Tuesday, May 25, 2010

Curatorial Vs Exploratory Game Design

Two big trends in game design these days are "cinematic" game design (ala Uncharted 2), and "sandbox" games (ala Grand Theft Auto). These are trendy things to talk about, and tons of people have really strong opinions about one or the other type of game for various reasons. However, I think that both cinematic and sandbox games are the latest incarnations of a much broader division in game design that goes back to the roots of the industry or before.

As the title of this post suggests, I'm calling them Curatorial and Exploratory, and I think that all games can be lumped into being one or the other -- or, really, varying mixes of both. It's rare that anything would be one or the other only.

Curatorial Game Design
Curatorial game design has something to show you. It wants you to follow a certain set of rules, and it will take you through a progression. There may be some little deviations here and there from the main progression, but overall it's just a matter of how well you perform during the progression that determines the outcome. It's a linear test of some sort of skill, be that reflexes or mental acuity.

Examples of very curatorial games:

The Original Super Mario Bros, which has a linear progression of levels (warp zones aside, but even that is just branching linearity, there's no variance in them.

Half Life 2, which funnels the player through a heavily (and wonderfully) scripted world. Their trick is that they make the game seem more exploratory than it really is by having a few little nooks and crannies, but really it's a very linear progression nonetheless. It feels way less linear than other classics like Quake 2 because of its innovative level design.

Peggle, which primarily consists of a long string of levels that you move through completely linearly. There seems also to be some sort of Quick Play mode, but the menu organization makes it pretty clear that this is hugely secondary.

Pac Man (and most other classic Arcade games) is simply a long series of levels like Peggle.

Exploratory Game Design
Exploratory game design tends to throw the player into some situation, or a series of situations, and let them figure out what to do. This broad description manifests itself in all sorts of ways, but basically the key factor is this: the player is very frequently able to choose what their next challenge is.

Players have overall goals, and numerous other smaller goals that they can pursue, and they choose what they want to do. Ultimately if they want to win the game they probably have a set curatorial-style progression that they have to follow, but there is so much other content that the curatorial parts are just one facet of the wider exploratory design.

Examples of very exploratory games:

Grand Theft Auto and all the other sandbox games are, of course by definition, very exploratory. There is a set of central story missions, and then a bunch of other side missions, and then a whole wide world to explore and simply play around in. Personally I loved the first two GTA games back in the day but never played the story missions -- I was able to exist in that world simply by exploring around and dodging the cops and so forth, and never felt compelled to do much with the curatorial parts. I believe I used cheats to get to the second and third cities, because I wanted to go to them but had no interest in the story progression that would lead me there.

The original Legend of Zelda is a great example of a highly exploratory game. Later Zelda titles emphasize the curatorial a bit more, but still have a highly exploratory element. The original game allowed players to complete the dungeons in somewhat variant order, and the players could get all over the overworld pretty much from the start. Even just finding the next dungeon was a game in itself, and going about that was left up to the searching strategy of the individual player.

My own game AI War: Fleet Command is hugely exploratory. A very few players have complained that there was not a scripted curatorial-style campaign, but that's rather missing the point. AI War is built as the strategy-game incarnation of how I played GTA: there is stuff all around the galaxy for you to find, and there are goals big and small for you to pursue, and ultimately you have a goal but you can go about achieving that goal (and preparing yourself for that goal) in absolutely any way you want. Various other strategy games, mostly in the 4X side of things, emphasize the same: Civilization IV is another great example.

A Wide Spectrum
As I noted in the introduction, most games don't fall solely into just one category. Exceptions might be games like Sim City on the exploratory side (there's not even an overall goal, no curatorial content whatsoever) or the rail shooter subgenre on the curatorial side (there's absolutely no way to explore).

But everything else falls somewhere in between, and figuring out the balance between the curatorial and exploratory aspects is a real challenge. Clearly there are players who are very much fans of both. Personally I found that I couldn't get all that much into curatorial game Uncharted 2, while I developed a completely unexpected attachment to exploratory game Red Faction: Guerrilla. But by the same token I absolutely loved curatorial Half Life 2 (like seemingly everyone else), while having a hard time finding much interest in GTA III and beyond despite their exploratory aspects.

So clearly there is much more going on with all these games than just their curatorial versus exploratory aspects -- the actual execution of a game, plus all its interlocking mechanics, are what make a game appeal to certain players or not. And I definitely don't think that one method of game design is more "correct" than the other. I probably have some exploratory leanings, but not as much as you might expect from someone who made a game like AI War.

My big challenge back in 2008 with Arcen's upcoming Alden Ridge came from me struggling with integrating exploratory aspects into a game that had started out completely curatorial. Our current game-in-beta, Tidalis, is a pretty much even blend of curatorial and exploratory, in that it has a linear adventure mode, but that's only a small portion of the wider game modes and options that are available (and even in the adventure, there are secret levels everywhere -- almost as many of those as regular levels in the main progression).

If you look at a lot of successful contemporary games, like the Mario Galaxy games for instance (or any of the 3D Mario games, really), they have a huge mix of curatorial and exploratory design elements. There is a linear progression of levels, sort of, but each level has multiple goals in it, and many of those are optional and/or can be done out of order. You can "beat the game" by simply collecting 70ish stars/shines out of the 120 that tend to be present, and you can do so in any order you want. Demon's Souls is another game like that: it starts out linear, and the individual levels are all pretty linear, but there are a lot of side branches that are optional, and the order in which you approach the levels is something you can switch up to a fair degree. More so than the recent Zelda games, and more on par with the original Zelda, at least.

In Conclusion
All this buzz exists around cinematic games and sandbox games, but the underlying principles at work here are as old as the medium.

The more curatorial games often create experiences that are more resonant in a storytelling sense because of the increased control they give designers; and when you pair that with the "cinematic" techniques of games like Uncharted 2, the result is definitely something Hollywood-esque. There are some seriously great things about that approach, but you also lose a lot of what makes for replay value in games, and a lot of player agency is lost as well.

On the exploratory side, there is a ton of player agency, but it is a really fine line to walk to not have players just feeling like nothing really matters and that the world is boring to inhabit. That was my challenge with the later GTA games, so clearly that's something that varies by player taste -- others were really compelled by those games in a way I was not. It's also pretty hard for players to feel emotionally invested in a purely-exploratory game if it doesn't have at least a single thread of curatorial progression. I think that's where the occasional complaint of the lack of scripted story in AI War comes from. It's also why players aren't emotionally attached to Sim City in the same way that they are to, say, Chrono Trigger (my favorite game of all time, alongside Final Fantasy VI, for the record).

I think it's telling that the games that I consider to be my overall favorites, and which I had the greatest emotional response to, where all mostly-curatorial with some exploratory sections: Chrono Trigger, FF6, and Silent Hill 2. But at the same time, the recent games that I most enjoy playing, and that I think back on with the fondest memories of inhabiting, are games that have large exploratory components: Red Faction: Guerrilla, Mario Galaxy 1 and 2, Far Cry 2.

To me, the debate over the merits of sandbox versus cinematic games could not be more pointless: it's like debating action versus comedy movies. Put simply, people should make more of all of the above!

Tuesday, May 11, 2010

Designing Emergent AI, Part 6: The "Tempo," and AI vs. Player Agency

In the fifth installment of this series, I talked about the reasons for not over-engineering AIs and thus letting them fall into traps by making them too predictable. That was intended to be the final article in the series, unless more questions came up that I needed to address. Recently a new topic has arisen, however, in the form of a question about player agency and why the AI in AI War is deviously reactive rather than taking matters into its own hands too much. This post is actually as much about game design in general as it is about AI design, but those two areas often go hand-in-hand.

Why Does The Human Player Always Have The "Tempo?"

Q: In AI War, you've made the AI play by different rules to the player. Much of this works, but the bit that I think is a bold red, font 100 minus is what seems to me could only have been a conscious decision to make it a passive, reactive and predictable creature.

In the classic pvp strategy games, after the learning curve, each player knows at least roughly how well each of their units do against their enemy's units. At the macro level, the core game is about out-thinking your opponent, deploying the strategy that counters the opponent's current strategy, aware of the likely reaction from the opponent because of your current strategy, and being prepared to change your own strategy in the near future to stay one step ahead. It is about acting and reacting. It is this dynamic which keeps people playing, which explains why Chess and Starcraft are still going strong.

A pvp game inherently has a huge imperfect information mechanic: you don't know exactly what your opponent is going to do. Since AI War is played against an AI, one which does not actually have a macro strategy, which only reacts to the player's moves, never initiating any, you don't have this mechanic or depth. I'm hoping to get across why I think AI War's gameplay is lacking that special something to get it from 'very good' to 'excellent'

Response from Chris Park, AI War's Lead Designer:

AI Modifiers and Other Existing Features

Are you aware of the AI modifiers for cross-planet waves and no wave warnings? A lot of what you are looking for is accomplished by that alone. You may even want larger waves on in order to tailor the experience even more that direction. The macro AI is actually quite devious in terms of what it does with ships once it has a bunch of them free as threat -- a lot of what you are describing happens already in those circumstances, but if you're effective at keeping the AI Progress too low then you wouldn't see that much; it's entirely possible you're on too low of a difficulty there.

Thoughts On the Tempo, And Why The PVP Model Is Not The Goal Here

Beyond that, though, you are entirely correct that the human players are given the "tempo" in this game. That makes it much less like the AI is a full opponent, as you are noting, and more like it is a cross between an opponent and a "game master" in terms of a pen and paper RPG or similar. This, as you might guess, is by design -- this game is about the player, and what they want to do. The AI does indeed throw in some monkey wrenches from time to time, and will kill you if you're not careful, but it's not as independent-acting in a macro strategy sense (most of the time having very few free Threat ships), because that's simply not the goal.

And why not have that as a goal? Well, here's why I hate pvp RTS games, as a blanket statement: the other player is doing stuff invisibly, I'm doing stuff invisibly, and then we finally see what we are doing when we meet, and whoever randomly did the best thing wins that encounter. If you built lots of horsemen and I built lots of pikemen, I win. If I built regular foot soldiers, you win. Then, depending on how unmatched we were, one of us might completely kill the other, or the losing player might scrape by and survive and battle back. Repeat.

To me, that's all those games are, and it hasn't been fun for me since Age of Empires II when I realized what the deal was. I was very much a fan of pvp RTS in the Warcraft II, AOE, and AOEII days, and even to a smaller extent with Empire Earth, but round about that time I was done with it and haven't looked back, and have been eking out an existence in a sort of scaled-down co-op purgatory against the AIs since then. I've had a ton of players write me to say that they haven't played any RTS since the Warcraft III days or similar, but then got back into strategy games via AI War, and I think this is a large part of the reason why.

Player Agency, Comparisons To Espionage Games and Tower Defense

Because, in the end, AI War is as much a puzzle game as anything else. It's a very complex puzzle that changes slowly over time and has some ability to throw monkey wrenches at you every so often, but overall it is an engine for letting you devise very complex and long-term plans, and then see them out. Of course you have to change your plans as the AI grows and the situation changes and such, but in the main it's about you and your team's cleverness in a complex scenario.

Think of it like one of those espionage games where you play as a team of commandos that has the building schematics where the terrorists are, and you plan out a route and then go busting in to save the hostages. The terrorists react and move around and generally foul up your plan, but if you plan is well designed then overall it still goes vaguely like you had hoped. The fun parts are the initial planning, the ongoing adaptation that the monkey wrenches cause, and that feeling of satisfaction at having bested the scenario.

Actually, that's what is great about the best tower defense games, too, like PixelJunk Monsters. And it was the thing that kept me going with all the other RTS games that I played over the years, from AOEIII to Rise of Nations, etc. The AI might be doing whatever on those maps, but overall it was just a matter of finding your way through the puzzle of their stock behaviors to grind them down. That was always very fun, to a point, but once the AI became too predictable and once I had my build patterns down, I was done.

How AI War Blends All That Together, And How It Has Grown

AI War was therefore built around having the players constantly off balance and having to adjust their strategies, and having much deeper and longer-term strategies compared to the "comp stomping" in those other games. That said, the AI is always growing an changing, too -- layering on many sorts of complexity in its behavior makes for a more interesting and varied experience, and makes it more effective at surprising the player.

There was a time when the AIs just had waves, special forces, guards, and "free" ships that ran right at you most of the time. Then Cross Planet Attacks and retreat/regroup behavior was added, and a lot of the finer mechanics were tuned. Astro Trains have always provided another layer of AI behavior, but they've had mixed popularity at best. Then we added minor factions of various sorts, which shook things up almost as much as the CPAs. The recent Border Aggression feature is another major step for the AI, adding on a lot of potential complexity for them in the later game in particular.

And we're planning further minor factions and have that new "entourage" behavior for starships and fortresses and similar in the works, which should also make things varied in yet another way.

All of these things combine to create an AI that is varied and that can surprise players, and that is always getting better and doing so. It has a lot of various emergent activities at this stage that genuinely surprise me when I play, because I never programmed anything of the sort in, and it's fascinating to me to see how that sort of thing comes out of those layers of complexity based on multiple simple overlapping rulesets. That's why I keep focusing on adding more rulesets as much as possible, because that leads to even greater variance. And having more ships with varying attributes and gameplay mechanics attached is also related to that. It's something that has to be done pretty carefully over time, though.

Even So, The AI Won't Have The Tempo, And Here's A Better Example Of Why

And, of course, none of that is ever really going to lead to an AI that has the tempo, as that's something I studiously avoid for the reasons noted above. There's a reason that the terrorists in Rainbow Six don't use squad tactics and come flanking you and executing deep strategies -- that would take all of the fun parts of the strategy away from the human players in those games. I've played many other FPS games where the AI adversaries do indeed use much more strategy, but since those strategies are largely invisible all it tends to affect is how many guys you wind up facing at a time, and where.

Perhaps my favorite example of great FPS AI is Far Cry 2, because they act like real people and respond intelligently (most of the time) to what I do, but I'm hiding in the bushes and moving around sniping them off, creating distractions, and so forth. If I make a mistake and break cover that will be my death, and if I do something stupid like rushing straight into the camp with guns blazing, then they will also kill me. Heck, if I approach a camp and just sit in one position firing on them, they'll flank me and kill me, so I have to keep moving and really act like a guerrilla. Come to think of it, Red Faction: Guerrilla did much the same sort of thing, too, and I really enjoyed that there. But in both of those games, as long as I keep hidden and/or behaved so that I'm not noticeable, the AIs are entirely no real threat to me. I have the tempo and can do recon, decide when and how to strike, and then am left with the challenge of dealing with the hornet's nest that I've just kicked.

AI War is like that -- that's the intended design, and it would be a pretty fundamental thing to change it and give the AI the tempo. If the bad guys in Far Cry 2 were constantly prowling the jungle on high alert for me, that would have just made it Half Life 2, which I found to be a wonderful game but much less satisfying tactically.

In Conclusion, The Tempo Thing Is Neither An Oversight Nor Accidental In The Design

Starcraft and Age of Empires and all those other games do what they do for good reason, and in a pvp type arena they provide a lot of what I used to enjoy in both Warcraft II and Counter-Strike. There is value in that, and I'm glad that other developers continue to make games of that sort, even if I'm not specifically interested in them myself anymore. But trying to bend the Rainbow 6es, the Far Cries, the Red Factions, and the AI Wars to be more like them is not something I'd want to do.

I hear what you're saying, but it isn't like it's accidental that AI War is crafted the way that it is. I think the reasons that you mention that hold it back from being "excellent" to just being "great" is what makes some people consider it really excellent at all.

AI Article Index
Part 1 gives an overview of the AI approach being used, and its benefits.
Part 2 shows some LINQ code and discusses things such as danger levels.
Part 3 talks about the limitations of this approach.
Part 4 talks about the asymmetry of this AI design.
Part 5 is a transcript of a discussion about the desirability of slightly-nonideal emergent decisions versus too-predictable "perfect" decisions.
Part 6 talks about player agency versus AI agency, and why the gameplay of AI War is based around keeping the AI deviously reactive rather than ever fully giving it the "tempo."

Wednesday, May 5, 2010

Like Chess, A Game Of AI War Has Three Abstract "Phases"

Normally I don't cross-post topics from the Arcen Wiki, but this one I wrote tonight, and was so fundamental and helpful in explaining many of the intricacies of AI War in broad terms that I thought it would be a worthwhile one to share directly here on my blog.

Q: The game doesn't seem to have any literal "ages" or anything like some other strategy games do, but players talk a lot about the mid game and the late game and so on. What are those like?

A: AI War, like Chess, is perhaps best thought of in three distinct phases -- despite the fact that neither of these games has any literal delineations (no "okay, the switch has tripped, so it's the endgame now!"), there are vastly different activities being undertaken in each of the three phases of both games (for Chess: opening, middlegame, endgame). We'll divide our discussion of AI War up in the same fashion


At the start of the game, things are comparably easy -- but don't mistake easy for unimportant. Your first half dozen moves in Chess are usually pretty easy, too -- to the point of being memorized/scripted in many cases, which is not true in any fashion for AI War -- but in both games those early moves can set you up well or poorly in terms of position, tempo, material, and so forth. Treating the early game too cavalierly is a leading cause of death later in games.

Human Activities

  • Early Land Grab
  • Early Scouting
  • Early Defenses
  • Early Economic Setup
  • Trying To Determine The Type Of AI Opponent If Playing Against Random
  • In Multiplayer, You Might Start The Earliest Parts Of Specialization Of Players

No matter what your ultimate strategy is going to be during a given AI War campaign, most likely the opening phase of the game is going to be centered around a land grab where you try to take territory to improve your economy, while also trying to gate raid, put up defenses, and otherwise protect your home planet.

Advanced players will generally try to strike as hard and as fast at enemy planets as they can, disabling them before they have a chance to reinforce. This has to be counterbalanced against your need to defend yourself, which can be tricky depending on the map and the AI personality. That's part of why scouting is so important: it lets you identify targets of opportunity, as well as planets that will develop into serious threats if left unchecked, as well as learn about the opposing ship mix in general use, as well as attempting to figure out which AI personalities you might be facing off against.

All of those early information-gathering activities are critical to making good long-term decisions and ultimately crafting a victory. Which planets you choose to attack in the early game might not make a whole lot of difference to how hard the average early game will be, but it can be absolutely critical for how hard or easy the middlegame turns out to be.

Often there are high-level enemy planets right near your territory, and the choice at this stage is whether to attack them early (and thus be over-matched while the other nearby planets all get lots of time to reinforce in your absence), or to just go for the really easy captures and thus let the high-level enemy planets get the reinforcements (often which causes them to be a permanent issue for you for the rest of the game).

Counterbalanced against those concerns of the potential offensive threat of the various nearby AI planets, you have the issue of trying to take planets that will best serve your economy. That means taking planets that are rich and resources, sure, but it also means taking planets in a strategic fashion that they create bottlenecks that help protect your resources. If your planets are constantly losing their harvesters due to inbound AI aggression, then the middlegame is going to be much harder.

If you don't scout at this stage, then you're just playing blind, and that's never good. Ideally if you are on an 80 planet map, you'll get visibility into the nearest 10-30 planets, depending on the type of map and how much effort you put into it. There are so many other critical things to do in the early game (land grab and defenses most of all) that you'll often be too busy to pay much attention to scouting beyond the immediate area around your home planet. That's okay, because the main decision points at this stage of the game almost always involve planets that are 1-3 hops away from your home planet at most. First priority is to "secure your hinterland" and set up a platform for your expansion into the larger galaxy, and then you can make decisions about where to go in the larger galaxy later on.

Rate Of Human Expansion

During the first hour of a game, usually most players will be able to take 3-6 AI planets for themselves with some degree of ease. This is faster than at any other point in the game, but might be slowed down significantly if players decide to take on a high-level planet very early in the game. In the second hour, the players might then take another 2-4 planets on average, usually arriving at between 8 and 9 plants total by the time the opening phase of the game is done.

Of course that all varies by play style, and some players will prefer to be much more defensive and thus taking far fewer planets in total. Or players might find a snug little corner of the map that they decide to clear completely out and thus have a single incoming bottleneck to a smaller group of planets. The defensibility of that sort of bottlenecking certainly makes it a worthwhile thing to consider if the opportunity for that presents itself.

As with all the phases of the game, there isn't one right way to play, so expansion rates will vary quite a lot from player to player. The above is perhaps the average for someone who is neither a turtle nor a rusher, and who is playing against a fairly evenly matched AI opponent on a moderately-tough map, though.

In general, by the end of this phase you probably have between 100-300 AI Progress, hopefully on the lower end of that spectrum.

AI Activities

During the opening of the game, the AI has pretty pitiful waves due to extremely low AI Progress. The waves will grow in size as the opening progresses, but overall they are small enough that advanced players can design automated turret defenses (maybe with a few mobile ships aiding on Free Roaming Defender mode) that will hold off most waves without any need to oversee the battles. Generally this would be incoming waves of a couple of hundred mark I ships in single player.

The AI is far from idle, however. Consider these just probing raids -- they are enough to make you have to spend metal and crystal on defenses, and they may cripple your economy a few times if you are not careful, costing you precious time/resources, but usually they aren't enough to really threaten a loss.

On the other end of things, the AI is also at its weakest point defensively, but it is also at its most concentrated. Depending on the type of map, there are likely only a few planets that are "on alert" thanks to being near your planets. The AI doesn't get very large reinforcements or very many reinforcement points at this stage of the game, but what they do get is likely to be pretty concentrated.

One effective strategy to intentionally prolong the opening of the game -- to your advantage -- is to intentionally put more planets on alert. Thus the AI feels like it has even more territory to defend, but it doesn't have the power to really defend it all quite yet. That can really buy you some time, which is helpful.

Time In Phase, What Causes Move To Next Phase

The amount of time spent in the opening phase really depends on the player and their play style. On average, it's usually going to last between one and two hours. But if the AI is of a very high level or a particularly difficult personality, that's going to make the opening phase shorter.

Generally it is to your advantage as a human player for the opening phase to last as long as possible. The opening pretty much ends when you hit an impasse and can no longer make a quick land grab in an effective fashion, and can no longer defend yourself completely on autopilot (unless you're really lucky or entrenched).

Risk Of Losing

This part of the game is when you have the least buffer between the AI and your home planet, but at the same time it's when the AI is the weakest. It's also the time when you have the least resources, but you also have the smallest territory to defend, which means that your resources can generally be spent more effectively. All of these factors usually cancel each other out in such a way that it's pretty hard to lose in the early game unless the AI opponent catches you just wrong, or a map is particularly brutal, or you're playing against an opponent far above your skill level.

Losses at this stage of the game do happen during normal play, don't misunderstand, but usually not unless they happen to have a big wave of bombers/tanks or ships that pass through force fields, and you happen to be really distracted or careless. In Chess, it would be rather like being caught in the Fool's Mate, the Scholar's Mate, or similar. It's actually easier for that sort of thing to happen in AI War than in Chess, as there are far more than just two common ways for that sort of early loss to occur, but the overall likelihood is still comparably low.

The opening of AI War is essentially your preemptive strike, and you have the definite advantage in your local neighborhood during this time -- but it's also easy to squander that opportunity if you don't make good decisions that set you up for later success. Despite having the advantage, there's nothing cavalier about this phase.

Chance Of Winning

At this stage of the game, you have basically no hope of winning. The AI outnumbers you by a vast amount, and if you try to go for an early victory against the AI (assuming you can even find their home planet at this stage in the game), your ships will be so outclassed that you likely won't make it even past their wormhole. Scratch that -- you'd never make it even past the "core" planets guarding their home planet with any force worth having.

If you even try to go for a win at this stage of the game, you're just wasting your time and squandering an opportunity to set yourself up for a more solid middlegame. The best advice here is typically to just not even worry about defeating the AI, and to just worry about setting yourself up for a later victory (along with making sure that you don't lose in the meantime).


Once you hit a wall with your initial rush of expansion, then the real long-term strategy begins. Of course, everything you did up until now is going to be echoing heavily throughout the rest of the middle-game, so your early strategies have hopefully set you up in a superior situation for the middlegame. If not, prepare to spend the next while trying to dig yourself out of the hole you put yourself into.

Human Activities

This is where the meatiest strategy of the game comes in, so it becomes harder to generalize here. Generally speaking, you'll have identified some desirable targets during the opening of the game (of economic, technological, or just general military value), and you'll pursue those targets. You're likely coming into this phase of the game still not knowing where the enemy home planets are, and still being blind to half of the map or more, so scouting will continue to be of importance.

Perhaps the biggest general goals of this phase would be:

  • Definitely try to find at least 3-4 of the Advanced Research Stations.
  • Definitely do a lot of knowledge gathering, possibly knowledge raiding, in general.
  • Maybe take some more planets for resources, or maybe adopt a resource-efficient raiding strategy if that's your thing.
  • Definitely try to find at least one Advanced Factory that you can adequately defend.
  • Definitely avoid putting the core and home AI Planets on alert if at all possible.

And then there's a whole mess of other possible goals that might be of huge or insignificant importance, depending on your play style and the map and the situation as you evaluate it:

  • You might want to locate the AI home planets so that you can start planning a route to them, or you might not care about that yet.
  • You might want to capture some large weapons, like stuff from Fabricators, or Golems, or Zenith Reserves, etc -- or at least have them ready to be captured if you later decide you need them. Or you might ignore them completely, opting for lower AI Progress instead of that extra firepower.
  • If Astro Trains are on for the AI, you might take a lot of care to re-route them to avoid having them constantly messing with your defenses.
  • In multiplayer, you might be really seriously specializing per player at this stage.
  • In solo or multiplayer, you're probably leaning more towards fleet ships or starships by now, possibly with some mix of both.
  • You might be spreading out the AI Alert in order to keep the AI from bunching against you too much.
  • You might go on an AI-Progress-reducing quest against Data Centers, Co-Processors, and the like.

And so on. There are dozens of capturables, especially if you are playing with The Zenith Remnant enabled, so there are literally more good things that might be helpful to you (and more bad things that might hurt you) than you can possibly hope to capture (or destroy) on a given map. That's part of why having a good picture of as much of the map as possible is important, because even if there are awesome (or terrible) things that you want to react to nearby, there might be more awesome (or more terrible) stuff further out. It's hard to make good long-term decisions when you have incomplete data, to say the least.

You'll also be really focused on defending yourself, but this will be covered more under the AI Activities section for the middlegame.

Rate Of Human Expansion

By this point in the game, you'll probably be down to capturing a new AI planet on average every 45 to 90 minutes or so. This also means that the AI Progress will be going up slower, but by this stage the AI is already a lot more powerful, and as you try to further expand your territory, you're spreading yourself increasingly thin.

Some advanced players on very high difficulties will practically stop capturing new planets all together, except to provide supply for transports or a way for them to capture Advanced Research Stations and so on. Those same players tend to abandon those planets they capture after they get the goods, so to speak, and also tend to focus on data centers to keep their AI Progress as insanely low as possible.

On the flip side, some other advanced players go for an incredibly completionist style of play that can make a single campaign last 40-60 hours or even more, 2x-6x as long as most campaigns should really take. These players tend to take all the planets in a carefully orchestrated strategy that sends the AI Progress sky high (and AI reinforcements, too), but which is managed by careful bottlenecking and incredible defenses. These players might capture 100+ planets in all, but that's certainly atypical.

Both of the last two paragraphs actually reflect very atypical advanced strategies, and one that most players would find boring or frustrating. But those two strategies reflect the extreme polar ends of what is possible in AI War, and what a few hardcore players gravitate to for whatever reasons. Most players adopt a strategy more about moderation, and keep a middling AI Progress and capture perhaps 20-30 planets at most by the end of the middlegame.

AI Activities

During this part of the game, the AI really comes into its own and is more of a threat than ever. You'll run into Cross Planet Attacks (CPAs) every few hours, Border Aggression will start becoming a problem to an increasing degree, and even the general waves will become more of a problem.

Most waves will still be dispatched fairly easily and with mostly automated defenses, assuming that you're keeping up with your gate raiding and creating whipping boys wherever possible, but the risk of a wave being slightly more powerful than you expect and thus overrunning your defenders is ever present. It is possible -- and somewhat common -- to lose whole swathes of planets to this effect.

This is where you need some degree of defense in depth, protecting your softer internal planets rather than just creating a harder outer crust. That in itself is a challenge, though, mainly for reasons of ship cap caused by limited knowledge and resources in general. The more you spend on defense-in-depth, the less you have for your frontline defenses and for your main offensive arm, so this is a real challenge. Some players use basic defenses paired with a mobile fleet that acts on offense except when being withdrawn to defend against a major AI incursion, for instance -- thus keeping things flexible and generally offensive-oriented.

Even on engagements with the AI where you win, individual ships can sneak away and then pose quite a threat if they are not noticed. Sometimes they'll retreat out into the galaxy and then return later with friends (or at a different location), and so it's important to keep an eye on your Threat meter in the upper right of the screen for this reason. Other times a small band of AI ships will break through your outer defenses and rush through your inner core of planets wreaking havoc while you're distracted with the main force.

It's important to watch out for that, especially if there is a lowish buffer between your outer shell and your actual home planet. As in Chess, it's easy to accidentally leave your "king" open and get checked or checkmated when you weren't expecting it. So it's definitely good to make sure that your defenses are generally as automated as possible in most parts, so that you can keep an eye on the overall battle and status of the galaxy, manage your offense, and deal with any breakaway AI ships that make it through to attack you.

During this time the AI is also on serious defensive mode. By 5-6 hours into the game, some AI planets will have been on alert for practically the whole game now, and they will have 4000+ ships in many cases by this stage. The ramifications of that are discussed here. Part of your offensive and defensive strategies will involve keeping an eye on what the AI is doing defensively, and adjusting as needed.

Time In Phase, What Causes Move To Next Phase

The amount of time spent in this phase is really hard to quantify, because (again like Chess) it makes up the bulk of the game. The opening for AI War is a pretty predictable 1-2 hours with relatively fewer exceptions. The endgame is also shorter, as it mostly consists of an assault on the two AI home planets. But the middlegame fills all the rest of that space.

Most games of AI War last for around 9-14 hours on an 80 planet map if you're using non-extreme strategies. On a 40 planet map, it's perhaps more like 7-12 hours on average, and it's a bit longer with even larger maps. The smallest maps are, somewhat paradoxically, neither easier nor particularly faster to play than a 40 planet map.

Some players that are very advanced and use game settings to make a more expedient game can get a faster experience, and those that are really completionist can certainly turn a single campaign into a marathon to rival the length of many RPGs.

The trigger for moving to the endgame is pretty clear cut: the human team decides that they are ready to start attacking one or both of the AI homeworlds. Of course, plenty of games never get to this stage at all... because the humans are already dead.

Risk Of Losing

The risk of losing is actually higher during the endgame phase, but the risk is high enough during the middlegame that a huge number of games never make it to the endgame. If you're playing at a difficulty level in keeping with your skill level, this is a pretty hard game. If you play on difficulty 10 or something else equally insane, it's pretty much Dwarf Fortress hard.

The three main causes of loss in this phase of the game are, in from riskiest to least-risky:

  • 1. Getting swept out of the galaxy by a massive cross-planet attack.
  • 2. Having a small band of AI ships break through your defenses at some point and checkmate you on the sly.
  • 3. Having your defenses unexpectedly overrun while your main fleet is elsewhere, possibly because of killing a high-level AI planet's command station early (with the risks that entails), or because of roaming Threat, Special Forces, the presence of some hostile minor faction, special forces ships, or some combination of the above.

To avoid losing, you want to make sure to do some or all of the following, in no particular order:

  • Keep the AI Progress as low as you can, via killing data centers or by only taking targets that are really worth it.
  • Keep really good defenses in place, and defenses-in-depth, using lots of turrets in particular, but also mines, fortresses, or whatever else you like.
  • If a CPA seems far larger than you can possibly defend against, consider nuking them if a few thousand of them are all on one planet at once. Usually it isn't worth it unless you can get about 3,000 or more of their CPA ships in one blast, but it just depends -- if it's the difference between a loss and surviving, go for whatever works.
  • Have as good an economy as possible, so that if your main fleet gets wiped out on the front lines, you can crank out replacements by the time the AI reaches your inner planets. If you're always running yourself at ship cap due to being that rich, you can also invest in mercenaries in the meantime, which expands your forces even more.
  • Consider all your options when faced with a seemingly impossible incursion -- use lightning warheads to strike the AI when they are planning to come through a wormhole, use multiple EMPs to stun them before you go in for the kill, use the fleet starships to boost your turrets and/or other ships, try to strike the onrushing AI ships in smaller batches where you outnumber them if at all possible, etc.

Chance Of Winning

Given the definition of the endgame, it might seem to be logically impossible to win in the middlegame (the endgame starts when you decide to go for the AI home planets, and you win by going for the AI home planets). However, it is possible -- if rare -- to be able to win during the middlegame.

Generally there's a pretty defined arc to when you hit the endgame: you've really beefed yourself up to the point where you feel like you'll be unstoppable or close, and you're ready to go kick some AI butt. There's always more that you could do to beef up more, but you're confident and ready.

However, the rare exceptions come in when you notice some sort of opportunity early, and decide to take it. Maybe you just got a golem or a giant zenith reserve cracked open, and the AI Progress is low enough that you want to take a whack at those home planets without fully beefing yourself up to where you'd normally need to be. That's harder than it might sound, even with a golem, because the AI has Mass Drivers, warhead interceptors, ion cannons, and core ships galore on those home planets and the "core" planets around them. So generally you'll need to be very clever and lucky with your main fleet and/or transports to make way for the larger hardware (even starships) that you're intending to strike with.

If you go that route, then your strategies would switch over to the endgame ones for a while -- but if you fail, then you can just go right to the middlegame, lick your wounds and continue beefing up your forces, and then go for the endgame a while (sometimes hours) later.


So you've survived to the point where you think you're strong enough to take out the AI home planets. Now it's time to prepare for the final assault, and then see if all your hard work and preparation will pay off.

Human Activities

You're out to kill the AI home planets at this stage. You know where they are, and hopefully you have scouts sitting on them and other key planets providing you a continuous stream of up-to-the-second intel. You can attack either planet first, or you can even try attacking both planets at once.

There's a huge jump in AI Progress (and often Threat) after the first planet, so it's wise to hit both planets at once if you can do so -- but usually it's a stretch just to get one at a time, even in multiplayer. You'll probably set up a forward base with docks and such 2-3 hops away from each home planet, which will put the "core" planets on alert but not the home planets, and then you'll likely use transports or otherwise carve a path to the AI home planets themselves.

There are a variety of strategies for actually taking the home planets out, but they vary so heavily on what ships you have and how you've built yourself up over the course of the rest of the game that it's even harder to generalize here than it was for the middlegame.

Rate Of Human Expansion

Most likely the only reason you'd take a planet at this stage of the game is for military advantage. In other words, for creating those forward bases 2-3 hops away from the AI home planets. Presumably you already have all the technology you need or want, you already have the resource throughput you desire, and everything is set up for you to try checkmating the AI. Granted, things may go wrong (and often do), so you may need to regroup and get more knowledge or resources or whatever than you thought, but as much as possible the goal is just conquest here.

AI Activities

Everything that was said about the AI during the middlegame still holds true, but the AI will also make special efforts to reinforce its home planet when threatened. If you've kept the AI Progress low and kept the AI home planets off alert until the last possible second, you might actually find the destruction of the first AI home planet somewhat anticlimactic, though certainly not a pushover.

However, this can be quite a false sense of security. Upon killing that first AI, you immediately jump around 100 AI Progress, a bunch of core ships likely go free from that planet and may make a beeline for your home planet (which you may not be watching very carefully at this stage), and if you're playing with AI Plots turned on you might get another nasty surprise of some sort (such as the Avenger appearing and wreaking absolute chaos until you either kill it or hurriedly kill the second AI, whichever seems easier).

One way or another, after you kill that first AI home planet, everything gets a fair bit harder all of a sudden. If you just barely scraped by with the first one, the second one might be even more of a nightmare. That extra AI Progress often pushes the tech level of waves and reinforcements up a notch, which makes your defenses suddenly strained even if they had not been before, and if a higher-level CPA is suddenly declared then you're in real trouble. This is why it's a great idea to be ready to do a one-two punch against both AI homeworlds if at all possible -- though often you just have to roll the dice, since each AI home planet might take 20-60 minutes depending on how well you've set things up for yourself versus how well the AIs set things up for themselves.

Time In Phase, What Causes Move To Next Phase

Really, this is pretty hard to quantify as far as time spent. It might be under an hour, or it might be a few hours. It's generally not a grind-fest, but if you alerted the home planet too early or were more ill-prepared than you thought, then it might turn into one. In those situations you can either cut your losses and regroup for a while, or you can just hammer against the AI with wave after wave.

This phase ends when either you kill both AIs, or they finish off your team. Or sometimes you might revert to the middlegame for a while, but that's less common once you've committed to trying to take out one or both of the AI home planets.

Risk Of Losing

Especially when you consider the normal danger from waves, CPAs, border aggression, special forces, and all that good stuff that's always around (and pretty dangerous by this stage), the time right after you kill the first AI is perhaps the most dangerous time in the entire game. You may have a few hundred core ships streaking towards your home planet right when some other calamity strikes, and that's a good example of when multiple events can combine to cause your defenses to get unexpectedly overrun.

And given that most of your fleet is probably either a) dead from the battle, or b) completely on the other side of the map from your home planet, your ability to bring your mobile forces home to aid in your defense is at an all-time low for the game.

Losing right when you're on the cusp of victory... hurts. So make sure not to get so caught up in your "inevitable" victory that you get stabbed in the back when you're not looking.

Chance Of Winning

This is the part of the game in which you'll win, most likely, if you were going to win. It's still quite hard, make no mistake. Some players that come to AI War from other RTS games think that victory is inevitable when really it is far from it. They look at their thousands of high-level ships, they have scout intel showing that the AI planets are undefended for the most part thanks to not being on alert, and they assume that they can't lose.

This is folly. You can lose at any point, even if you have overwhelming odds. Don't be like the Buggers at the finale of Ender's Game. The AI can -- and will -- do that to you if you're not careful. You could outnumber their home planets 10:1 and still lose horribly because your fleet was in the wrong place and your defenses weren't enough.

So proceed with caution, at any rate. Another wrong assumption that many players of other RTS games carry over to AI War is that the game will get easier once the first AI player is dead. Quite the opposite, as noted above -- AI War is a team game, when it comes to both the humans and AI players. No human or AI players are ever "out of the game" early; the game ends for everyone at once. So when you kill the first AI home planet, all you've done is really make them angry and more powerful, and both of them are still right there ready to kill you (same principle holds true when they kill one of the human players in a multiplayer game). Of course, killing one of their two home planets does get you that much closer to actual victory, so it's a needed step, but it is a very dangerous step.

If you do manage to win, it's a pretty impressive achievement, as this is quite a hard game on the standard difficulty levels. Congratulations!

Interview with Novelist Stephen Parrish

With us today is my blogging buddy Stephen Parrish, whom I've known for a couple of years through writing circles. I first met him back when The Guardian was seemingly nearing picking up an agent, but then obviously all that fell through for me, and I've lost contact with the bulk of the writing friends that I made during that time.

But Steve and I have kept in touch, and I've been enjoying seeing him break into print -- I picked up his book, The Tavernier Stones, just yesterday from my local Borders, and will post about that sometime in the coming weeks once I have a chance to finish it. Steve is going 'round the blogosphere giving interviews, and so I took my chance to ask him a bunch of rude questions.

Q: Okay, so the obligatory "for those who don't know Steve" question -- can you tell us what your book is about? I bet you have an awesome elevator pitch by this point.

Stephen Parrish: It's based on the true story of seventeenth century journeyman and trader Jean-Baptiste Tavernier. He made six voyages to the Orient, chiefly India, and brought back to Europe many of the world's most famous diamonds, including one that would eventually be recut into the Hope. He mysteriously disappeared during his seventh and final voyage to India. My novel capitalizes on that mystery: it postulates that he arrived at his destination, amassed the largest cache of gemstones in history (including several actual stones that haven't been seen since his time), and was robbed during his return trip to Europe. The legend of the "Lost Tavernier Stones" swelled during the centuries following Tavernier's disappearance, until the day a body floated to the surface of a bog in northern Germany, with one of the stones clutched in its fist. That's where my story begins.

Okay, so it's a little long for an elevator pitch, but just imagine we're going up several floors.

Q: We've all heard the stories about famous books and authors whose work was rejected a bajillion times before they went on to be rich and famous. So how many rejections did The Tavernier Stones rack up? How many times did you rewrite it?

Steve: I kept a log, but I refuse to look at it. I also didn't keep any rejection letters, even from luminaries; I don't believe in saving mementos of failure. I wrote twelve structurally different versions of the novel and rewrote them probably thirty times. The book was turned down by over a hundred agents and editors combined, but I don't know the exact number. I'm baffled when I read about writers giving up after a dozen or two rejections. Agent Query lists over a thousand literary agents. Why stop at double digits?

Q: And this is your third book that you've written, but first to be published, right? Statistically, that would mean you're a bit ahead of the curve of most authors who actually get published. Do you feel ahead of the curve?

Steve: You know, I keep hearing such statistics, that most people have to write five or six or ten books before they break through, but that doesn't match my own anecdotal experience. Most of my friends published their first or second book. Maybe I just hang out with smart friends.

Q: What about unfinished books? How many of those would you estimate you have racked up over the years? And, for that matter, at what point in your life did you start writing?

Steve: I can't pinpoint a time I started writing, it kind of ramped up over the course of my life. Unfinished books? If you count a synopsis and a scene or two, a bunch. If you mean I got halfway through and petered out, a couple.

Q: All right, all right, I'll stop trying to make you look like a loser now. In all seriousness, though, I've read your blog for years and have always felt like you have a really effortless command of language, humor, and all the little techniques that make prose sparkle. Too often, I think that non-writers see a new author coming out and think of them as some sort of overnight success, where that is really almost never the case. A lot of mega-famous authors struggled enormously before they were published, and I think that's part of why their work is so solid. All that "diamonds out of coal and pressure" stuff, right? Do you feel like that's been the case for you?

Steve: My work is still coal. At any given point in my writing journey I've felt I'm better than ever before. No doubt a few years from now I'll look back at this stage and . . . well, I won't want to look back at this stage.

Q: So, realistically, now that you've made it into print, do you think you'd have wanted to short-circuit the process in the past? Speaking personally as someone who has not yet made it into print, but has had a very near miss, I definitely can say that I'm glad my earlier work never made it past the agents barrier, because I want my first published work to really represent me as well as possible. If you could play God and go back and get a magic "yes" from some agents that turned you down in the past, do you think you'd do it?

Steve: Hemingway's first published story was called "Up in Michigan." I'm hard pressed to say anything good about it. Of course, he went on to write great things. I think we should be putting our stuff out there, posting in on our blogs, sharing it with friends, submitting it, hoping it gets into print. Even if people are hard pressed to say anything good about it. Because doing so gives us practice in a way that working in isolation and storing things in trunks don't. If you can publish your "bad" stuff, more power to you; it'll be behind you then, and you can move on. If you wait until your bad stuff is good, you might have to wait a long time.

Q: We all know that you think that being a writer is the best job in the world, and that everyone should be doing it. But you've had a lot of different jobs -- unusually varied for any one person, actually. Can you tell us a bit about that? Was it restless youth, or what? Aside from writing, what was your favorite job?

Steve: Caddy, janitor, bartender, cook, newspaper deliver, factory worker, soldier, typist (clocked at over 120 wpm), teacher, jewelry salesman, cartographer, and a couple of others. Once I was even a paid lab rat: now anytime somebody says "Rosebud" I get hungry for a piece of cheese.

Restlessness probably sums it up. But I have a lot to draw from now, lots of places, experiences, and perspectives. When I run out of things to say I'll go back to cartography. I loved sharing a tradition with people like Eratosthenes, Ptolemy, and Mercatur.

Q: You've made references in the past to the fact that you draw on your past, and very varied, experiences for your writing. That seems like a killer idea. What would you therefore recommend to someone, say, who is just heading off to college to study creative writing or somesuch? Would you advise them to put their writing plans on hold for 10 or 20 years, go out and experience the world more, and then come back and write? Seems like a lot of writers are in their 30s or 40s at the earliest when they really make it into the business, so from my perspective that seems like valid advice, and is basically what I decided to do, myself. But I'm curious what your take is, having traveled that road -- do you feel like you missed out by not writing more instead of working all those other jobs?

Steve: I have mixed feelings. The writers who were already cranking out books in their teens just never seem to have enough depth for my tastes. I think it's most important to read, as much as possible, and to start reading very early in life. It will tune your writing style, for whenever you get around to actually playing some notes. You'll know when it's time; you won't be able to put it off any longer.

Q: Personally, one of my favorite things from reading your blog has always been the wry, often self-deprecating humor that tends to pervade it. Does that show up in your longer prose to the same degree? From the posted snippet of The Tavernier Stones, which I greatly enjoyed, I couldn't tell -- but that was pretty early in for the book.

Steve: Wait for the next one. That's when I stretch my voice.

Q: Most writers become known for writing a certain sort of book, or in a certain style. Stephen King is horror or magic with lots of character analysis, Stephanie Meyer always writes romance with reasonably young characters, Michael Crichton always had themes of the dangers of science left unchecked, etc. Statistically speaking, it's impossible for us to make valid conclusions from looking at your one book. So tell us: what aspects of The Tavernier Stones do you feel like are universal "Steven Parrish" attributes that you'll be known for? Are you planning to be a historical thriller writer in general, or is there some other focus?

Steve: It's always been my ambition to write what is generally considered "mainstream." Nobody seems to use the term anymore, though. I know the arguments in favor of genre labeling, nevertheless I'm opposed to it: labeling is disabling. A wonderful science fiction novel, for example, will be passed over by people who don't normally read science fiction. I don't care of your story has spaceships or cowboys or lovers or vampires; if it's good, I want to read it.

And I don't think that's just me being eclectic or something. I think everyone wants to read a good story; they're just afraid of stereotypical spaceships, stereotypical cowboys, etc.

Q: I have my own feelings on this, as I think every author does, but I'm curious on your take: in the age of movies, television, and video games, what do you think novels offer that the other mediums do not? For people who don't read much fiction (shame on them), what are the reasons they should give it a try?

Steve: My argument isn't going to change anyone's behavior, except maybe parents looking for ways to develop their children. Much as I love movies and games, and I do have a passion for both, they're passive. Stories, on the other hand, require work: the brain has to illustrate the scenes being described. Spielberg isn't there to do it for us. That effort, subtle though it may sound, is not only why reading is better training for the mind, it's also why literature can move us in ways no other medium can. Even without special effects and background music.

Art is the highest achievement of civilization, and literature is the highest form of art. That's my opinion and I'm sticking to it.

Q: Anything else you'd like to add?

Steve: When I was twelve or so (caddying) I asked a rich man for his secret. He said, "Work your ass off. And hand me the nine iron."

Not A Q: Thanks for joining us, Steve! And of course, best wishes for your success with this title and those to follow!

Tuesday, May 4, 2010

All Advice You Hear (Even From Me) Is Wrong

Bold statement, no? But this is a pretty simple post, actually. It shouldn't come as surprise to anyone that we live in a complex world, but we tend to rely on mental models that oversimplify things. Even without discussing heuristics and the problems those can create, we can simply think of this in terms of the "perceptual screen." In other words, we all form preconceptions about all manner of things in the world, and we tend to discount or even completely ignore data that doesn't fit our model.

Yeah, there are exceptions, and yes, many people try to make a point of working around their perceptual screens (myself included), but the fact is that people tend to do so much subconscious filtering that there is only so much your conscious mind can override that. I drive down the same streets every day, and miss a thousand details that are there because I'm distracted, watching the road, or otherwise. Only if something really violates my expectations of what I should see on that road am I likely to notice it, and I might not even notice it then.

Perhaps that's getting overly technical, but my point is simply this: none of us knows what we are doing, really. If someone asks me why I think AI War sold as well as it did, or even why it didn't sell better than it did, I can give my theories, but those theories are wrong and incomplete. If you ask a developer, an executive, and a marketer that were involved with a AAA game the same thing, they'll all give you slightly (or majorly) different answers, too. None of them know, really. To some extent it's unknowable: if you ask the average consumer why they're so excited about some game or another, often their response is going to be pretty vague and unhelpful.

This isn't related to games, either. Nobody knows why one movie is more successful than another, or why one author is suddenly a bestseller overnight when another similar author is not. Sometimes there are obvious things: quality differences, timing differences, the amount of marketing support, etc. Having a quality product that is well marketed -- or an established brand that people are already loyal to -- is a well known way to ensure some degree of success in a project.

Until you start looking at all the exceptions: Rock Band Beatles, anyone? That was way less successful than anyone expected, and it had absolutely everything going for it. In retrospect you can point to this or that factor, but the fact is that no one can be really sure why the game didn't do as well as expected. The problem: there is a market filled with hundreds of millions of consumers, and they have thousands of reasons why they may or not buy something. And not only that, individually they all have various timing issues (being too busy, being out of a job, whatever) that are impossible to collectively predict.

Statisticians may be scoffing at me around now, as many seem quite confident of their ability to predict what large groups of people will do. To that, I say: meh. The best statisticians are pretty accurate in the main, but it's those pesky outliers that confound all the regular models. They're using heuristics or mental (or mathematical) models like everyone else. Their models might well be far better than the norm, but they're still so far from perfect it's not funny. Not that I'm trying to insult statisticians or their work, which is still important, but I'm making a point here: the future is unknowable, and the reasons for past events when those events involve millions of people are also pretty well unknowable.

Here's a fun example: if I were releasing AI War now, instead of last year, how would that have affect things? Would the proximity to Starcraft hurt it? Would I have had an easier or a harder time forging partnerships with digital distributors? Would less or more people have bought it do to economic factors with the recession? Would competition with other AAA and indie games have been better or worse for AI War in 2010 compared to 2009? Even the weather might have had some effect, indirectly, in terms of how much people were inclined to stay inside and play games last year versus this year.

And those are just the major factors. There's also all the hundreds of personal-life factors for every person who bought AI War, or who didn't-but-might-have-bought AI War. It's a mess of probabilities, of emergent patterns based on multiple overlapping systems. In short, it's that whole Chaos Theory thing.

So if you ever ask anyone advice about, well, anything, bear that in mind. We're all walking around behind our own perceptual screens, making wrong and/or incomplete assumptions and models of the world, and acting based on an accumulated multitude of beliefs. And even if your beliefs, history, etc, closely line up with those another person, if they're advising you about something like (say) how to be a successful indie developer, bear in mind that there are still so many different factors that can't be helped: the timing of their release versus yours, the fact that you both developed games that are probably quite different, and so on.

Perhaps this post sounds pessimistic, but I don't mean for it to. Advice, wrong and/or inapplicable though it might be, is still better than no data at all. Seeking advice is a great thing to do, just like the study of statistics is extremely valuable. You just have to understand what it is you're getting out of either process: a model, or a way of looking at some small slice of the world. All models are incomplete, so you can't just rely on any single one. That's why you want to listen to the advice of others, then make your own decisions.

In the end, the only way to find out how AI War would do was to release it and see. Same with Rock Band Beatles. I made my predictions and I'm sure Harmonix made theirs, but in the end all predictions tend to be at least somewhat wrong. If you're an aspiring indie developer, you need to worry about two things:

1. Making the best game you can.
2. Doing the best job you can in promoting it and getting people interested in it.

Then, based on a thousand factors -- many of which you can't control or even really even perceive -- your game will do well, or it won't. Most likely it will fall somewhere in the statistical norm for the type of game you are making and the market in which you release it. There's a chance you might be an outlier in a positive or negative sense, but planning to be an outlier is like planning to win the lottery. Awesome games flop all the time, while their inferior counterparts get accolades piled on top of them. More often than not quality rises to the top, though, so by making a good game you're certainly playing the odds.

I'm sure that I'll continue giving out advice for indie developers, just like other indie developers or industry veterans do, but in the future when you read advice from anyone (me included) keep this post in mind. Too many commentators get hung up on extolling the virtues (or imminent demise) of this or that model -- casual or not, social or not, sandbox or not, are the current hot topics -- and what you need to remember is to do what makes sense for you, and to not try to trend follow. Forge your own path, and see what happens. That's all any of us are doing, really, if you get right down to it.

Tidalis Beta 0.407 Released, Major Updates and New Videos

Version 0.407 of Tidalis, which has been inwork for a surprisingly long amount of time for one of our beta updates, is now out. If you look at the full features list, you can definitely see why it took more time than usual. The major new feature here is multi-well co-op play on the local computer, which has a lot of exciting possibilities that we're just starting to tap. This also paves the way for our Vs and network play modes, which will be making an appearance soon.

Also new in this version are more art/music background themes, more special blocks, the first few usable items, updates to the level editor, added animation techniques and general polish, more block skins including even better colorblind support, and dozens of more minor features. Things are moving fast as usual with Arcen, and we're looking to be right on schedule for our planned full release in early July (Windows and Mac OSX at first -- other platforms are under consideration for the fall and beyond).

We've also put together a new pair of videos for the game:

View Larger Version

View Larger Version