Gameplay thoughts from a newer player

[lanstro]

Yes, the chess analogy is very good Doddler.

Anyway, we'll play a few games utilising the new things I've learnt in this thread and report back.  Each game does take a while though, so it may be some time before additional meaningful feedback!

[superking]

I often think the least interest part of the game is killing the last AI, when you have already got one and you have to slowly grind upto the other, which takes a while because the other AI is renforcing heavily now your AIP is high. I love the early AND midgame, but the endgame vs the last AI sucks.

I agree with much of what is said in this thread concerning the lack of 'the AI hitting you back'

Just my 2 cents, and there are plenty of people in this forum with 2 cents, but what I would love to see:

when you kill one of the AIs, the whole dynamic of the game changes; it switches from attack the AI to tower defence. The ai fires out CPAs constantly, and virtually empties its systems over time. This means, if you survive, that then finishing the AI is fairly swift, while ensuring the endgame is a huge showdown. This would be great while removing all the huge mkIV garrisons that are making the endgame last forever.

[WinterBorn]

On the next DLC poll I am nominating AI recolonizing planets - a group of threads that has been around about a year now.

http://arcengames.com/forums/index.php/topic,459.0.html
http://arcengames.com/forums/index.php/topic,2609.0.html

New thread
http://arcengames.com/forums/index.php/topic,897.msg29181.html#msg29181

I think this could be an very good new agressive behavior for the AI. As it is now if I lose a planet to the AI it does nothing with it. Recolonizing AI's would mean that I could lose a planet and the AI would rebuild turrets, minefields, etc. Forward staging bases could get wiped out and become very hard to retake. I would also expect the AI to rebuild stronger since it had to fight to retake the system.

------------------------------------------------------------------------------------
I also like this Idea from Keith from earlier in this thread

Ah, I meant percent of a given reinforcement, rather than all ships on all planets (that could kill your framerate even worse than it would kill your planets).  Ever watch an AI planet for a while and see a bunch of ships and turrets and whatnot pop out of nowhere?  That was a reinforcement; I'm saying that a certain percent of the ships added in one of those could be immediately freed to go after you, rather than clinging to a guard post.

And perhaps it could be set up so that if a planet is at/close-to it's "cap" that instead of redirecting those reinforcement checks to other planets (and thus building up tons of crust on them too), it would just release all of those (or more realistically just a higher percent of those, and reinforce fewer other planets).  This would obviously be a niche-option (and perhaps later an AI type), as balancing will be tricky and the no-warning nature of it would be off-putting to much of the player base.

[Doddler]

I also like this Idea from Keith from earlier in this thread

Ah, I meant percent of a given reinforcement, rather than all ships on all planets (that could kill your framerate even worse than it would kill your planets).  Ever watch an AI planet for a while and see a bunch of ships and turrets and whatnot pop out of nowhere?  That was a reinforcement; I'm saying that a certain percent of the ships added in one of those could be immediately freed to go after you, rather than clinging to a guard post.

And perhaps it could be set up so that if a planet is at/close-to it's "cap" that instead of redirecting those reinforcement checks to other planets (and thus building up tons of crust on them too), it would just release all of those (or more realistically just a higher percent of those, and reinforce fewer other planets).  This would obviously be a niche-option (and perhaps later an AI type), as balancing will be tricky and the no-warning nature of it would be off-putting to much of the player base.

Can we nominate this idea? :p

I've been experimenting a bit with super-agressive AI settings, trying to design a game setting that would last no more than 6-8 hours.  My last run was something like this:

- 40 planet grid style map.
- +200% handicap set for the player.
- Two Lv 8 AI's, (I set it random, I got Alarmist and Raider).
- AI Progress set to 1 per minute.
- All minor factions minus dyson.
- Game options Cross Planet Waves, Unannounced Waves, Schizo Waves, 2x Wave Size.
- Of course fast and dangerous combat.

It actually works out quite well.  While you can build ships really fast even right off the start with the handicap, the AI hammers you really hard every couple minutes with large attacks.  By around 200-300 AI progress the AI starts to overlap waves, where they are attacking before the previous wave dies.  I eventually fell over at about 2000 threat at 3h, but I'm confident that it can be beaten.

1 AI progress per minute sounds like a lot, but it turns out that AI progress reduction is affected by handicap, so you gain -60 progress for breaking a data center (1h of auto progress).  Given the size of the map, you should be able to hit centers like that without huge issue.

Handicaps work kind of weird though.  I didn't expect it to increase the rate of researching.

[Mánagarmr]

It's not actually wave size that is doubled, it's rather the time between waves that is halved. So you get twice as many waves within the same timespan, although with unannounced and CPA wave types, they almost end up being double size anyway

[lanstro]

Hi all

Sorry to revive an old thread, but I'm back with some more feedback after a few games with the expansion.

We've been playing 3-4 player games on difficulties 7.0 to 8.6, with 2x waves, schizephrenic waves, no wave warnings, CPA waves and 1 aip/5 mins on realistic 80 planet maps, against 2 random all non technologist AIs.

Firstly, the gameplay in the expansion is massively better than in the original, especially with the above options on.  It feels a fair bit more dynamic, though of course the effect is simulated through the addition of neutral factions and buildings, rather than any improvement in the AI's behaviour.  That is, once you've figured out which of the random AIs you've rolled, you know how it's going to behave.  But your plans do change more often than it did in the original because of the larger range of neutrals.  It's not ideal but still very much better than the original.

We especially like the new ways to control the AI progress: the co-processors, the building that drops aip but raises the aip floor, etc: this adds an additional dimension to the game and we think that it should be further expanded.  I guess ideally we would like to see the AI being able to do things that raise its AI progress unless you go and counter it.

One thing that AI Wars lacks which makes it feel less of a strategy game is that the AI does not react to your unit combinations.  What I mean is that the AI really should pop out waves that counter the ships you have, and/or send its troops to the player who is least able to counter its ship mix, forcing the players to cover each other or create better balanced fleets.  A player who's gone heavy frigates should be subject to bombers attacks.  A wave that has a lot of frigates shouldn't rush blindly into the defences of the player with an abundance of bombers, it should try to go an easier way, etc.  Yes I've seen some feints and retreats, but they are rare and rarely effective.  As it is, the AI just seems to spawn some random collection of ships, which goes straight for the nearest planet, and most fights are 'my blob vs your blob' or 'my blob chasing your blob'.

Another thing in AI Wars that is more restrictive than in other strategy games is the ability of the player to affect the AI's means of waging war.  Sure, you can gate raid/guard post raid/control the AI progress, but it's not quite the same as the mechanic in other games of hitting the opponent's economy or their key troop production buildings to prevent them from being able to produce x y or z.  To simulate the effect, how about having certain buildings that you can destroy which lowers the AI's reinforcement rate on a global scale, and other buildings that you destroy which will hamper or completely destroy the AI's ability to build one class of bonus ships?  And how about instead of the AI automatically being granted a new bonus ship at certain times/ai progress thresholds, it spawns a number of these buildings, which if destroyed will then deny it that bonus ship.

Finally, the part of the game we most enjoy is CPAs, and we think they don't happen often enough.  How about a 2x and/or 4x frequency CPA option?  That'll make us pay more attention to defences, and also make taking planets less of a grind since the enemy guard posts will be freed up every few hours.

Regards
Lanstro

[keith.lamothe]

Thank you for reviving the thread, actually, I was hoping to get more feedback after you'd tried some of the options we'd implemented for people like you

1)

I guess ideally we would like to see the AI being able to do things that raise its AI progress unless you go and counter it.

Yea, that's definitely a preference that varies from player to player.  We've got the mining golem and the rebel colonies for this (you may not see that as the AI doing things, but it is functionally similar), and more minors are planned.  I have some ideas for more AI-initiated stuff in the plots section (which don't all need to be endgame, though probably won't ever be early game as we want to let people get on their feet first).

2)

One thing that AI Wars lacks which makes it feel less of a strategy game is that the AI does not react to your unit combinations.

Hmm, that could get into some gap-in-the-wall issues (which is one of the big, huge, nasty problems that AI War's design prioritizes the prevention of), but I do have some ideas for letting the AI send some extra attacks that are somewhat more focused on getting something done in player territory, and if they get gap-in-the-wall'd, well, you've still got to deal with everything else the AI is doing.  All that would need to pass the Chris-test, though

3)

Another thing in AI Wars that is more restrictive than in other strategy games is the ability of the player to affect the AI's means of waging war.

Personally I would absolutely love to be able to smash up the AI economic and industrial infrastructure (and to see a separate economic, industrial, and research component to AI Progress, but that's just not going to fit the game)... but this is getting into one of the other big, huge, nasty problems that AI War's core design is dead-set-against: getting to the midpoint of the game and you've clearly won, it's all just easy mop-up.  Granted, some players still get to a mop-up point by mid-game, but they have to be careful even during that phase or they can get torn to bits.  So any exposure of the AI's infrastructure would have to either be really superficial or the destruction of it would have to shift the challenge to another aspect rather than just remove it.

4) For CPA frequency, you are not the first to request modifiers for that, and I hope to add them after the official release coming out shortly.  I'm thinking 0x, 0.5x, 2x, 4x, 8x (for the insane). 

By the way, have you noticed much in the way of ships being freed from the new border aggression mechanic?  It's quite good as-is, but I'm thinking some players would like modifiers and/or AI Types to really crank that up; that way you're constantly having to fight for your life but taking AI worlds is not nearly as siege-oriented (though that could be textured by making various key AI worlds not contribute as-much or at-all to border aggression releases).

[superking]

One thing that AI Wars lacks which makes it feel less of a strategy game is that the AI does not react to your unit combinations.

Hmm, that could get into some gap-in-the-wall issues (which is one of the big, huge, nasty problems that AI War's design prioritizes the prevention of), but I do have some ideas for letting the AI send some extra attacks that are somewhat more focused on getting something done in player territory, and if they get gap-in-the-wall'd, well, you've still got to deal with everything else the AI is doing.  All that would need to pass the Chris-test, though

Well, the wiki suggests that the AI often is able to do the optimal behavior, but a less than 50% chance of doing so in order to ensure unpredictable behavior

Why not give the AI a similar chance of using the optimal unit in waves?

By the way, have you noticed much in the way of ships being freed from the new border aggression mechanic?  It's quite good as-is, but I'm thinking some players would like modifiers and/or AI Types to really crank that up; that way you're constantly having to fight for your life but taking AI worlds is not nearly as siege-oriented (though that could be textured by making various key AI worlds not contribute as-much or at-all to border aggression releases).

why not change the border aggression mechanic to be affected by AI type as with wave sizes?
I can see the Mad Bomber using more border aggression than the Fortress Baron, for example

that would have the advantage of also making a planets ship count before border agression even less predictable, especially with random AI types

[keith.lamothe]

Why not give the AI a similar chance of using the optimal unit in waves?

The problem is that there is no way of computing the optimal unit that could not be predicted and exploited by the human player.  Thus, a gap-in-the-wall.

[superking]

Why not give the AI a similar chance of using the optimal unit in waves?

The problem is that there is no way of computing the optimal unit that could not be predicted and exploited by the human player.  Thus, a gap-in-the-wall.

I disagree! 

For example:

The AI is programmed to choose the optimal unit by splitting all units, turrets and buildings into 3 catagories: weak to fighter, weak to bomber and weak to frigate. It decides the optimal unit to choose.

However, there is only a eg. 25% chance that it uses the optimal unit; 75% of the time it chooses randomly.

I do not beleive players would bother carefully micromanaging their fleet composition between their warpgate worldand a niegbouring for a 25% chance of gaining an advantage and a 75% chance of micromanaging pointlessly. The infrequency of the AI making the gap-in-the-wall decision makes it too unpredictable (and random AI waves are invariably of stupid compositions and unit types anyway!)

This creates a situation where the player does not greatly change his playstyle, but the AI has a 25% of adapting to player disposition and launching a far more effective wave than normal - making the AI a more intelligent opponent without comprimising its unpredictability.

[Doddler]

You'd THINK that schizo AI modifier would make waves more random, but in fact it does the opposite.  Once the waves get to a certain size, more or less every wave is the same even composition of all the AI's available units.  I'm not sure this is specific to CPA waves, but because the AI won't group move their ships or wait for all ships to gather before sending them in, a wave of 600 ships tends to break down into small groups of 100-200 ships, which makes it easier.

What I'd like to see in future expansions, more than anything, would be more options to customize galaxy creation and AI behavior.  It's pretty common in 4x games to optionally allow players to configure most major aspects of the game.  AI war has such a robust system that allows for a huge variety of gameplay, but offers no way to make any of those changes.

[keith.lamothe]

Well, iirc, one of the currently-planned "anchor" features for the next expansion is a scenario editor, which would probably include much of what you want there.

[lanstro]

I agree with Superking that it doesn't have to be a hole in the wall issue.  And while we're on the point, I have to say gate raiding is the most glaringly hole in the wall feature I've seen in strategy games for some time.

As for smashing up AI economic/industrial infrastructure leading to an endgame rout - again, strongly disagree that this will be the effect.  In my view it'll add more interesting strategic options: do you go straight for the core planet, which is going to be a pain to break as-is, or do you take your time, to go on a few tangents to erode the AI's economy and ability to produce certain ships, at the cost of more aip?  At the moment there is only one correct way to approach the final core planets: the shortest possible route.  Why?  Because it minimises the grind.  Add in the ability to soften the AI up, dot that infrastructure around the map, in particular further away from the humans' starting planets, and you add many options. 

In any case, even if the game becomes a bit easier, everyone can just move up a difficulty level or two.  Or you could compensate by making the AIs a little stronger.  As far as I know, nobody is regularly beating standard 80 planet settings games on 9+ difficulty right?

As for the new border-aggression mechanic, I'm afraid we haven't noticed it much.  We move fairly quickly and rarely do frontier towns get that many ships.  Further, since we play CPA waves, there really isn't much difference between a wave and a released border aggression 'wave'.  Suggest an option to drastically turn down the threshold for the border-aggression mechanic to turn on (to put some real pressure on border towns next to Mark IVs).

Doddler, I have indeed noticed that schizo waves tend to separate out and become easier to kill than a 'blob'.  Maybe we should return to non schizo waves. 

[superking]

As for the new border-aggression mechanic, I'm afraid we haven't noticed it much.  We move fairly quickly and rarely do frontier towns get that many ships.  Further, since we play CPA waves, there really isn't much difference between a wave and a released border aggression 'wave'.  Suggest an option to drastically turn down the threshold for the border-aggression mechanic to turn on (to put some real pressure on border towns next to Mark IVs).

or alternatively, http://arcengames.com/forums/index.php/topic,5048.msg35318.html#msg35318

/plug 

[keith.lamothe]

And while we're on the point, I have to say gate raiding is the most glaringly hole in the wall feature I've seen in strategy games for some time.

Yes, that one is actually intentional   It's negated by CPA waves, though.

On the other hand, as you've now found, CPA Waves + Schizo Waves is actually easier than having neither of those options (or neither), since it disrupts the AI's concentration of force.  That's one area where we might be able to make a change, by having a quasi-schizo modifier that tries to pick like-speed-ship-types from its available set  and group move them until it reaches human territory before un-furling.  There would be some exploitability problems there with ambushing and lightning warheads, but probably nothing worse than CPA waves in general.  And even if exploitable, it'd just be a modifier for those desiring it, and could be fixed (or possibly just scrapped) as needed.  Of course, I'm not sure how Chris feels about the idea

On the infrastructure stuff, one of our planned alternate endgame scenarios ("plots") is to have special AI structures scattered throughout the galaxy by the destruction of which you can cause the AI to shift towards a more aggressive stance, which could be accomplished by applying a modifier to the border aggression cap.  Basically a galaxy-wide version of smashing up guard posts and command stations to get the enemies to come after you.  The balance will be a little tricky, probably buffing the AI core world defenses to make some degree of it necessary, and having the individual structure-destructions result in a fairly significant AIP boost, etc, so that it doesn't just make the game easier... anyway, you might find it more to your liking.

For a similar plot, or perhaps instead of (still working it all out, maybe they could work well together), I'm hoping to implement a sort of "Showdown" plot where once you're on a more-or-less even footing with the AI (either when you destroy the first AI home command station, or something like that), it basically makes an all-out attempt to kill you, and if you survive you can probably win in an hour or less.  The idea is, for those who want this, to condense the challenge of that second half of the game into one really intense series of battles.  The design challenge there is manifold, in that simply freeing all AI ships to attack will basically kill your framerate, and even aside from that there are many players who set up such effective defenses that this would be suicidal behavior on part of the AI.  So either it would need ways of bypassing normal defenses (spawning new exo-galaxy wormholes, special siege-weapon type units to smash stuff, etc), which I personally would find frustrating (not the end of the world), or we'll need to find some other way of making sure it's still fun.

Back to the infrastructure thing, I think it would be possible to have a sort of minor-faction-type-event where instead of "Mining Golem Going to Pwn Murdoch in 1:34:22" it's basically "AI 2 Finishes Construction Of New Fabrication Complex in 0:25:00", and when they finish it they get a new bonus ship type, and when you destroy it it goes away.  This wouldn't change the basic AI behavior, it would really be just as much a surprise to it as you from a code perspective, but it would provide a new "mission type" for the humans to go out on, and give you a way/motivation to strike at the AI infrastructure without reducing the difficulty (to make it a little less one-sided, the destruction of the complex could reduce AIP a little, or maybe the humans could use it like a core fabricator, etc).

Anyway, I think that's enough from me right now