Another month gone (tempus fugit – it feels more like a week) and it’s once again time for an info update. I’ll begin with practical news – new tiers have been added to our website, so if you want to support us, you can now pledge up to King level and get a beautiful handmade woodblock print, a (real) silver coin, a t-shirt, an action figure and other goodies. King is limited and there are only about 250 pieces left.
So now back to telling you about how the development of Deliverance is progressing. Although it looks from the previous blogs that the work is going without a hitch, the time has come to cool down a little and look at all the stuff that is not going so great and that we’re seriously struggling with. Not that I want to make us look like incompetent amateurs and lead you to a state of despair, but I think you will be interested to know about all the things we come up against.
As you may have gathered from the previous posts, Warhorse has doubled in size in the last two months from less than 30 people to 60 of us. The new recruits came not only from our own little country, but also from the US (even from Bethesda), Poland and Sweden. So theoretically we can now do twice as much work. The problem, though, is that work on a game is based on the design document. Naturally, our original plan was to write the design as we went along. According to the design, the core features were supposed to be designed first and as the designs gradually expanded we would take on more people as needed.
But our situation in the last, quite dramatic year drew a stroke through our budget. Since I was looking for money and shooting the Kickstarter video instead of designing, our design document has a few, quite significant gaps and even though I am now far from alone on the job (there are eight of us now), it is only coming together as a whole very slowly. The new people have to be trained, we all have to get on the same page, write the design in the same way, set up a system of work and define patterns of how we will write so that other people apart from us will be able to find their way in it, and all of that is demanding. Especially when, like me, you have to roll in front of you a massive boulder of backlog stuff.
Got a cloning device?
Don’t get me wrong. Our design runs to several hundred pages - we don't pull the game out of thin air. Most of the features are described down to the minutest detail. Only then a situation come along where you are desperately trying to write the last few missing, but quite important features for the programmers, the designers meanwhile are working on lacking craft mechanisms and in the middle of it all ten new graphic and concept designers are asking for assignments. But to assign work to the graphic guys, you first have to read and comment the crafting design from the designers, which after two weeks of work by six people “surprisingly” runs to a hundred pages, and that you cannot read in five minutes.
So you make an agreement with the graphic artists that instead of creating assets for minigames they should first design situation plans and white boxes of the villages on the map and then start on the crafting next week after you've done your review. During the course of the revision, however, you discover that some of the designers haven’t quite gotten the idea of how the crafting should look, that two very similar activities from two different designers have completely different controls, and you will have to go over it with them, redo it and add some stuff that you didn't think of when conceiving the crafting, but turn out to be quite fundamental obstacles to its functionality.
So in the end it turns out that the graphic guys have to wait a week longer and the designing of features and handling the backlog will have to wait, too. The Boulder of Sisyphus has rolled back a bit. And then when you've finished all the crafting and show it to the programmers, they throw their hands up and tell you it can’t be done like that and they’re not going to waste two months of work on some nonsense like cooking and we should go and simplify it. So the graphic designers...
Along comes anarchy, which in the majority of game studios is on the daily agenda and to some extent is inevitable in something as complicated as game development, but which I honestly hate. Especially when it is I who am the main cause of it, and the fact that our woes of last year had a lot to do with it and everything would be different if things had gone according to plan doesn’t change anything. What’s done is done. At our regular leader sessions the heads of the individual departments complain that I ignore them and they feel like the ship is tossing in the waves without a helmsman.
So for the last month we have been gradually establishing order and trying to get into a routine. We updated our roadmap. We started fundamentally reworking the planning system and since we use agile planning, we even had a guy here from Hansoft to help us set up the right processes. Even that did not go ahead without shouting matches, because even though agile planning is great for programmers, planning design and its implementation in it is something like writing the script of a big-budget movie while shooting. It happens sometimes, but it’s not OK. Here, too, we had to work out complicated compromises and hopefully we've done that. Only now we have to update our entire backlog. And my Boulder of Sisyphus rolls back another little bit…
Don’t count your chickens
There’s one more thing that is bothering me quite a bit just now - I have to constantly keep my eyes on the ball. The problem with Kickstarter projects is that you are promising something that’s not ready yet and you are making that promise at a time when even with the best planning in the world you can never be quite sure that everything you’re planning and everything you would like to have in the game will succeed in happening. When you’re planning a game of such massive dimensions, even with a crystal ball you can’t have 100% certainty.
A game on the scale of Deliverance has thousands of graphic assets and animations, hundreds of features that have to be programmed and thousands of lines of script, and all these incomplete things influence each other. To assess at the beginning of development how long some asset will take to program and how easy it will be to program some feature is simply impossible and so there’s no choice but to estimate, refine the estimates during the course of development and if they don’t add up, then make cuts or extend the development and make it more costly. Obviously, from previous experience your estimation gets more precise, but it's still an estimate and usually at a time when the person who's going to do the job in question isn't even working in the firm yet and the technology you want to use for it has yet to be programmed.
While in the case of a commonplace game you go and peddle your wares a few months before publication, when all the chaos is pretty much behind you and it’s clear to you what you will succeed in implementing and what you can still get done in the six months left to publication, in the case of Kickstarter the most you can do is show a prototype and your plans. Of course it’s not written anywhere those plans will succeed and experience tells you that all too often they don't succeed. Just look at stuff like Broken Age, Wasteland, Divinity and Star Citizen and their original estimated schedules.
By keeping my eyes on the ball I mean not getting dazzled by your own awesomeness and coming up with more and more new ingenious features before you realize that to implement them you would need a five times bigger team and budget and double the time. Then comes the cold shower, the crossing out and the annoyed fans who didn't get what you promised them. The more new people joined our team, the more ingenious ideas there were about how to improve just about everything or add some awesome new thing.
At the same time, the basis and main features alone are a big enough mouthful for a team of our size. It’s up to me, then, to curb the over-optimism and ban the addition of new stuff. Sometimes it’s a real battle (with myself). When you know something would really improve the game, you can hardly try to think up a worse way of doing it to make it easier, especially when the people who would be working on the feature in question are enthusiastic about it and try to persuade you that there’s plenty of time for it. There isn't.
Just today we were talking about how cool it would be if the player could have the job of innkeeper. He would simply take over from the NPC innkeeper, the other NPCs would give him orders and he would bring them beer. We already have it all in the game, so why not add that? Well, maybe because the NPC innkeeper has precisely defined places to stand when he puts the beers on the tables, while the player can come from anywhere and it would look ugly if we didn't have extra animation (extra work) for it, so we’d have to deal with all sorts of new situations. Like for example if the player ignored an order, which the innkeeper never does (extra work), how the player pulls the beers at all (GUI, animation and extra work) and lots of other things, and so we immediately dropped the subject - to the great chagrin of the designers and scripters who came up with the idea.
Right now we are in a situation where we have most of the stuff written, recruited the people and now we’ll see whether we can get the written stuff implemented as fast as we thought we could. And if we don't get it done fast enough, we'd have to simplify or move to the next act, and you probably wouldn't like it...
That’s the big drawback with presenting a project too soon.
What to do?
I wouldn't want it to look like I’m just crying here over my own incompetence, so I better share with you what we intend to do about it. The design, which so far has been the biggest stumbling block, is finally getting to a desirable stage. The stacks of notes I've accumulated over the last few months are gradually getting integrated into the design. Hopefully we will finish the remaining unwritten of the design this month. It seems we may have finally sorted out our writing system. It might seem like a trivial task, but drafting an open world RPG is really a lot more complicated than writing a linear script. Multiple NPCs have the same dialogues, the quests are composed of smaller fragments and roles and all of that has somehow top be processed symmetrically so everyone is on the same page and nothing is being unnecessarily done twice over.
We also scripted dozens of events for the world and activities that the player will be able to do outside of the quests. As soon as we’re finished with all of that we will get stuck into the final versions of the quests, the scripters will start working on them and the graphic artists will compose the world according to their requirements.
So keep your fingers crossed for us and I sincerely hope that the list of things we planned that will not end up in the game will be as short as possible :-)
Dan Vávra, Creative Director