AI War state of the game

[chemical_art]

20 minutes into an 8/8 game on 2.0, i got a 1.6k unit wave warning. I ragequit.

For comparison, with my 10/10 mad bomber game, I think I was dealing with 400ish ship waves (granted, MK II ships, so 800ish effective MK I ships). Granted, in the 4.0 era that would be a double wave as well...

[TechSY730]

One thing I have argued is that the wave sizes at AIP 10 are too small, and need to go up, in return for the wave size not growing quite as quickly with AIP. This will make the AI a bit more consistent with its difficulty, and give a more interesting situation in the early game, where you don't have enough that you can ignore the subtle tactical stuff like matchups, and don't have enough to just throw away large numbers of ships. Also, this would reduce the dominance of ultra-low AIP games, as you don't get as much reward for avoiding AIP (though it still would be important, but maybe not "ALWAYS hunt data centers before taking any planets" level of importance, which seems too high for all but AI difficulty 9.8 or 10).

[LaughingThesaurus]

One thing I have argued is that the wave sizes at AIP 10 are too small, and need to go up, in return for the wave size not growing quite as quickly with AIP. This will make the AI a bit more consistent with its difficulty, and give a more interesting situation in the early game, where you don't have enough that you can ignore the subtle tactical stuff like matchups, and don't have enough to just throw away large numbers of ships. Also, this would reduce the dominance of ultra-low AIP games, as you don't get as much reward for avoiding AIP (though it still would be important, but maybe not "ALWAYS hunt data centers before taking any planets" level of importance, which seems too high for all but AI difficulty 9.8 or 10).

Something that worries me is that I did hop into a 10/10 game with friends around just to see how quickly I would lose. Got right on to producing a whole bunch of fleet ships and defenses, grabbed Mark III harvesters as was recommended in some kind of AAR, and I just barely survived what I swear was an unannounced wave of 200 raiders and two starships and a guardian who all just decided to hang out on ye olde human home planet. Maybe the AI detected an audience and all that jazz, and wanted to make things interesting.
What actually worries me is, if you up that wave size, then the waves at difficulty 10 get nigh on impossible to survive at the start, possibly. You'd almost need some kind of grace period before the AI starts attack waves in order to live. Although, I guess, cross that bridge when you get to it... or if it's even an issue, given that I don't quite know how 10/10 is supposed to most efficiently be tackled.

[Gallant Dragon]

Maybe at 2.0 it was just a whole bunch of little things?  Like different cap sizes and balance, fewer options, turrets, AI defender behavior, all put together?

[Wingflier]

Dammit, stop making such good counterarguments.   

They're really not even that good.

He keeps bringing up 10/10 as a measure of balance, when 10/10 is literally supposed to be impossible to beat.

If I can beat that difficulty with any strategy it would be a miracle, because you're not supposed to be able to beat it.

He's using extreme examples to try to prove his point.

If the game only becomes a strategic, non-build-order spam on 9+ difficulties, then the design still sucks, plain and simple.

You keep assuming that the game has no strategy, despite evidence otherwise. Just because strategy is in choices in planets and techs instead of tactical choices doesn't make it any less strategy. I'm REALLY curious what version of AI wars didn't encouraging blobbing, since it certainly wasn't before 3.5ish era when I joined up.

People think the design of the game has improved since 2.0, but in actuality most of what has improved is the interface, the controls, and graphics, etc.

I mean for God's sake, the horrible galaxy pathfinding has been around since the game was released 3 years ago, and thousands of patches later, it still hasn't even been addressed.  If the damn pathfinding was a minor issue for all these years, consider what had to have been worse than that to come first.

It was a one-man project that tried to cover a huge scope and scale, and in many ways it succeeded, but at the cost of a very unpolished and ugly game, with several bugs and issues.  Part of the problem is that we continue adding bonus ships, continue adding scenarios, AI Plots, minor factions, AI personalities, etc., and still ignore the major underlying issues with the game.

[Gallant Dragon]

Ehm...  perhaps.  Although there is the point that you keep playing the same way over and over could be taken in different ways.  On one hand one could argue that its your own fault that the game is boring.  Then again, though, there is the very compelling (and probably more valid, come to think of it)argument that if you can beat a supposedly unpredictable and complex AI with the same strategy over and over it's the developers' duty to reexamine the game and figure out what's going wrong.

[chemical_art]

Dammit, stop making such good counterarguments.   

They're really not even that good.

He keeps bringing up 10/10 as a measure of balance, when 10/10 is literally supposed to be impossible to beat.

Your arguements have been soundly more repulsed then mine, but if you want to comfort yourself that playing the game identiically then getting bored when the game reacts in a similar fashion then go ahead. We can pound each other with insults all day, but we both know that will go nowhere.You are the one saying you can win any game with your formulaic strategy and are not noticing a difference a strategy. I prove you cannot, then say I take it to the extreme? We have played this game for years and know all the tricks, we are far, far beyond what should be considered balance point if you want to bring it up. This is evidenced by the plenty of new players who find 8 and below challenging. Try your formulaic strategy with 9.3/9.3 then. As difficulty level raises the number of viable strategies falls.

If I can beat that difficulty with any strategy it would be a miracle, because you're not supposed to be able to beat it.

Then prove it. The game has never claimed 10/10 it to be unbeatable, its always claimed it requires a  mix of cheese, intricate knowledge, and patience. A very, very tiny fraction of players have won without complete cheese.

He's using extreme examples to try to prove his point.

No less extreme then claiming you can win with the same strategy while claiming to not notice difference. I'm simply challenging your assertion that you can use your formulic approach to win any game.

If the game only becomes a strategic, non-build-order spam on 9+ difficulties, then the design still sucks, plain and simple.

Is the most effective strategy the same for every game? You can use the same strategy every time, without fail? Or are you mad that you have played the game enough you know how to counter the varying AI's with varying tactics, and now want a random game?

You keep assuming that the game has no strategy, despite evidence otherwise. Just because strategy is in choices in planets and techs instead of tactical choices doesn't make it any less strategy. I'm REALLY curious what version of AI wars didn't encouraging blobbing, since it certainly wasn't before 3.5ish era when I joined up.

People think the design of the game has improved since 2.0, but in actuality most of what has improved is the interface, the controls, and graphics, etc.

Get real, you think that. I haven't seen 2.0 even mentioned before this thread for anything more then nostalgia factor, and now you claim its this wonderful game for some reason.

I mean for God's sake, the horrible galaxy pathfinding has been around since the game was released 3 years ago, and thousands of patches later, it still hasn't even been addressed.  If the damn pathfinding was a minor issue for all these years, consider what had to have been worse than that to come first.

Again, YOU think that. I haven't seen the pathfinder issue even brought up before this week. I don't know what gets your knickers in a twist that your aggrevation of the week is the downfall of the game, but most of us get along fine with pressing "shift + click" if we want more specific instructions. There has been no great clamor over this issue. Now that you have brought it up it is being discussed, but you and only you find it more then a minor issue.

It was a one-man project that tried to cover a huge scope and scale, and in many ways it succeeded, but at the cost of a very unpolished and ugly game, with several bugs and issues.  Part of the problem is that we continue adding bonus ships, continue adding scenarios, AI Plots, minor factions, AI personalities, etc., and still ignore the major underlying issues with the game.

I don't see anyone but you lamenting that the game is getting worst over time. If you don't like it that is fine, but there does come a point where complaining doesn't add anything.

[Wingflier]

Your arguements have been soundly more repulsed then mine, but if you want to comfort yourself that playing the game identiically then getting bored when the game reacts in a similar fashion then go ahead. We can pound each other with insults all day, but we both know that will go nowhere.You are the one saying you can win any game with your formulaic strategy and are not noticing a difference a strategy. I prove you cannot, then say I take it to the extreme?

What part of 10/10 is supposed to be impossible to beat do you not understand?

If I went and played a PvP strategy game where the teams were 5v1 (me), then it wouldn't matter how good my strategy was, I would still most likely lose.  Just because I don't attempt to beat a difficulty that is continually buffed to be impossible to beat, does not mean the game's design is good, it just means the odds were stacked against me so heavily that without some really meticulous strategy like staying on my home planet, sending my champion around the entire map, taking out all the progress reducers for 3 hours (someone actually did this), and eventually winning by default.  That is not fun.  Keith has specifically said it's not the way the game is intended to be played, and Chris has spent untold amounts of time attempting to remove the effectiveness of deep-raiding from the game.

We have played this game for years and know all the tricks, we are far, far beyond what should be considered balance point if you want to bring it up. This is evidenced by the plenty of new players who find 8 and below challenging. Try your formulaic strategy with 9.3/9.3 then. As difficulty level raises the number of viable strategies falls.

That's exactly what I'm going to do, and I'll report back when I'm finished.

No less extreme then claiming you can win with the same strategy while claiming to not notice difference. I'm simply challenging your assertion that you can use your formulic approach to win any game.

Dude, if you can use the same strategy on difficulty SEVEN (which is supposed to be normal) against all the different AIs, there is still a problem.  Why wouldn't you balance the game around the average difficulty people use?  Why would someone have to jump to 9+ in order to start having to try new things?

I don't understand.

[chemical_art]

What part of 10/10 is supposed to be impossible to beat do you not understand?

The part where I see, in writing "10/10 is supposed to be impossible, and not just very difficult"

If I went and played a PvP strategy game where the teams were 5v1 (me), then it wouldn't matter how good my strategy was, I would still most likely lose. 

You are still bringing up PvP situations into a PvE game? When will you understand you are comparing apples to oranges?

Just because I don't attempt to beat a difficulty that is continually buffed to be impossible to beat, does not mean the game's design is good, it just means the odds were stacked against me so heavily that without some really meticulous strategy like staying on my home planet, sending my champion around the entire map, taking out all the progress reducers for 3 hours (someone actually did this), and eventually winning by default.  That is not fun.  Keith has specifically said it's not the way the game is intended to be played, and Chris has spent untold amounts of time attempting to remove the effectiveness of deep-raiding from the game.

This is called cheese. AI wars has plenty of it. No more cheesy then pulling up a 10/10 game with 10 worlds and the player taking 8 worlds, or the many many other cheesy things you can do. AI wars has so many options so you can choose what difficulty you want to use aside from the numbered difficulty. Why else would the exo wave difficulties have a multiplyer? Or why you can set mad bomber to send no waves at all? You have a toolbox of options to set up your games, and you are mad that there are combinitions to make the game easy? 

That's exactly what I'm going to do, and I'll report back when I'm finished.

That's great. But remember, if you play a different AI with that same difficulty and use anything but the exact same research tree, then you are making varying strategic decesions that dramatically effect the game.

Dude, if you can use the same strategy on difficulty SEVEN (which is supposed to be normal) against all the different AIs, there is still a problem.  Why wouldn't you balance the game around the average difficulty people use?  Why would someone have to jump to 9+ in order to start having to try new things?

I don't understand.

Because most people, especially new ones, don't ever play the game enough to fully understand the very complex game. We here are the forms have among the most intimate knowledge of the game. This is evidenced by plenty of new players coming in saying they have great games at under difficulty 8. These players don't cheese the AI for the most part, because they don't know how.  Until you take away the options that allow cheese, like 1/10 fallen spire games or mad bomber + no waves, then there will always be cheese.

[Wingflier]

I'll report back after my 9/9 Crosshatch game against a Heroic AI and a Special Guard Posts Captain.  Ta ta!

[Histidine]

Histidine's crazy idea time!

if you have a little bit of everything

You always have a little bit of everything, because the way ship caps work enforces this. This nominally enforces variety, but as noted in the schizo wave discussions and the introduction of per-planet ship focus types, everything everywhere is functionally as homogeneous as one thing everywhere.

My proposal:
Fleet ship (and maybe starship) caps are shared across ships of the same mark, weighted by the ship type's "natural" cap. Say I start a game with normal caps (96 of each triangle ship), with Spire Blade Spawners (cap=5) as my bonus ship. I can build:

• 384 Fighters
• 384 Bombers
• 384 Missile Frigates
• 20 SBSes
• 96 Fighters, 96 Bombers, 96 Missile Frigates, 5 SBSes
• 144 Fighters, 48 Bombers, 10 SBSes

or any combination in between.

The AI would need to be able to respond to this, so if you decide to just blob bombers (because who wouldn't?), it starts focusing Fighters in reinforcements (where possible), waves, SF reinforcements, hybrid drones, exos, the like.

If coupled with TechSY730's idea for emphasizing planetary ship focus types more strongly (which I think is a good idea on its own regardless), we could turn the game into a bit of a strategic puzzle. This planet has a lot of 'ling Nanoswarms, that one's heavy on the Zombards, the one with the CSG seems to like Raptors with some Gravity Drains for flavor... how to best deal with each of them?

This idea won't prevent tactical blobbing - to be honest, I'm not sure we can fix it with our current toolset even if we wanted to - but it introduces strategic considerations in deciding the composition of your blob.

Obvious problems:

• If you have a bonus ship you really like , you don't care what new bonus ships you get from an ARS because you can just use it to feed your ship cap if you can't find a use for it otherwise.
  (probably at least partly a symptom of the following issue)
• Emphasizes ship imbalances - imagine going pure Tackle Drone Launcher for your home defense fleet.
  I think that for the most part, if the game encourages this it's because of an imbalance that should be fixed anyway. But as someone put forward in the discussion of AI low-cap ships (sorry, forgot who!), low-cap ships may well become non-linearly stronger when in numbers beyond the ship cap available to the player. Some ships will probably need non-sharing caps, or at least to share caps only within a limited subset of ship types.
• It'll be rather micro-intensive to maintain a certain composition through short-term losses. Say I have a Space Dock making 5 Fighters/5 Bombers/5 Frigates on repeat. My 96/96/96 blob loses 30 fighters, which are replaced by 10 Fighters, 10 Bombers and 10 Frigates. My fleet composition has changed without me meaning it to.
  A"maintain up to X units" control for each ship type would work, but that's probably overcomplicated for a UI...
• All that fancy strategic stuff could well degenerate into "take one planet, spend 20 minutes refitting blob, take next planet" (i.e. grind).
  (this is probably the biggest issue)

[TechSY730]

An intriguing proposal.

Make ship caps (at the very least for fleet ships) a relative ship "cost" against a per mark "supply".

This most certainly would more closely mirror the system the AI uses (again, without a hard cap in supply, but we would gain the opportunity to substitute some ship count for another if we chose, like the AI can)

I'm not sure if something this radical should happen, but it is most certainly a cool idea.

[Wingflier]

Histidine's crazy idea time!

if you have a little bit of everything

You always have a little bit of everything, because the way ship caps work enforces this. This nominally enforces variety, but as noted in the schizo wave discussions and the introduction of per-planet ship focus types, everything everywhere is functionally as homogeneous as one thing everywhere.

My proposal:
Fleet ship (and maybe starship) caps are shared across ships of the same mark, weighted by the ship type's "natural" cap. Say I start a game with normal caps (96 of each triangle ship), with Spire Blade Spawners (cap=5) as my bonus ship. I can build:

• 384 Fighters
• 384 Bombers
• 384 Missile Frigates
• 20 SBSes
• 96 Fighters, 96 Bombers, 96 Missile Frigates, 5 SBSes
• 144 Fighters, 48 Bombers, 10 SBSes

or any combination in between.

The AI would need to be able to respond to this, so if you decide to just blob bombers (because who wouldn't?), it starts focusing Fighters in reinforcements (where possible), waves, SF reinforcements, hybrid drones, exos, the like.

If coupled with TechSY730's idea for emphasizing planetary ship focus types more strongly (which I think is a good idea on its own regardless), we could turn the game into a bit of a strategic puzzle. This planet has a lot of 'ling Nanoswarms, that one's heavy on the Zombards, the one with the CSG seems to like Raptors with some Gravity Drains for flavor... how to best deal with each of them?

This idea won't prevent tactical blobbing - to be honest, I'm not sure we can fix it with our current toolset even if we wanted to - but it introduces strategic considerations in deciding the composition of your blob.

Obvious problems:

• If you have a bonus ship you really like , you don't care what new bonus ships you get from an ARS because you can just use it to feed your ship cap if you can't find a use for it otherwise.
  (probably at least partly a symptom of the following issue)
• Emphasizes ship imbalances - imagine going pure Tackle Drone Launcher for your home defense fleet.
  I think that for the most part, if the game encourages this it's because of an imbalance that should be fixed anyway. But as someone put forward in the discussion of AI low-cap ships (sorry, forgot who!), low-cap ships may well become non-linearly stronger when in numbers beyond the ship cap available to the player. Some ships will probably need non-sharing caps, or at least to share caps only within a limited subset of ship types.
• It'll be rather micro-intensive to maintain a certain composition through short-term losses. Say I have a Space Dock making 5 Fighters/5 Bombers/5 Frigates on repeat. My 96/96/96 blob loses 30 fighters, which are replaced by 10 Fighters, 10 Bombers and 10 Frigates. My fleet composition has changed without me meaning it to.
  A"maintain up to X units" control for each ship type would work, but that's probably overcomplicated for a UI...
• All that fancy strategic stuff could well degenerate into "take one planet, spend 20 minutes refitting blob, take next planet" (i.e. grind).
  (this is probably the biggest issue)

I think doing something like this, in addition to adding some kind of new AI Plot - Super Guardians:  http://www.arcengames.com/mantisbt/view.php?id=9450
that discouraged blobbing...by well, absolutely destroying everything that blobs, would make the tactical aspect of the game so much more exciting for people like me, who want a constant challenge, not just a minor challenge every few hours.

A HUGE part of the problem is the current state of the hull type and armor system.  People try to downplay how big of an issue this actually is.

The fact of the matter is that about 25% of the bonus ships are just flat-out better than the rest.  In any given situation, the best bonus ship (at least in my experience) is extremely easy to choose.  Instead of having say 3 bombers all with different strengths and weaknesses, or 3 raiders, 3 long-range bombarders, etc., you've got 1 ship that's easily the best and the other two that...well why bother?

The entire thing has become so homogenized.  I really wish Keith would just take my advice and put in a bonus ship omission file.  It would make the game 1 million times more balanced and fun for both the player and the AI (how many times do they get garbage ships?)

Then there would be no pressure getting to the armor and hull type rework.

[Kahuna]

He keeps bringing up 10/10 as a measure of balance, when 10/10 is literally supposed to be impossible to beat.

At the moment 10/10 is not impossible so it's ok to bring that up. It's not impossible so go ahead and blob your way to victory vs 10/10 before it's buffed again. Or at least vs 9,6/9,6.

[Kahuna]

I do agree that we need to look at the 2.0 days and see why the waves were more threatening and the AI seemed to deal with "blobbing" and deep strikes better.

Was it because the AI got turrets back then?
Was it because the player economy was so much weaker? (thus, the AI was competitively stronger to the player at most points in the game as a result)
Was it because the waves started out stronger in the early to mid game?
Was it because the humans had far less tools from the start back then?
Was it because a different balance model between ships, and different "balance targets" for each ship class?
Was it because of the different armor/shield system?
Was it something else?

I honestly don't know.

Someone needs to look at what it was back then and see if any of the concepts, balance goals, and limitations (on the player especially) will fix some of the issues the game has now, while trying to integrate them with the solutions to the issues the game had back then (especially when it came to AI ferocity, even if it is found that it was artificial due to player limitations)

Again, complicated stuff.

Did the AIs use fewer ship types? Now the AIs use everything they have unlocked. All AIs are basically "Everything" AIs. And as we know non Schizophrenic waves are harder.