A pre-post mortem on a seriously low budget title's final push to release

A pre-post mortem on a seriously low budget title's final push to release
It's August 22nd, 2018. We're about 2 months from releasing Blood, Sweat, Gold. A game I've been working on, in some form for years.

As the deadline approaches, the stress does as well.
The release isn't going to be ideal, there will be bugs, the full range of combat scenarios isn't done. So many things need polish. There's been no marketing at all. We're late on play testing. There are still game breaking bugs.

But underneath that rough exterior, there's a gem of a game that just needs love, polish and time.
Let me catch you up on the game. Around 5 months ago, @BIGTIMEMASTER joined me working on BSG. This offer to help came at a particularly low point in production and re-energized the effort. From this point on, the level of visual polish in the game skyrocketed as sketchy low poly asset store character models were replaced by fantastic hand made art.

Generic, dull characters could be replaced by characters that really sold the gritty, dirty theme. Months ago, when I saw @BIGTIMEMASTER's first peasant, I knew we had something.

The first inkling of goodness: breaking all RPG tropes for player characters
​

We had a character that didn't fit RPG tropes. He was a mean ass, poor, starving peasant. For the first time, BSG had character art that was good enough to really highlight in game. I added a proper model to what was once an empty inventory screen.

First iteration of revised inventory screen

Seeing this kind of progress was a bolt of energy motivating two major changes to the game systems. One, a move from analog control to point and click.​

Point and Click Movement
The main motivation for this new input system came from feedback I got from video. Too many people didn't understand that the game was a Turn Based Strategy RPG, the fluid animation and analog control simply confused viewers and sent the wrong message in terms of genre.

People kept thinking that the game was real time, and found it weird that people were frozen.

Additionally, analog movement was so slow that anything involving moving 4+ characters really dragged down the experience.

Early point and click movement test

The next major improvement involved a re-imagining of how the details of turn based combat might work. I began a rough prototype first, as proof of concept.

The Combat Prototype

​

The core idea was to transition from more traditional attacks and special moves, to a more fluid system. Something almost akin to a fighting game. @BIGTIMEMASTER and I played this prototype repeatedly and I collected feedback from some friends.

This new system was at heart a collaborative effort, and is, I believe a small but very real contribution to the turn based genre. Simply: it feels great and it adds some interesting depth to the process of attacking.

The idea itself is very simple, once you attack another guy with melee, we switch into a zoomed in view. Each keypress is attached to an attack type (each of which uses different amounts of stamina, etc). Animation freezes as each attack completes and you chain together a combo of sorts.

An early test of the new combat system

I love this combat system. It feels fantastic. I believe that it's good enough to potentially become a genre staple going forward, and that if BSG actually sells enough copies that other devs experience it, some will clone it.

Our work continued forward, updating the UI, updating the art. @BIGTIMEMASTER began working on clothing sets and material variations.

Our Army of Poor Starving Villagers


​

On the other side, we began to rework the archer combat system into something more fluid. Replacing simple hit changes with a more realistic deflection based targeting system.

Testing potential impacts using proper shot deflection for archers
​

Collaboration isn't easy
Changes to archer mechanics and the system in total was difficult, not only because of the work needed, but because we disagreed on the direction for archers.

Fundamentally we disagreed on how exactly Archers should play (and on what power level they should have). Every team butts heads now and then, but on a project where nobody is getting paid - these kinds of disagreements can really add a lot of strain. The weight of that strain was pretty significant and fell into play testing. Was I looking for potential exploits, or was I really looking for arguments in favor of my vision for archers? The lines might have gotten a bit blurry.

The combat mechanics as a whole became a bit more contentious. How much defense should a guy have, how desirable is realism, how exactly should the numbers work?

I really believe that these kinds of disagreements are healthy, and that arguing in good faith is probably the best way to avoid making huge mistakes, since both sides weigh disagreements. But good for the result or not, they take a toll and add stress.

Which brings us to the final stretch, the last two months
Crunch time. I'll be honest, the game as a whole is in an uncomfortable place. Too unpolished for this point in the timeline. Too many bugs, too many rough edges. But there is so much goodness, so much potential. It can really be a great game.

@BIGTIMEMASTER is finishing up the art, and putting together hair and beards that simply look fantastic. He's going to do some work on improving the environment. I'm going to be squashing bugs and trying to get things fully organized.

"Will it make enough sales to justify finishing it?"
Of all things, fear really is the mind killer, and when you have so much work at stake, and you're at such a risky point, it's incredibly hard to stay focused and motivated. You have hopes and dreams and fears all swirling in the back of your mind.

It's at these points when it's helpful to remind yourself that these things are natural, that anyone, anywhere would feel the same stress and same tension.

I believe in what we've accomplished. The game is at heart, a solid, fun game. But will we be able to polish up the loose ends in time to make a decent release? Will we be able to scrounge up enough media to sell copies?

Nobody will know for certain unless we push this beast onto the market and see what happens.

If nothing else, it'll be a hell of a portfolio piece.

 

The last mile is always twice as long as everything before it....

Great write up! It's been a long road since I started helping with this project. I've seen it come a long way. I could easily write a novel about what I've learned so far, but, you know, no reason to give away all the secrets.

When I started this project, I was somewhat experienced in 3d character creation (having began doing 3d less than a year ago) , but I was still very green. I really had no clue how to make modular characters at all. But, there's plenty of info out there if you are just willing to take the time to research it, and then of course you've got to do the work, make the mistakes, and manage your own expectations so that frustration doesn't eat you alive.

Having an experienced programmer who is able to quickly prototype things out is probably what made it possible for me to actually develop a modular character system that works. It's somewhat simple in theory, but the programmer side of things is not something that I could have done without at least a year to study just that. So the characters you see in action are very much a collaborative work -- especially considering the nuanced animations that have been carefully customized by @frosted to respond to hit direction, timing, etc.

As mentioned, the games in a rough state right now, which doesn't makes us feel so great being so close to our launch deadline -- however I think we are both very pleased and confident that the foundation of the game is something both unique and -- most importantly -- downright fun. I've never been a big fan of turn-based games, but the combat in this game really strikes a unique balance between real-time speed and responsiveness, but retains that deep, chess-like strategy you'd want from a TB game.

About the characters, they had to be developed as quickly as possible. I didn't have any experience to gauge how long things would take, so I determined to work as efficiently as possible and get a minimum quota delivered as fast as possible. This way, I could hopefully have spare time at the end to devote to unforeseen issues. All in all, this plan worked out pretty decently as I will finish the characters this week and have plenty of time to help polish up other parts of the project as much as possible before launch.

Although most of these characters were made in two days or less (from blockout to rigging...), I constantly used @frosted for feedback to make sure I was hitting the kind of aesthetic he had envisioned. Thankfully, I think we both have similar artistic taste, so pretty soon I was able to just confidently make characters however I thought best without any friction between us in those regards.

What we are hoping to achieve artistically is a game that does the opposite of what most other games do. We want to shed away all idealism, and give the player realistic, emaciated, ugly, desperate characters who you could really believe might cut a man's throat for a sack of coins. I have to say, making characters like this has been too much damn fun. Sometimes I am laughing for hours as I groom the filthiest, most awful hair styles while these beady eyes and rotten teeth stare at me from the computer screen.

Had I more time (any time at all really) to plan things out more, concept more, I have no doubt the artwork could be a lot better. However, feedback so far has been in line with what he had hoped it would be. People either say, "I've never seen nasty, grungy characters like this in a game. Awesome!" or they say, "Gross!" Either way, this makes me happy.

 

One issue that cannot easily be solved when you are undertaking a new journey is this:

You learn so much that the work you did at the beginning cannot hold up to the work you are doing later. In the image below, the center character is the initial peasant model, and immediate right is the same model but with updated textures I made just a few months later.

Whether you are doing the art or the programming, I think this is one of the unforeseen things that can bite you in the ass if you are doing such a large-scoped project. Luckily I had the time to update all of my characters, but it is a real issue beginners will have to contend with. Another good reason to keep the scope of your first games tiny. Or, at least, be sure you plan time to redo much of your initial work.

 

Cool thread guys, always interesting to hear about the difficult push to release! Good luck!

 

Congratulations to both of you! I gotta say, I am impressed! This looks a lot better than the last that I had seen from it. I think you've made all the right calls with easier to understand concept/controls, a unique visual identity and innovation in the combat mechanics!

Have you found proper music for it yet? This will be a big part in shaping the "feel and identity" of the game.
I wish I had time to offer some help there, but there's no way I'd get anything done in the coming months.

 

Good work guys. The environment looks pretty solid already. What terrain shaders and tools did you use?

 

Can you guys add a month or two before release where you just do light QA but mostly marketing and rest? Because if you don't respect the project enough to do that you will probably lack the energy to spot important opportunities or problems.

When you have finished it, you should gain some rest long before launch. Don't just launch it when done, because you will not be operating at the right level to launch.

 

to all, Nick has just left out of town for the next day I think. He can really answer most of the questions better than I can.

@Martin-H , about music, no I don't think we have anything more than what you have probably already heard. This is simply due to a lack of money/time to invest in that department. Of course, the power of a solid soundtrack is immeasurable -- I think both me and Nick agree on this, but without some money coming in from EA it's not something that we can do.

@AndersMalmgren , the environment I believe is a dated version of MapMagic, though I know Nick has done a lot of custom work to it. The assets themselves I believe are just speedtree and vegetation studio stuff. I agree that it doesn't look half bad as is, but I think it can still be improved considerably. It's a time investment I think will pay off, as just a few days of work will make for much prettier screenshots and demos.

So, in the next weeks, I'm going to be attempting to make some big time improvements. In short, we'll be condensing the playable area considerably in order to make very dense and realistic forest, and I'll be focusing on making an environment that both makes sense ecology, and also benefits ambush-tactics focused gameplay. At the same time, we want to minimize possible introduction of new bugs, so we may actually be simplifying the terrain in a way to make navmesh issues less of a thing.

@hippocoder , yes I was just discussing this with Nick. Ideally, we'd have at least a month to just sit on the project, do marketing, do playtesting, etc. However, personal obligations have made for a pretty strict deadline, and the project simply can't go on much longer without funding. I'm really keeping my fingers crossed that Nick gets the break he needs to give ourselves time to deliver the thing properly... but at some point it just has to get pushed out the door.

 

As Alex (BIGTIME) mentioned, the environment is procedural, all maps in the game are generated using map magic. Map Magic itself is well enough made to allow for a great deal of customization, so I extended it out from fairly generic stock functionality to pretty specific outputs.

This for example is a town and road network that we generate for the world map and is obviously custom work:

What makes map magic great is that you can literally replace all the guts and operators with custom ones, and just use his slick GUI for the node based design. There is some code there that is a bit questionable (the threading makes me cringe) but in terms of the core design he got it right, and he had the foresight to keep it highly extensible.

Example of the fully rendered world map:




In terms of the environment:

Vegetation Studio's grass rendering performance is unmatched and professional quality. There is no comparison between it and other asset store offerings. It is fully professional tech.

MicroSplat replaced RTP for the terrain shading. For years RTP was the only really full featured terrain shader on the market, MicroSplat and MegaSplat are the only ones that offered the same level of functionality with better performance and workflow. The author is a serious, skilled developer.

 

Cool, I have played around with both VS and Megasplat (Jason have also made Microsplat). Great tools.

 

I really recommend both. I have tons of respect for Jason's work. He's a brilliant guy and I am really picky about code.

If MegaSplat is overkill, MicroSplat will fit perfectly.

 

At some point I will look at using Jasons texture array implementation to reduce draw calls

 

Really suggest taking a quick look at MicroSplat (base pack is free) in terms of usability it's great. Unless you have really specific needs for a custom implementation it's hard to beat.

 

Megasplat has the ability to use texture arrays on normal mesh renderers, but maybe micro have to?

 

I might need to hire you guys to create a open map for us

 

If your goal is blending objects into ground - MicroSplat offers this functionality but I haven't used it and haven't seen the internals. I don't think this functionality is possible without rendering the texture array onto arbitrary mesh (but I can't be certain).

RTP had similar functionality but it was such a pain in the ass I decided the extra fidelity wasn't worth the hassle, so I never tried it with MicroSplat. That said, Booth probably figured out a way to do it that wasn't such a nightmare.

Lol, we offer training courses at low discount rate

 

I'll talk the wife into letting me buy this. You've been busting your butt for years on this.

 

What training do you offer?

 

I was kinda kidding, but honestly if you're interested in some sort of training - I'd be willing to walk you through the steps required to do just about anything outside of modelling. I know Alex (@BIGTIMEMASTER) wanted to do some tutorial videos on the process he uses for art.

In the not so far future, I'd like to put together a series of articles on many, many things I learned over the process of making this game. Both technical, and in terms of the real process of being a solo/duo developer working on something overly ambitious.

 

If you're just curious about anything generally, just post it here and I'll show you how we did it - in whatever level of detail you're interested in.

 

When will we see a longer gameplay video with audio?

 

I am putting a pin in your offer. I would honestly pay for training on modeling, skinning, rigging and handling character animation states in-game. I can't get my head around doing it well.

 

Good question, @Martin-H ,

I can't say for sure, hopefully by mid-next week. I'm about to deliver the final characters fbx to Nick today. Probably just as soon as we get them integrated and any bugs taken care of that arise, we'll make a video.

There was supposed to be some already, but we figured it would be better to show a fuller range of content rather than seeing the same two characters fighting each other -- along with some unfinished materials and such.


@Unknown33 , to learn that much stuff you'd need a mentor, which will cost big money. And you'd also need 6 months to a year of focused study. Could be double or triple that time, if you aren't seriously focused.

I'd recommend getting started with pluralsight or 3dmotive. Really cheap, tons of content. That's where I got started. The rest you have to scrounge and learn by trial and error.

 

Also, I see stuff like this from beginners all the time. They are eager to learn and create awesome content, but they have too many problems along the way and begin to feel like the only way they can learn is to have somebody else spoon feed them.

This is a major mistake. There is an absolutely foolproof way anybody can learn (even if you're an inbred hick from alabama public schools like me), in an efficient and thorough way. It's called the scientific method. You likely learned it in high school, but forgot all about it after the test. This is a shame, because the scientific method is mankinds best way for determining the right questions to ask and then finding solutions.

Seriously, if you brush up on the basic scientific method, and take that approach with a healthy dosage of patience, you could be making awesome art in a year or two if you've got plenty of time to sit down and focus daily.

 

You guys better get some rest though. Or I'll pretend to ban you from forums or something. This game is awesome. It deserves the best release

 

Seriously, thank you.

 

That reminds me - don't forget the MWU page, can take a bit of time to get it published etc...

Edit:

It used to be people could show off their games here: https://unity.com/madewith
But it looks like it's just curated now.

 

What's an MWU page?

Also, if you guys have any advice on how to approach the next steps - they're super welcome. Neither of us have ever done real PR or actual releases, so any insight on the next steps could be great.

 

@hippocoder

When you see us both dicking around on the forum, that is us resting! Procrastination is underrated.

 

I don't have any hands on experience either, but my 2 cents would be:

Personally I think a title that is easier to remember, harder to confuse with something else (this one is easy to confuse with "blood, sweat and tears", and if people missremember it they won't find the game through searches) and handles better in SEO terms (less generic terms that already have thousands of games with such words in the title) would be a good idea.

What games do you think would someone like, who likes your game? Find youtube reviews of small to mid tier youtubers who rated those games highly, make a list of their contact details, send them "press release" type emails to let them know in advance what and when is coming.

Have a "press kit" ready at release, that you can send to more traditional games media outlets. It should contain screenshots, videos, and text. Do as much of their work for them as you can, they should be able to assemble a proper news article about the release of the game (not a review) without having to actually play it. There's probably tons of advice for presskit best practices out there. Look here for "presskit()" : http://www.vlambeer.com/toolkit/

 

this looks amazing.
what tools were used to make the 3d models?

 

Maya LT -- blockout, UV's

Zbrush -- mostly just used for details on the faces. most of the outfits are just old-fashioned block modeling without custom normal maps.

Substance Painter -- textures. some of the outfit textures come from gametextures.com, though I heavily modified those. Everything else is just done by hand or with various mask that come with SP by default.

Marvelous Designer -- some of the clothes

Marmoset Toolbag -- all baking

google -- references

Because I had to make the characters so fast, the process involved as many shortcuts and cheats as I could figure out (nothing illegal, all content is my own). I think it's natural to want to make everything as detailed and perfect as possible, but I had to be really careful to only do just as much as necessary but no more. For instance, it would be easy to get carried away sculpting wrinkles and pores on the faces, but because we are using 1k maps and the camera is never so close, none of that level of detail would ever be seen. So banging out some fast prototypes and doing your look development in engine as early as possible is really important so you can identify what things will really look like in the final project and you don't get sucked into useless time sinks.

I think the processes I used would be useful for people who want to make basic character assets for their games but not necessarily try to become professional artist, so at some point in the future I'll make a tutorial suited for that crowd -- assuming I have time.

 

It looks great, I love the combat system and the way the camera angle changes, the pause is nice too before the damage is dealt...

I second what Hippo says around having a QA pass before release. Have at least one good QA pass and bugfix/improvement pass before releasing to the public.