Documentation, you need more of it.

[TechSY730]

Well, let me be straight up and honest.

Your documentation is lacking. Oh sure, your tutorial and manual covers the basics, and the wiki has a bunch of explanations, but it lacks many of the obscure mechanics that can easily make or break a game.

And even if the wiki covers it, it will still annoy people when they lose because they triggered an obscure defensive mechanic by the AI that your pretty much had to know about ahead of time to avoid it. Yes, it may be on the wiki, and they might find out what just happened to them there, but these important (like, you will quite likely lose if you don't know about them important) but you would have no idea if it you didn't read very detailed technical stuff or experience a loss (aka. trial and error gameplay, which is almost always bad design).

This can be a big deal, as if a new player loses because of a non-intuitive and not very well explained in game mechanic, AND the outside the game documentation (tutorials, the manual, and to a lesser extent, the wiki) does not make it clear that you should know this, that doesn't make new players think the game is clever, it makes them think that the devs are just making stuff up and "springing" on you for a surprise "ROFL-curbstomping", which you look like devs who can't keep their own mechanics straight, or worse, jerks. Either misunderstanding are great ways to lose new players.

Admittedly, the problem is much worse in the betas, but that is to be expected.
But here are some things that you need to know to survive (or have a decent chance of winning) but are very poorly explained in game and hard to even notice in the wiki until it is too late.

Solutions, hmm, that is much trickier. Maybe an "advanced" tutorial showing off some of the more obscure but "you need to have at least some understanding of them to stay alive" mechanics. Things like armor and how it influences durability, radar dampening, raid engines, alarm posts, border aggression, AI eyes, black hole generators, force-field immunity, and other things that can easily take a new player by surprise.

Also, there should be an "obscure and/or somewhat uncommon things to look out for or risk losing or making the AI near unbeatable" section in the manual AND the wiki, where they can get an advance warning about these things. In addition to what is covered by the "advanced", it should also list stuff rare enough not to justify putting in a tutorial but still can kill you if you trigger it accidentally (minor faction and AI plot stuff would belong here)

Also, continue work on the tool-tips. Many funny unit behaviors still don't have satisfactory explanation in their tool-tips

Here are some examples I can think of, in addition to what I listed for what needs to be put in an advanced tutorial. (Limiting myself to stuff in the last stable version of course):

Caps on defending AI ships per planet and what they scale with: I know that the number of defending ships that can be on any one AI planet is proportional to the number of guard posts, but this isn't clear at all from either the tutorial or the in game documentation. As such, you wouldn't know why neutering can be important until it is too late (all your border planets get surrounded by near impenetrably defended AI planets), and you have to come to the forums to ask how to deal with situation and then you find out about neutering.

Freed ships waiting on the other side of the wormhole: It isn't explained very well in game why AI units tend to "ball up" around the other side of the wormhole. The intuitive response would be to send more defenders to that planet, in fear of an immanent attack. However, thanks to the not very well explained behavior that the AI will wait until it thinks it has a decent chance of winning (thanks to firepower calculation shenanigans), these extra defenders make the challenge worse. The right responces are either to plunge into enemy territory and take out what you can and then pull back to take care of what went through, or to pull you ships to another planet to "bait" them into the planet, then come back in and crush them. Neither one of these is intuitive, but there is nothing in the game to explain why the intuitive solution is the wrong solution.

The "slowdown and you can't unload here zones" that you hit when you are far away from the center of the planet: No further explanation needed.

Hybrids: How they work is very, very opaque. As such, if a player takes the intuitive strategy with them (take some extra time to build up more defenses so they can survive hybrid attacks), this hurts them in the long run. Thanks to how caps are tied to hybrid spawners, the proper strategy is to find these spawners and kill them decently early in the game. But this connection of hybrid caps to spawners, or even the existance of spawners, is something that is not clear at all.

Fallen spire event attacks: I frequently find myself hesitating to build something new with the fallen spire campaign units (especially the buildings), because I don't know whether starting to build it will cause the AI to send out more event attacks.
This bit me in the butt the first time I played the campaign. When you build the scanning thingy from the first shard, no new event attack is spawned. As such, I thought that if you get the shard to your home, take out any lingering threat, and then start building what the shard gives you on your home, your are safe. Then imagine my unpleasant surprise that this does not hold for the second escort mission. In fact, the wave that spawns when you start building the spire colony ship seems to be even more brutal than the one you got escorting the Spire survivors, and there was no warning about it. Because I thought I was safe, based on my experiences from the first shard, I thought I could build it while my defenses were rebuilding as well. ...Yea, I was wrong, DEAD wrong.
It gets worse, another event attack is spawned when you place a city hub, and when you start building the first Spire reactor on a planet. All with no warning.

Yes, I understand that you guys are limited in man power, but non-intuitive, unclear from in-game sources, or you wouldn't even know about it until it is too late stuff needs to have a clear warning somewhere, or new players will be driven off.

Yikes, this post has gone on for WAAAY too long. I'll stop for now.

[x4000]

I disagree with your premise that new players will be driven off.  I've not seen any evidence of it.

In terms of advanced players with questions, as they come up we answer them, and we try to get them into the wiki.  We also make wiki accounts available for folks that want to add to it (PM me if you want one, anyone reading -- we do it indirectly there to prevent spambots, etc, which had been wreaking havoc before that).  We also update tooltips as outdated or too-vague ones are pointed out to us.

This is the best we can do.  There's nothing to particularly discuss, eh?  You could wish that we were a larger company, or that we didn't need to keep making new games to survive, but neither of those are going to come true. 

"The documentation is lacking" does not, to me, equate to "lacks many of the obscure mechanics," in the first place.  I've seen professional code APIs that were incredibly less documented than AI War is.  The documentation is actually quite good, it's just not perfect or all-inclusive.  In that sense, sure, it is lacking every last little detail about the game.

Here's some food for thought: what sort of documentation did you get with Supreme Commander 2?  Starcraft II?  Etc.  Those all give you a tiny little pamphlet that highlights a couple of units, if they're anything like most games I've played lately (Age of Empires III and Supreme Commander 1 were definite examples I can cite for sure).  They don't even tell you all the hotkeys.  Players compile their own lists based on experimenting and finding them.  Our documentation is fan-freaking-tastic compared to most strategy games, in terms of what is provided by the first-party developer and not third-party (paid!) strategy guides.

[TechSY730]

The thing is, in Starcraft II, the obscure, advanced, poorly documented mechanics won't make or break your game except at high levels of play. (I can't speak for Supreme Commander II, as I don't have it) In AI war, even lower difficulty levels can start really damaging you if you trigger one of these types of mechanics.

But you are right, I was being a bit too pessimistic. Most of my "rant" was about documentation that comes with the game, e.g., the manual, the tutorials, and the tool-tips.
Both the wiki and the forums are MAGNIFICENT. FAR better than almost every other game out there, especially the wiki. When picking up a new game, after playing through the tutorials, reading through the manuals, and playing a few games to learn the ropes, I start wondering about the advanced mechanics and complicated behaviours I have seen, and I start looking for the more detailed online documentation (1st party, 3rd party, or otherwise). All too often, I find that there isn't such a resource.

I guess you guys take the "open source approach" to documentation. Even if your built-in documentation isn't something to sing praises for (though mature open source projects have gotten that nailed down pretty well too), you more than make up for with the online documentation and the community.

Sorry if I gave the impression that I think things are horrible with the current state of documentation. I was merely pointing out that, IMO, the in-game and offline documentation seems to miss important stuff until it is too late. You're documentation overall (including both offline and online sources) is magnificent.
 

[x4000]

Oh, one more thing.  To the specific examples you gave:

Caps on defending AI ships per planet and what they scale with:

This does NOT need to be explained in the tutorials or in the game, particularly.  For your average player, they aren't going to meta-game to the degree where this is even relevant.  For those who would need it, we can put it on the wiki.  But it's also not a particularly actionable piece of data.  Why do you need to know?  In the interest of complete documentation, sure, it could go on the wiki.  But what I mean is, how would this affect your play, even?  For a rookie, it's worse than irrelevant, it's clutter data, actually. The wiki explains, in depth, how reinforcements happen and how to reduce the ship caps the AI has in general on the planet.  If you want to limit a planet, you know what to do.  If you don't want to limit it, you'll see what its cap is soon enough.

Freed ships waiting on the other side of the wormhole: It isn't explained very well in game why AI units tend to "ball up" around the other side of the wormhole. The intuitive response would be to send more defenders to that planet, in fear of an immanent attack. However, thanks to the not very well explained behavior that the AI will wait until it thinks it has a decent chance of winning (thanks to firepower calculation shenanigans), these extra defenders make the challenge worse.

Wow, but you like to meta-game.  The average player just isn't going to want to know the innards of the AI decision tree.  They treat it like a human, and go "what is he thinking now... hmmm... looks like he will soon attack me, etc."  When they attack the AI, the AI then may counter-attack their planet during the confusion.  This is fun!  This is a tricky AI to the player.  This is the AI being clever and human-like, as far as they know.

For the advanced player... the info is freely available, or they can simply learn the patterns.  It probably could stand to be on the wiki, but I don't think it's hurting anyone's ability to play the game not having it there.  Part of the fun of a game is watching your opponent and learning how they act and react, not just having everything told to you up front.  In fact, most of the AI behaviors are not documented, and I have no intention of documenting them on the wiki.  I mention them in passing in release notes for testing purposes and so that players know what changed.  

But isn't the point of an AI system to provide the illusion of greater decision making processes at work?  Doesn't documenting each and every last little thing sort of undermine that?  And isn't that totally irrelevant for anyone who plays long enough to care, anyway, because they'll have figured out the same stuff through observation without having to be told?  Isn't this a lot of questions I'm writing?

The "slowdown and you can't unload here zones" that you hit when you are far away from the center of the planet: No further explanation needed.

You go way away from the planet, and it tells you your speed has been reduced based on being way away from it.  Problem...?  In terms of the unloading thing, that could stand to be on the transport tooltip.  That's something it takes all of ten seconds to tell us on Mantis, and it's just one of many small things that needs a bit of tweaking.  It's a living system, so something will always be slightly out of date.

Hybrids: How they work is very, very opaque.  As such, if a player takes the intuitive strategy with them (take some extra time to build up more defenses so they can survive hybrid attacks), this hurts them in the long run.

The "intuitive" strategy is going to vary by player -- trust me.  And we keep these guys opaque for a reason.  Part of the fun of playing with hybrids is figuring out how to beat them.  If we just tell you how to do it, where's the game?  There is nothing there that isn't perfectly reasonable for an advanced player to figure out through experimentation.  And make no mistake, Hybrids are the most advanced thing we have in the game.

Fallen spire event attacks: I frequently find myself hesitating to build something new with the fallen spire campaign units (especially the buildings), because I don't know whether starting to build it will cause the AI to send out more event attacks.

Again: our goal isn't to help players meta-game.  It's meant to seem more organic than that, though it's scripted.  If you just absolutely can't stand that, then you can always read the very long walkthrough on the wiki, or you can savescum.  Or, you can treat it like a realistic and unknown situation, where you don't always know what happens with the enemy based on you actions.  Personally, we think that's way more fun.

[x4000]

It's no problem.  In terms of the tooltips, they are meant to give the novice and mid-level player what they need (and for advanced players who aren't meta-gaming, past a certain point they are all you need, too).  The tutorials are meant to get people past the novice stage, and that's it.  The manual is meant for people who absolutely have to have a manual, and it's even more novice than the tutorials. 

The wiki is where it's at, as the game changes to frequently for a printed manual to be at all feasible to maintain (seriously, InDesign and similar really stink), and having more or more complex tutorials WOULD turn away new players, no doubt.  It's a very challenging balance with the tooltips and the tutorials in particular to get just the right balance of enough info to play, but not so much info that it's filling your screen or overwhelming.  I think that, as of 5.0, it's at the best stage it's ever been.

Anyway, hopefully my last post on the specific examples and my responses are illuminating what our goals are for that sort of documentation.  We provide as much information as is needed to get started playing and experimenting in all cases, and in some cases we provide some strategy-guide-like details, but that's hit-or-miss as you noticed, since it's a huge game and writing a full strategy guide would be a gargantuan process that would definitely be worth $20 a copy if it were printed up at the end.

[TechSY730]

Hmm, sounds reasonable. I still think that the existence of raid engines and alarm posts and their importance could be made a little more clear in the tutorial, but I can see now that maybe my idea of "advanced" is closer to the average players' idea of "pulling the game apart". Another side affect of my tendency to over-think things.  

Anyways, now that I think about it, when you consider the sheer number of mechanics and behaviors that are in this game, you're offline and in-game documentation is actually really good. Your ratio of "well explained" to "poorly explained" mechanics is quite likely as good or maybe even better than mainstream titles. Its just that thanks to the sheet amount of stuff to cover, there are more things that fall into the "poorly explained" category, even if you are doing better overall. Darn statistical reasoning fallacies that come about from confusing counts with ratios.

[x4000]

Another thing to remember is that a lot of that complexity isn't even in the base game, it's in the expansions.  Hybrids, raid engines, alarm posts, etc -- none of that is in the base game.  It's kind of assumed that those players that get into the expansions are a bit more advanced, since the base game wasn't enough for them, and thus that they can partake of the wiki if they want, or they can just learn "on the job."

The important thing is having it be clear what is happening as it is happening, when it comes to alarm posts and such.  In the past, that was very unclear.  You could trip one without knowing you even had.  One of the big things in 5.0 was making a lot more messages so that it was clear what was happening, etc.  So now a new player that runs into an alarm post for the first time hears the alert, sees the message, and then wonders what an alarm post is.  They then use their scouts to find it on the planet that was already named in the first message, and then they can read about it.  And from then on, they know to avoid them.

Falling into the trap of the AI every so often is pretty cool, actually.  It's that back-and-forth of the learning process, versus being an expert right from the start and never being surprised.  I don't have any beef with documenting this sort of thing on the wiki (and I think it is), and for those who want to study the unit charts that info is all in-game, too, but that's simply not the standard pattern for most people, I don't think.

I'm flogging a dead horse by now, but hopefully you see my point.  I always compare AI War to Chess, but this is one place where the two games really differ.  In Chess, you need to know pretty much every mechanic before you can effectively play even your first game.  In AI War, there's an intentional fudge factor built in, so that you're always expected to learn on the job.  Ender sees a planet full of innumerable buggers, and has to figure out a way to win.  The Terminator is coming for you, and as far as you know that thing is deathless.  So you experiment, and you learn, and you feel justifiably proud of yourself when you actually figure it out on your own.

I think that's one of the things that makes so many people gravitate to AI War, is that they always feel like there is more to the experience than just what they know by heart already.  That's not to say that we can't have a huge and sprawling strategy guide on the wiki, but it is my hope that most players will use that sparingly, if at all, at least at first.  Otherwise that's kind of short-changing them, unless they are hoping to play on the AI War World Tour or something (that doesn't exist, for anyone confused ).

Anyway, "poorly explained" is kind of a misnomer.  Sometimes, like specifically with the fallen spire, we are incredibly vague on purpose.  It's a story, and some stuff happens that you can react to or ignore -- that's all you need to know to play it and have fun.  You'll probably lose your first time through it, unless you are quite good.  Or at least you won't complete that campaign itself.  The problem is...?

I think most AI War players lose at least 50% of the time, if not 90% of the time, and that's another big difference between this game and other games.  That's sort of the point, too.  The AI is good, overall -- you have to learn quite a bit more, and can't just rely on simple tricks.  Even if we had a full strategy guide, few new players could just sit down and digest all of that well enough to win the first time on a reasonable difficulty.  Losing is fun, and is a process toward being a better player.  Winning teaches us nothing. 

I've had players who played the game for over a year before they won their first game, and they were really ecstatic when they did.  That sort of experience, that feeling of accomplishment... that doesn't come when we just hand the answers on the plate.  If you think about many other games of all sorts, it's much the game.  Are the monsters in Donkey Kong all documented with explanations of what to do?  No.  You are taught what moves you have, and then you have to watch the monsters and experiment in order to figure out how to win.  Surely a strategy game should require at least that, as well.

I don't defend the pamphlets that most RTS games give out with their games, as it's just not enough info -- which is why we have a wiki in the first place -- but my hope is always that people will use the wiki just for what they really need, rather than trying to just binge on it.

[Zeba]

Tbh I have never read any of the wiki or any documentation for ai war. I just dove in and learned by playing the game. If I ran across a game ending mechanic I was unaware of I just go back to an earlier save and don't repeat what brought the hammer down on my command station. I mean if you know ahead of time every single mechanic and ai tactic that makes for a rather boring game yes?

[superking]

while I wouldnt say documentation is lacking, I think it would be nice to get more feedback from the game when such events do happen (eg. counterattacks, alarms, AI homeworld fleets etc), it can sometimes be frustrating when a sea of ships pour onto the attack for reasons unknown

[zoutzakje]

I did read the entire wiki before I started my first real campaign, but I had forgotten most of it right away. When I encountered my first AI eye, I vaguely remembered reading about it on the wiki. But it wasn't until hundreds of ships starting to pour out of it that I learned and remembered to be carefull whenever I would encounter another one. Discovering things like that and making mistakes like that is what has made this game extremely fun for me.
I see every planet I want to take as a puzzle. And I have to solve the (always different) puzzle before the planet is mine. Are there any AI eyes, alarms, raid engines or counter guard posts? how many guard posts are there? enemy starships? ship types? are important targets close or far away from the wormhole?
defending your planets is another puzzle on it's own. Where to place my turrets? should I leave a part of my fleet here? military or logic command station? only got 3k knowledge, higher mark turret or higher mark fighter for example? 2 of your planets that are far away from eachother are about to get hit by a powerfull wave. turrets won't hold. your own current fleet is just big enough to handle one of the waves. split up and hope for the best? save one planet and sacrifice the other? It's decisions like these that will always keep me intrigued with AI war
I have played AI war over 150 hours and there are still so many things that can get me by surprise and things I don't even know about. I find it challenging and fun. You need to lose a few times before you can win (I know I did). Which made me really happy and gave me an adrenaline rush when I finally managed to beat neinzul viral enthousiast and technologist raider on 7.3/7
I don't think the complexity of this game is scaring people off. I think it's attracting people. It's what made me a fan.