Core design: What makes AI war, AI war?

[Kahuna]

Pretty much everything I can think of has already been mentioned.

10.  Asymmetry.  I really wish more games of this sort did this.  This makes things extra fascinating to me, and it means that it wont feel like a "mirror match" also.  I'm not fighting against JUST the units that I myself can also use.  Sure, I'll definitely see some of them on the enemy's side... but at the same time the AI has it's own unique threats that I cannot use, and the same the other way around.

Imo this (in addition to the amount of content and ways to approach the game) is one the most important aspects of AI War. The problem with PvE mirror matching is that the AI will never be as good as the player unless the AI is given "unfair" advantages like extra resources, reduced build times etc. which easily make it feel like the AI is "cheating". Thus it feels like the game is no longer hard but "cheap hard".

There are a lot of things already mentioned here but I'll add this one. 

Sense of scale.

You can have superbly massive battles raging across any one of a hundred plus systems, or you can take it down and relentlessly raid with handfuls of ships, and it can happen at the same time, in the same game. 

Also, this.  these moments.  Those things that tell you you done messed up.  Stuff like this.

Or you can build a massive beachhead and nuke hundreds of thousands of ships at once. I'd have waited a bit longer and nuked even more ships at once but the game started to freeze after the ~400k ships. Doing crazy things and succeeding against all odds is just something really awesome and rewarding when you manage to pull it off. I don't think you can do stuff like this in any other game so that too makes AI War unique.

Sometimes, I will buy the stuff that seems the most fun

Yet another point for AI War. You can do/choose all the "fun things" and still be successful/effective. There are so many viable strategies and options that everyone finds something they like. In other games I've run numerous times into the problem of liking something that is vastly underpowered or unviable. For example enjoying playing a class in an RPG that has interesting mechanics and playstyle but ultimately finding out that it's underpowered/unviable and that I'd be much better of just choosing another class or build. In AI War I can take pretty much everything the game has to offer and build my own play style/strategy from that.

When I read that name, I assume anything is possible 

Funny, I think the same thing when I see a particular super-cat avatar...

I'm flattered.

[Cinth]

There are a lot of things already mentioned here but I'll add this one. 

Sense of scale.

You can have superbly massive battles raging across any one of a hundred plus systems, or you can take it down and relentlessly raid with handfuls of ships, and it can happen at the same time, in the same game. 

Also, this.  these moments.  Those things that tell you you done messed up.  Stuff like this.

Or you can build a massive beachhead and nuke hundreds of thousands of ships at once. I'd have waited a bit longer and nuked even more ships at once but the game started to freeze after the ~400k ships. Doing crazy things and succeeding against all odds is just something really awesome and rewarding when you manage to pull it off. I don't think you can do stuff like this in any other game so that too makes AI War unique.

 

[kasnavada]

One of the beautiful thing in AI war is that even when automatically moving around (looking for targets), via X movement, v movement, automatic patrol, every ship chooses its target smartly, calculate kills on the fly and so on.

[Draco18s]

More hacker uses is always an interesting choice.

Speaking of....

Hacking as a resource has been one of the best things to happen to AI War.  The metal/crystal thing made less and less sense as time went on (it was virtually impossible to create a fleet that required 10 times more metal than crystal, or vice versa, leading to very balanced costs, leading to very balanced income, leading to virtually no distinction between the two).  If resources could run out then it might've made sense to keep the distinction, but again, AI War was like "there's resources here, they're infinite, don't sweat it" and that was good.

[Cyborg]

More hacker uses is always an interesting choice.

Speaking of....

Hacking as a resource has been one of the best things to happen to AI War.  The metal/crystal thing made less and less sense as time went on (it was virtually impossible to create a fleet that required 10 times more metal than crystal, or vice versa, leading to very balanced costs, leading to very balanced income, leading to virtually no distinction between the two).  If resources could run out then it might've made sense to keep the distinction, but again, AI War was like "there's resources here, they're infinite, don't sweat it" and that was good.

There are limited resources in the game, but not crystal, and that was important. I think we got to a good place with the resources after a few years. I don't see any reason to completely scrap that, although I'm certainly listening.

[Billick]

I don't know if this is a case of people ignoring the elephant in the room, but the core of AI War design, to me, is the AI progress mechanic.  It's kind of the keystone of the whole design.  It's so simple and elegant. 

[eRe4s3r]

That is very true.. the AIP mechanic is the core of AI War, if anything it needs even more reactive AI stuff tied to it because even at AIP 600 I don't really feel the dread (I know how to build a decent defense..)

[Draco18s]

==Note: this post may need its own thread==

I suppose I can say the one thing I would like to see changed, somehow, is for games to not take bloody forever.  Three to six hours would be ideal.  And I realized that epicly long games is one of the things AI War is known for, but I don't think it's really needed to make AI War AI War.  Should they be possible?  Sure, there are all kinds of playstyles and some of them favor longer games.  But it's not about the time component, but the nature of progression, the ability to make further gains.

One of the things the game doesn't do well is forcing an end-game.  You can grind to a long, drawn out stalemate easily where it's virtually impossible for the player to progress, but the AI doesn't do anything to actually crush the player.  That is, the difficulty doesn't continue to ramp up when a stalemate occurs (as the player is no longer causing AIP increases, so they languish at defending against a statically sized wave after wave of attacks that cannot breach their defenses).

I don't know what can be done about it.

If we look at the Steam Achievements, we can set a baseline for "how many people have played for more than ten minutes" with the #1 most achieved achievement:
1. Rainy Day Savings: 50,000 Energy

Given that the game is currently "do you have energy? good, you get this achievement" as the 50k value used to be a lot and now it's half what you get for building a single reactor.  Anyway, 28.2% of the people who own the game have this one.  Generally the most common achievement gives you an idea of how many people are actually playing as it tends to be one you get on accident (makes me want to put in a "you ran the game" achievement to a game, just to see the numbers).

But let's look at the "win/loss" numbers.

"1. First Loss" 12.4%

Half!  Half of the people who have played at all have ever lost a game!

"1. First Victory" 3.6%

And only 1 in 8 players has ever won a game.

Even if we add these together, that still leaves 3/8ths of the people who've played the game having not played it long enough to reach a conclusion.  Instead they abandon the game and start a new one (or never come back).  And that drawn out stalemate is likely the problem.  No one is going to let the AI whittle away at their defenses just to have the game tell them that they've lost: they already know it.  So they quit and maybe try again.

A lot of board games don't have this problem, where the end-state occurs right at the moment that its becoming apparent who's winning.  Race for the Galaxy does it perfectly, managing to thread the needle between "arg, one turn!" and a runaway winner.  That is, if there had been one more turn, only one person would have been playing: the guy everyone already recognizes as the winner, but if there'd been one fewer turns, it could be a tossup between Player A and Player B depending on what they did that last turn vs. what they would have done on a one-more.

Other RTS games usually have a tipping point as well, where one player achieves that upper hand that leads to an ever-widening gap in power, leading to a quick loss.  The middle game there isn't a stalemate where neither player can progress, but rather a series of probing strikes, looking for a weakness to exploit, that when found tips the scales suddenly and decisively.

AI War though, I tend to find that if the AI throws a giant wave of ships at me, I can fend it off.  But it leaves my own armada weak and unable to push back and in the time it takes me to rebuild, the AI has rebuilt too and we clash head on and neither of us accomplish anything.  Some players here have found ways around this (cough, nukes, cough) but I tend to be AIP-averse and don't resort to warheads as my policy is "a nuke now means two nukes later" as the AIP rises.

[chemical_art]

==Note: this post may need its own thread==

I suppose I can say the one thing I would like to see changed, somehow, is for games to not take bloody forever.  Three to six hours would be ideal.  And I realized that epicly long games is one of the things AI War is known for, but I don't think it's really needed to make AI War AI War.  Should they be possible?  Sure, there are all kinds of playstyles and some of them favor longer games.  But it's not about the time component, but the nature of progression, the ability to make further gains.

One of the things the game doesn't do well is forcing an end-game.  You can grind to a long, drawn out stalemate easily where it's virtually impossible for the player to progress, but the AI doesn't do anything to actually crush the player.  That is, the difficulty doesn't continue to ramp up when a stalemate occurs (as the player is no longer causing AIP increases, so they languish at defending against a statically sized wave after wave of attacks that cannot breach their defenses).

I don't know what can be done about it.

If we look at the Steam Achievements, we can set a baseline for "how many people have played for more than ten minutes" with the #1 most achieved achievement:
1. Rainy Day Savings: 50,000 Energy

Given that the game is currently "do you have energy? good, you get this achievement" as the 50k value used to be a lot and now it's half what you get for building a single reactor.  Anyway, 28.2% of the people who own the game have this one.  Generally the most common achievement gives you an idea of how many people are actually playing as it tends to be one you get on accident (makes me want to put in a "you ran the game" achievement to a game, just to see the numbers).

But let's look at the "win/loss" numbers.

"1. First Loss" 12.4%

Half!  Half of the people who have played at all have ever lost a game!

"1. First Victory" 3.6%

And only 1 in 8 players has ever won a game.

Even if we add these together, that still leaves 3/8ths of the people who've played the game having not played it long enough to reach a conclusion.  Instead they abandon the game and start a new one (or never come back).  And that drawn out stalemate is likely the problem.  No one is going to let the AI whittle away at their defenses just to have the game tell them that they've lost: they already know it.  So they quit and maybe try again.

A lot of board games don't have this problem, where the end-state occurs right at the moment that its becoming apparent who's winning.  Race for the Galaxy does it perfectly, managing to thread the needle between "arg, one turn!" and a runaway winner.  That is, if there had been one more turn, only one person would have been playing: the guy everyone already recognizes as the winner, but if there'd been one fewer turns, it could be a tossup between Player A and Player B depending on what they did that last turn vs. what they would have done on a one-more.

Other RTS games usually have a tipping point as well, where one player achieves that upper hand that leads to an ever-widening gap in power, leading to a quick loss.  The middle game there isn't a stalemate where neither player can progress, but rather a series of probing strikes, looking for a weakness to exploit, that when found tips the scales suddenly and decisively.

AI War though, I tend to find that if the AI throws a giant wave of ships at me, I can fend it off.  But it leaves my own armada weak and unable to push back and in the time it takes me to rebuild, the AI has rebuilt too and we clash head on and neither of us accomplish anything.  Some players here have found ways around this (cough, nukes, cough) but I tend to be AIP-averse and don't resort to warheads as my policy is "a nuke now means two nukes later" as the AIP rises.

Will echo this. There have been improvements in this regards in terms of salvage, but I feel like a more intense factor unique to late game needs to occur. There is a definite "listless" factor to the mid game, where there is simply to little direction on what to do, so it feels like a list of "do this, then this, except for this." After this mid game list has been done, you are stuck with "either I can kill both AI HW, or I cannot and get bored"

So as the game reaches its close, both the AI and the Human need to be more exposed. "Plays" [sidebar, I NEVER expected to say that about this game] need to occur. Either the player or the AI makes a meaningful, purposeful gain that forces the game to end.

The game has to FORCE games to end. Not die in the stalemate. That also means if the player is using there is a very real hope that one successful defense can turn the tide.

[Cinth]

I find stalemates and acceptable endgame.  And I like that I control the pace.  I like being able to invest over 100 hours (game time) into one match.

Though I will say that usually when I stalemate, it's because I made a mistake somewhere along the way. 

[Steelpoint]

I'm pretty sure why the energy achievements are more unlocked is because years ago you got points or something for unlocking the achievement with one of the earlier Steam Summer/Winter Sale events.

[chemical_art]

I find stalemates and acceptable endgame.  And I like that I control the pace.  I like being able to invest over 100 hours (game time) into one match.

Though I will say that usually when I stalemate, it's because I made a mistake somewhere along the way.

I highly respect you Cinth but I have to ask:

Just how many players invested 100 hours in their_total_career of AI Wars?

5.5% of players did 10 hours in one game.
3.0% of players did 15 hours of a game.

For me, the norm should be closer to 5 hours per game given the new medium. Yes, I know that means the average number of worlds decreases. It is a shocking idea. But I am challenging basic facts.

WHY does a game need to be OK with 100 hours. I cannot think of any modern game except an hardcore GSG going for that long! Yes, by choice someone can have an 100 hour game. Just like some players can take 120 worlds in one game.

That does NOT mean the game should ever be designed where such a thing is the norm. For it is frankly not, in the current game environment.

[Cinth]

I think it's a great part of the scale of this game.  I can play my way, and you yours and it's all possible. 

And let's remember that Steam isn't the only means of acquiring the game.  I got mine off D2D back in the day :p   It might provide a metric, but it isn't the only metric.

[chemical_art]

I think it's a great part of the scale of this game.  I can play my way, and you yours and it's all possible. 

And let's remember that Steam isn't the only means of acquiring the game.  I got mine off D2D back in the day :p   It might provide a metric, but it isn't the only metric.

True enough, but steam is the most "mainstream" method of acquiring the game. I doubt other sources provide sources of less skilled players with greater sales, so they would only skew the results about the average skill level of players.

AIW never proposed playing 120 planets for the sense of scale, but it was provided so that the few who had the means and will to enjoy it could. That does not have reason to change.

But for most players, the ideal of 70 worlds needs to be cut radically down. The overall time scale needs to drop dramatically. Players cannot drop 10's of hours into a campaign. It seems 10 hours should be the max for new players.

[kasnavada]

I suppose I can say the one thing I would like to see changed, somehow, is for games to not take bloody forever.  Three to six hours would be ideal.  And I realized that epicly long games is one of the things AI War is known for, but I don't think it's really needed to make AI War AI War.  Should they be possible?  Sure, there are all kinds of playstyles and some of them favor longer games.  But it's not about the time component, but the nature of progression, the ability to make further gains.

This.
Really, this.

It's one of the design flaws of SBR & bionic dues too. A game of bionic dues take forever and more, and SBR was designed from the start to start for thousands of turns...