rebuilding) and a "tank" (mobile forcefields), your main fleet's overall ability to do its real job ("dps") was massively magnified. Not because it did more damage, but because it could actually survive to do that damage.

Players who never figured that out just went out with a bunch of dps classes and were at a serious disadvantage. They were still able to win, due to balance, but the players who could exploit other "classes" did much better.

So with AIWC I wanted to make sure that you had ready access to those massively-important tactics, so that we could balance one game instead of balancing both for "players who know the key combos" and "players who try to do a 5-man dungeon with 5 mages" (in MMO terms).

The end result was that ultimately "tanking" is just OP, and was cut with the departure of shields. And if one looks at equivalents in modern naval combat and the more rigorous sci-fi analogues, it does make sense that "a ship that can redirect all incoming damage against an entire fleet to itself" would be massively OP.

In AIWC tanking is still OP, it's just usually only available to players with one particular bonus type (shield bearers) or who are playing Spire.

But is it good to have those kinds of OP tactics kind of off to the side? For AIWC, yes, it's reasonable. Sure, some players will take Shield Bearers, Riot Starships, Protector Starships, and Enclave Starships and will just run away with things (because the game has to be balanced to allow players not doing all those things to still win). But it's fun.

My hope here was to have a game without that degree of balance-outlier. Either the tactic is balanced enough to be mainstream-ish, or it needs to not be available outside a Fallen-Spire-type option. Does that make sense?

My hope with the "healer" role is to maintain the user-convenience of it, while bringing it within reasonable levels of balance.


I take your point about the AI not having good ways of dealing with healers. It can be taught to assassinate flagships if it has the local firepower to do so. That's less of a murder-sentence than it would be in AIWC due to the lack of infinite-range weaopns. But we'll see if that kind of kill-the-healer tactic turns out to be desirable.

Your suggestion of having production on enemy worlds cost more is I think the best solution given time constraints.

/Yea, we may try something like that soon. Where your metal costs are multiplied by X+1 where X is the number of hops back to your nearest planet. So rather than deep-striking being a binary proposition it just gets progressively more difficult to support logistically.

Alternatively the metal costs could be kept the same and the build-rates could be divided by X+1 or something like that. So it wouldn't ruin your metal stockpile, but eventually your healers just wouldn't have enough throughput to help enough.

And we could have techs that allow you to reduce that penalty (or penalties), so we get back closer to AIWC's model of "you can have really good healers, it just costs K". Since those "stats" would be computed on the planet instead of on the ships we wouldn't need to massively complicate the math. Though having techs that don't unlock actual ships is a hurdle we'd just need to jump for that one (not too bad, if it solves a problem like this).

Part of this is a hold over from AIW 1, where starships compartively had smaller dps in exchange for a much larger health pool. So if I see starships dying relatively quickly my first knee jerk reaction is that they just be weak altogether. It seems in AIW 2 they are more dps focused compared to 1 so I have to remember that their deaths are more inevitable. This is some mental retooling I have to do.

Well, I don't want starships being flimsy here. Except for the ones that are really supposed to be kept out of enemy range (Siege and Sniper). So it may be that I just need to make starships have a higher balance baseline than fleet ships, and put most of that increase into hp instead of dps. That's what I did turrets recently.

So it might be something like:

1) Turrets: balanced at 5x the normal amount of "strength" (baseline balance points) per unit of Science, but immobile.

2) Starships: balanced at 3x, mobile, but cannot be replaced in the field.

1) Fleet ships: balanced at 1x, mobile, and can be easily replaced in the field (if you aren't too far from home without the tech for it).

I decided to throw up some data points compared to AIW 1 (7/7) in terms of speed:

Time to make a cap of fighters, bombers, and frigates: 5 minutes with me adding 10 engineers.

I ran a similar test in AIW2 a moment ago, and my fighter/bomber/corvette fleet was ready to go out the wormhole at 1m 10s. Not surprising, given the streamlining there.

Enemy ships on first MK I planet: 20. 3 turrets, this possibly included special forces.

Popped command center and massed on a second planet by 9:00.

Much more resistance in my case, of course (though I plan to halve mk1/2 planet turrets for next version), and I had cleaned the planet of all resistance by about 10m 30s. Much longer than it took you due to your longer build-time, but not 20 minutes or something like that.

The AI units didn't really aggro against me at all until I got close, due to the low strength of the no-starship fleet.

Second MK I planet: 22 ships. Half did not aggro at the start. Was able to pop the world within 12 minutes. First wave warning at 10 minutes.

I researched 4 mkI starships using my starting science before attacking the second one; that started around the 11m30s to 12m mark, and was done by the 16m mark. The AI did generally aggro after me there because I had so much more strength than they did.


I didn't get fancy, except to pull my missile corvettes out of the way so the enemy plasma guardians didn't get in range on the first planet (none of those on the second planet, thankfully), to avoid a cap-wipe of those.

Oh, I did get a little fancy on the second planet by dropping 10 squads of needler turrets between my fleet's initial position and the first chunk of AI fleet ships that came after me. I don't know how much that helped, but I'm sure it reduced the number of enemy bombers that got within range of my fleet.


Anyway, I think we're in range of finding a reasonable middle-ground on the issue of the first few attacks. I think AIWC's first-neighbor planets were so weak that it was more of an interface chore than a real challenge to take them, so I was going for something a bit less tame.

[etheric42]

Well, I don't want starships being flimsy here. Except for the ones that are really supposed to be kept out of enemy range (Siege and Sniper). So it may be that I just need to make starships have a higher balance baseline than fleet ships, and put most of that increase into hp instead of dps. That's what I did turrets recently.

So it might be something like:

1) Turrets: balanced at 5x the normal amount of "strength" (baseline balance points) per unit of Science, but immobile.

2) Starships: balanced at 3x, mobile, but cannot be replaced in the field.

1) Fleet ships: balanced at 1x, mobile, and can be easily replaced in the field (if you aren't too far from home without the tech for it).

If anything, many fleet ships get a bonus because they are more "granular" and receive the balancing bonus from it, so a cap of fleet ships (before taking damage) has better stats than a single starship (to repeat: until the fleet ships take damage).

Turrets I get being balanced upwards (due to lack of movement not receiving enough of a balancing bonus... I'm assuming because you didn't want buildings to have a bajilion HP) and I get the point of "starships can't be built in the field", but I am wondering if you couldn't just use an existing lever for them.  Things like assault starships have "normal" metal/fuel costs compared to "high" for bombers, so bombers get some extra balancing quality to them compared to that starship.  Perhaps all the starships need to be moved into even higher levels of metal/fuel costs.  3x metal in exchange for 1.5 multiplier (that could get spent on HP by upping it to a higher tank class).  x1 fuel for another 1.5 multiplier.  A fleet of high-mark starships has CONCENTRATED power that gives until the last drop, but is harder to replace and manage than a swarm of fleet ships.

The science multiplier system makes the absolute disparity even worse.  Bombers are high fuel, high metal, midlowcap granularity for (1.15*1.1*0.9=1.14 (assuming multiplicative stacking)), wheras the starships are normal fuel, normal metal, and but at least receive a buff for their granularity (since the "extremely low cap" has 2 per mk and a 0.7 multiplier, starships cheat a little and get 1 per mk with a 1.0 multiplier) for (1*1*1=1)

So a Mk1 bomber has ~14% bonus over a Mk1 starship (before paying for things like range, etc).  Mk2 doubles your stats and doubles your cap, Mk2 bombers are 4.56 compared to Mk1 and Mk2 starships are 4 compared to Mk1.  Still a 14% bonus over those Mk2 starships, but 14% of a bigger number and since you had to pay to even get the Mk1 starship, even more reason to continue to invest in fleet ships.

If starships receive a boost for being exorbitantly expensive to create/maintain, you wouldn't have to "cheat" with their granularity bonus and they would be a lot more imposing on the battlefield.  If you give too much free power to starships, then that could invalidate fleet ships (for people looking for optimal).

/Yea, we may try something like that soon. Where your metal costs are multiplied by X+1 where X is the number of hops back to your nearest planet. So rather than deep-striking being a binary proposition it just gets progressively more difficult to support logistically.

Alternatively the metal costs could be kept the same and the build-rates could be divided by X+1 or something like that. So it wouldn't ruin your metal stockpile, but eventually your healers just wouldn't have enough throughput to help enough.

And we could have techs that allow you to reduce that penalty (or penalties), so we get back closer to AIWC's model of "you can have really good healers, it just costs K". Since those "stats" would be computed on the planet instead of on the ships we wouldn't need to massively complicate the math. Though having techs that don't unlock actual ships is a hurdle we'd just need to jump for that one (not too bad, if it solves a problem like this).

It's already dangerous to take your flagships deep striking (since they have to fly through enemy planets if they retreat and if you lose them they are harder to recover), I'd be cautious about hitting them with another penalty for an already high-risk maneuver.

No, I'm not saying this because I'm having trouble coming up with a UI element to show the player the increased costs/build time for deep strike building.  There's that too, but....

The end result was that ultimately "tanking" is just OP, and was cut with the departure of shields. And if one looks at equivalents in modern naval combat and the more rigorous sci-fi analogues, it does make sense that "a ship that can redirect all incoming damage against an entire fleet to itself" would be massively OP.

Yeah, AEGIS in the US Navy is kind of like that (as a kind of damage-sharing-mechanic) and is pretty OP.  The other way carriers (and healers or squishies in RPGs) protect themselves is through creating safe spaces for themselves.  Whether that is through getting air superiority in the navy or various "locking" mechanics in RPGs (attacks of opportunity, controlling aggro, etc).  Mobile tractor ships (and I guess field-created tractor turrets) create that kind of "tanking aggro" by holding x strength ships out of range of your core.  Without tools like that easily available, I'd hate for the AI to send strike teams straight at my flagship where my only hope to deal with them was to hope to DPS them before they got in range.

(As a note, flagships were previously noted as "too squishy" by Badger just last month, so I think there were some AI changes to deprioritise them.)

p.s. power distribution nodes seem to be incredibly plentiful and not give much in the way of rewards when attacking them on Mk1 planets.  I can micro bombers around to each one, but it takes forever compared to just taking out their turrets.  I can set some ships on FRD to try to damage them simultaneously, but it seems to assign one fighter to each one to be pretty useless.  And then if the planet controller starts repairing them I really feel like I have wasted my time.  If there were about half as many, I think they might seem like a more reasonable alternative.

[chemical_art]

My hope here was to have a game without that degree of balance-outlier. Either the tactic is balanced enough to be mainstream-ish, or it needs to not be available outside a Fallen-Spire-type option. Does that make sense?

It does, but it also makes the game ultimately shorter lived in terms of its game's life cycle. It is precisely those outlier strategies that keep a game fresh and allows for a larger number of total number of strategies. Part of making the game drag is that it is the same tactic game after game as right now. Homogeneity is great for balancing but can make for repetitive gameplay.   

...

Alternatively the metal costs could be kept the same and the build-rates could be divided by X+1 or something like that. So it wouldn't ruin your metal stockpile, but eventually your healers just wouldn't have enough throughput to help enough.

I would go with build cost. Making it build slower will just increase the need total time of skirmishing before a player can feel comfortable clearing a planet. With price there is a trade off in that a player has to decide if it is worth the economic downtime to make it on the fly will increasing time just means a player is spent waiting in a secure position (there isn't really a trade off, it just is).

And we could have techs that allow you to reduce that penalty (or penalties), so we get back closer to AIWC's model of "you can have really good healers, it just costs K". Since those "stats" would be computed on the planet instead of on the ships we wouldn't need to massively complicate the math. Though having techs that don't unlock actual ships is a hurdle we'd just need to jump for that one (not too bad, if it solves a problem like this).

You could make the prototype ship have its own MK line. There is plenty of precedent of ships not just getting better combat stats but additional benefits as well. Nothing wrong with spending K to have a ship that is better able to handle a MK III or even IV planet.

Well, I don't want starships being flimsy here. Except for the ones that are really supposed to be kept out of enemy range (Siege and Sniper). So it may be that I just need to make starships have a higher balance baseline than fleet ships, and put most of that increase into hp instead of dps. That's what I did turrets recently.

So it might be something like:

1) Turrets: balanced at 5x the normal amount of "strength" (baseline balance points) per unit of Science, but immobile.

2) Starships: balanced at 3x, mobile, but cannot be replaced in the field.

1) Fleet ships: balanced at 1x, mobile, and can be easily replaced in the field (if you aren't too far from home without the tech for it).

That works and would put it more in line with AI guardians.

...

Anyway, I think we're in range of finding a reasonable middle-ground on the issue of the first few attacks. I think AIWC's first-neighbor planets were so weak that it was more of an interface chore than a real challenge to take them, so I was going for something a bit less tame.

I think those first neighbor planets are very key in their current strength. For three big reasons:
1. A new player isn't likely to know the tricks of the game. Tricks like low starships to avoid agro, not attacking without full caps, or bringing in the prototype starship to replace losses are bread and butter tactics but they have to be learned. If you don't do any of the above three things for example I think any attack on AIW 2 will stall. It should get to that point, sure, but not at the start. That will frustrate a player and be the door for the 2 hour refund if a player cannot get a single planet.

2. The game actually was pretty well balanced for reasonable games if someone wanted to secure the immediate neighbor planets so as to provide a layer of defense. At 7/7 there was no problem in grabbing them. They provided a quality of life anchor in that the AI would have to penetrate the outer worlds and gave a player an economic base to boot. In addition, it gave enough K so that a player could start to flesh out what they wanted and thus a player could start making the game interesting. The AIP increase that resulted meant that waves could actually stand just a bit of threat and often the AI would get a new unlock as well which further fleshed out things. This in many ways actually made balancing easier because rather then trying to make AIP 30 interesting you could instead balance it at 100 being interesting which gives a lot more breathing room for the (new) player.

3. The most important is the most fundamental in making what AI Wars is. The concept of an AI starting with ignoring you entirely but eventually getting to the point where it will always crush you*. That former part is something that can be forgotten for an experienced player but it is just as important as the latter part. The AI has to feel weak at the start, it is almost part of the game's psychology. You have to give that initial high of being able to swipe with ease against the AI.  It draws you into the game, wondering how it is the AI is finally going to get you. It really allows you as a player to choose how fast you accelerate the game. If you want to go all in and pop 4 worlds in 20 minutes, you can! Or you can take your time taking those 4 but not because the AI is stopping you. Eventually the AI will stop you, earlier if you are new and later if you are experienced, but the journey to that point where things "get serious" is a key part of the game. So much of the later gaming content is not explored right now because the early game is too long. To put it another way I cannot think of  many occasions were a new player thought the early game was too long. Maybe it felt long to an experienced, because they were doing some relatively gamey strategy, but the general response I got if anyone were to say that the early game was too easy was "Advance, it won't stay that way for too long  ". I feel like that vibe needs to return to AIW 2. In my AIW 1 games, even despite the thousands of hours I played on it, I felt the pull to play more because of that ease I was smashing planets. I wanted to see what tech that next planet might hide, that golem that could shape my game, that new obstacle the AI might try to stop me (that didn't involve massive garrisons). I don't get that right now in AIW 2 and it really is a detriment.

[keith.lamothe]

You convinced me on making the mark 1 neighbors easy to take

[keith.lamothe]

0.718 is out now; not all discussed changes are in it (it was mostly just a matter of cleaning a few things up to get the new in-game-ui out). Release notes to follow in the morning after I get some sleep, just fyi.

[chemical_art]

You convinced me on making the mark 1 neighbors easy to take

That last paragraph of mine was a lot more feelings based then numbers based. But it was heartfelt. The fact I could clear away planets with such ease at the start of AIW 1 compared to the instant slog of AIW2 made me want to continue playing a lot more despite playing it dozens of time. Psychology is real, and the thrill of bashing the AI to the point they finally get smart is just so enjoyable in terms of an early game. I want to emphasize it is a curve of this AI getting smart, I was able to knock one AI planet after another going increasingly deeper into AI terroritory 3 times (meaning the third planet was three hops away) because the AI doesn't care how far away I get from my planet in the big picture. They would garrision their homeworlds tightly and their homeworld border worlds (mk IV) almost as much, but they didn't have the care if the player plunged into "human border" worlds or they moved in a single direction. Another key aspect from AIW 1 is that aside from V and IV's the worlds started with minimal garrisons, like maybe 10% of their max. They had the potential to be filled in and would gradually over time (with border worlds getting the lions share), but never did they start feeling like they were even 20% full. Right now I feel like the worlds start feeling 75% full. To place it a different way, the AI garrisons for all worlds started almost empty aside from the most absolute key worlds but over time they will fill up with border worlds getting the most. Right now it feels like all AI worlds are already filled which just overall increasing the feelings of grindyness and removes the value of the player trying to control AI reinforcements. I remember a strategic strategy of breaking an AI border world and swarming 2-3 worlds behind it before they could garrision up, a blitzkrieg of sorts. With AIW 2 such a thing is not possible, I must grind them all out. Likewise there was even a value of leaving an strategically isolated border, an odd planet that had 5 enemy wormholes. I would hold it not because it could take waves well, but because it could force the AI to deploy reinforcements across 5 additional planets to that isolated border while the main offensive would continue in a different direction. Again, if the worlds are all filled such a strategic nuance is lost.

While writing this I saw your new post. We can all use sleep. But I feel like good data was exchanged today. If you have the stamina I want to keep slugging these ideas  .