We meet again, respected ally of the 'At The Gates' tribe!
A couple months ago I hinted at the possibility of some big changes. Well, the big news today in AtG-land is that said "possibility" has turned into reality! That means some exciting new features to talk about - but before getting into the nitty-gritty I think it's best to explain why we even have "big changes" to talk about.
Iterative Design - Not Just a Buzzword!
I'm sure some of you are thinking "What do you mean 'big changes'? Wasn't the game supposed to be done by now? Has AtG succumbed to feature creep? Has Conifer run out of money? Do you guys have any idea what you're doing?"
Given the state of Kickstarter these days I begrudge no one for having perfectly-justified concerns of this sort (hell, I'm in the same boat with quite a few still-unreleased projects I've been looking forward to!). Thankfully, I can state with zero reservations whatsoever that AtG is in great shape. There are no gaping holes in the gameplay that may or may not ever get filled, nor dark clouds portending a studio closure looming over the horizon. The game is fun, all features are at least roughed in and we still have plenty of money (mmm, ramen...).
Make no mistake, we're going to overshoot the projected release date I came up with back in late 2012 by a pretty healthy margin, but I've never by shy about the fact that our one and only priority is delivering a great game - regardless of how long that takes. I know I sound like a broken record here, but that truly is Conifer's "mission statement". No one remembers when a game is late, but no one forgets when a game is bad!
Okay, okay, let's all assume that AtG is in fact as amazing as I say - why are we making "big changes"? And how do we know the game actually is in good shape? The answer to both of these questions is simple: external feedback.
As one might expect from such a mature and supportive community, a number of amazing playtesters have stepped forward as huge contributors to AtG's development. Not only have these individuals provided great insight and suggestions, but they've also provided honest assessments about the state of the game. I really do appreciate constructive criticism, and the AtG Test Group has certainly delivered on that front.
A few months ago and back before the "big changes" much of the feedback we were getting could be summed up as: "The game is good... but it feels like something is missing." After journeying to a mountaintop and meditating in raging blizzards for a couple weeks I returned to my desk having come to the conclusion that they were right.
AtG was kind of like eating a candy bar when you're so hungry your stomach is growling. The first bite is great, but a half hour later you're still not really satisfied (sorry Snickers commercial, we're going to have to agree to disagree). Exploring the map, dealing with the ever-changing seasons and migrating was fun, sure, but what was it all for?
If a game is to lure you back to play again and again you need to be able to achieve something, to earn trophies you can point to and say "Look what I did!" AtG 'v1' was a game with several cool mechanics which tested how well you could keep your head above water, but little else.
There's nothing wrong with that if your goal is to create a simple $15 indie game, but we're aiming much higher with AtG. It was clear that for the game to really, truly be one of the best strategy games ever it needed something... more. We're not just a few peripheral features bolted onto the existing chassis, but a full rebuild. A whole new center of gravity. Small tweaks here and there can work when your pacing or balance is off, but when your entire game feels hollow you have no choice but to go back to the drawing board and rethink your core vision.
What made this particularly challenging (and necessary) was that instilling a meaningful sense of progress in a game about tribes which never stay in one place for long... ain't exactly easy. The main reason why information about Germanic tribes of this era is so scarce is that they didn't leave archaeologists much with which to reconstruct their societies. In a traditional 4X game all of your achievements are laid across the map. Cities, wonders, buildings, roads - they're all placed one-by-one by your own hand. That was not an option in AtG.
And so we had our million-dollar question: what do we replace all of that with?
Power to the People
The answer we came up with? People. Instead of developing cities and structures you would be developing your followers.
While this approach seems kind of obvious in hindsight it was tough to see at the time, in large part because we're entering territory very few strategy games have ever set foot in. Two notable exceptions are Crusader Kings 2 and King of Dragon Pass, which, of course, both happen to be among the genre's most revered titles. With such illustrious company I was feeling pretty good about what this new direction might do for AtG, but how would it work in terms of actual gameplay mechanics?
Step 1 was fundamentally reconstructing the game's core vision and 'holistically' integrating this new concept into the new one. AtG's original themes of migration, dealing with a hostile environment and overcoming hardship would still have key roles to play, but joining them would be something completely new: "Clans", each with a unique name, personality, talents and desires.
Gone was the dry mechanic where a settlement's population stat would tick up turn after turn, destined to be fodder for the future's generic, interchangeable playing pieces. No, Clans would be actual characters living in a place safe from the whims of the player's godlike mouse cursor. A gentle Clan with an agrarian leaning is unlikely to be too pleased if trained as front-line warriors. But hey, if the Huns are coming you can still force them into service. Just make no mistake, there will be consequences.
With this new people-centric focus I decided to lower the number of settlements you control from a max of ~5 down to one. Yep, one. That's it. Ever. Your area of influence can still be expanded by other means, but AtG's economic engine has now been consolidated down to a single centralized system (SSI's Imperialism is a good example of how this can work).
To be fair, owning multiple settlements had always felt a bit odd in a game like this, where there's never been any way to improve or customize them. I briefly considered changing this but shelved the idea in short order, as it was clear that upgrade-able structures are fundamentally opposed to AtG's theme. The big design shift Clans represented gave me the excuse I needed to finally cut the cord.
Reducing the number of settlements doesn't mean the game is any simpler though - in fact, I'd say the opposite is true. Clans arrive in your lone settlement and can then either be trained in "settled" Professions or sent off to harvest resources, explore, fight, etc. Directing their careers and guiding their stories provides a massive amount of new gameplay that simply didn't exist before.
Alright, Clans might seem like an interesting idea, but how would they actually provide a sense of progress to a game sorely in need of it?
Professions & Pacing
AtG's Clans can have desires, become unhappy, get into feuds and more, but for now we'll focus on mechanics. Clan development is represented in two primary ways: the Professions they're trained in, and how good each is at doing their job.
Professions are the replacement for the distinct, unchangeable 'Unit Types' common in other 4X titles. In those you might build a Scout in one of your cities, but in AtG you train Clan Adelhard as Scouts - and should the situation change you can always send them home to be retrained as something else. While it's possible for a Clan to completely switch gears, doing so means sacrificing the experience gained in the old discipline and starting over from scratch. In the early game this is no big deal, but in the final few years you'll have some tough decisions to make.
For Professions to offer a truly meaningful avenue of development we needed to have either a lot of them or a way to enhance them. After all, if you only ever retrain a Clan once or twice it really won't feel like you've made much progress! In the end I opted for the 'breadth' approach of having a large roster of Professions, as switching between them already strings and I didn't want players to also lose whatever Profession-specific upgrades they might have invested in.
The question of how players would unlock all of these Professions was a tough one though, and this occupied the team's mental energies for several weeks. For a progression system of any kind to be satisfying your pacing must be nearly perfect; hand goodies out too quickly and you dilute the entire system, hand them out too slowly and your game turns into a frustrating grind devoid of interesting decisions.
One of the simplest ways to model progress in a game is a basic 'prerequisite tree'. If this were utilized in AtG this would mean to train a Clan as Weaponsmiths they'd already need to be Blacksmiths, which in turn could only be trained from Laborers, and so on.
The problem with this approach is that when you want Weaponsmiths to make weapons for you what you want is, you know, Weaponsmiths making weapons. This might sound like an obvious and meaningless statement, but it's often the most stupid simple concepts you lose sight of when wading through the waist-deep swamp that is game design. What purpose do prereqs serve? To slow players down, or at the very least gate them in some way. Follow this to its logical conclusion and you realize that Professions like the Blacksmith are little more than 'speed bumps' designed to slow how quickly you can get between what you have and what you actually want.
Worse, the deeper you make your tree the more players run into this. Over the course of an entire game players could be forced to hurdle speed bump Professions dozens or even hundreds of times. Instead of players getting excited about training a Clan in a brand-new Profession as we'd hope, they simply sigh, shrug and queue up yet another Blacksmith. It doesn't take a professional to tell you that this ain't good game design!
Don't get me wrong, the venerable prereq tree certainly can and does work well in many other situations. It's simply a bad fit for systems which require players to make parallel decisions or when it's possible to backtrack (both of which are the case with AtG's Professions).
The second idea we considered was having advanced Professions require advanced resources. In the early game you won't have access to coal... without which you can't make steel... without which you can't train Armorsmiths to make armor... without which you can't train Heavy Infantry. This sounds good and makes sense in theory, but it fell apart quickly once we actually tried it out in-game.
The issue here is that you just don't know what resources are going to be nearby. Sure, we could spread them more evenly across the map, but that dilutes their importance. The whole point of having resources like coal at all is gating access to cool stuff that everyone covets but can't have. Making coal a vital link for a large number of important Professions is basically the same as funneling players into situations where there simply may not be any real decisions to make - if you have coal you train Clans in those Professions, otherwise you might as well pretend they don't even exist. Yawn.
Our next stab at solving the Professions Pacing Puzzle was requiring Clans to have a certain amount of experience before it was possible to train them in high-level Professions. We started with seven different 'skills' experience could be gained in, but this became unwieldy once you had more than a handful of followers. "Okay, Clan Raimond is level 2 in Construction. Oh yeah, didn't they also have some experience in Learning? Maybe I should save them to be a Surveyor instead. Err, wait... am I thinking of Clan Adelhard? Hmmm, I'd better go back and check for an eighth time..."
Trust me, that's not an exaggeration! Even so, the core concept was sound and work keeping in one form or another. In the end we streamlined the design a bit: Clans now have a single 'discipline' which they accumulated experience in, and (as I mentioned above) although this can be switched actually doing so means starting over from the beginning.
This was much, much more promising than our earlier attempts, but as often is the case in game design we were derailed by an unintended side-effect: it suddenly became very difficult to adapt. One of my core design tenets is that players should be encouraged and sometimes even forced to adapt to changing circumstances, so this drawback was no joke. To address it I decided to bring back an old friend I'd said goodbye to and never expected to see again...
Growing a Backbone
4X is one of several sub-genres of what we call "sandbox" games, where the basic idea is that the flow and pacing is driven not by developers but by the players themselves. This provides an unrivaled sense of freedom, enhances the thrill of exploration and adds incredible replayability - but there is a cost.
Topping the list is that sandbox games are, to put it bluntly, really hard to make! As a developer you have to trust that your abstract, conceptual rules will hold up and keep the game on-track when mixed with the completely unpredictable behavior and tendencies of your players. A good analogy is how driving a car yourself differs from writing the unbelievably-complex AI logic for a car which can drive itself. While I'd say making games isn't quite as challenging (!), things rarely go quite as planned.
AtG had major pacing problems, and my attempts to fix them 'cleanly' by using existing systems driven by players and randomness had failed. It was time to roll up the sleeves and make sure the pacing was right.
One of our playtesters wisely noted that every 4X under the sun has a self-contained research system/Tech tree, and that this isn't coincidental. Research provides our genre with a pacing 'backbone' that, having thought a lot about it lately, I honestly think may have a good substitute. Unit and structure-based prereq trees can work well in 30-minute RTS matches, but in a 4X you're either going to burn through them in no time or run into a glut of speed bumps. So how is research special?
The fact that it's self-contained is the key. The rate you acquire Techs can be completely independent from whatever resources you might or might not have been lucky enough to find nearby, or structures that you may or may not have bee-lined for (or forgotten about!). As a designer I can very easily set a couple numbers in XML and know for certain that players will get a new Tech no fewer than every 6 turns, but also no more than every 12. I can also dramatically increase the cost of Techs playtesters have found to be particularly powerful, or lay out the tree in a different way to ensure that there's literally no way to get them before turn 150.
Another, less obvious advantage of the traditional 4X research system is that it's also 'self-propelled'. No matter what, players are always studying something, always making at least a little progress regardless of whatever else they might be doing. This is important, and not the case in a similar system where you instead purchase new upgrades with money.
Giving players that kind of full, godlike control over pacing means some will inevitably fixate on unlocks even as the rest of their empire falls into ruin (*raises hand*), while at the opposite extreme other players will neglect them completely. Neither of these is necessarily a problem when you're talking about a non-essential gameplay system, but they absolutely cripple one that provides a game's pacing backbone. Just think about what it would be like to play Civ for 500 turns and never leave the stone age!
It's important for us to remember that limits are a big reason why games are fun. There are times when being a good designer means grabbing the wheel from your players and making sure they don't inadvertently careen off the road!
Now that AtG's core gameplay is firmly in place from top to bottom my focus for the next several months will be, as promised in the last update, the AI and diplomacy. After that we're talking all polish, all the time. The first part of this lengthy phase of development will be wheeling back around and making another pass on the game's Professions, Techs, etc. What we have in right now is fun, but very, very rough, and it will take some serious iteration before all the prereqs, bonuses, costs and more are in a ship-able state.
I know this update was completely devoid of pretty pictures, but the plan is for the next one to be dedicated to AtG's new art style. Not gonna lie, it's got even me pretty excited. I won't spoil the surprise early though, as we're just about done with it.
Alright, well, I think it's about time to put a bow on this one. When I started writing this post my intention was for it to be a short one, but... well... here we are. I suppose I should know better by now and learn to love the bomb! (Maybe it's time I finally bit the bullet and found someone else to help me write these so that I can spend all that time programming instead!)
We opened with me praising our awesome Test Group, and we'll close things out with the reverse, just to raise the excitement level a notch or two. These are excerpts from the two most recent 'First Impressions' playtest reports, both written after the big redesign described above:
"The game is very interesting. Right off the bat... the combination of knowledge/terrain/clans gets the gears in the head spinning. It feels well thought out, with a nice balance between options. There was a good 'one more turn' vibe going. Enough so that my notes from the first playsession are very sparse! You've done a great job crafting a fun game with depth."
"Wow, it's 4am now. After my initial eight-hour play session, I'm extremely impressed by the implementation so far. The game feels like the rough-cut of a precious gem. I can already tell this is going to be one of my favorite games of all time."
Thanks again for your support and patience, everyone. Like you, we can't wait until AtG reaches that incredible potential our playtesters have already gotten a taste of and we can officially call it "done". While that moment may still be far ahead of us it gets closer and closer every day!