Long time coming, but here’s Part II of my long winded post-mortem of The Last Stand: Union City.

---

Characters

I’ve made characters the same way since The Last Stand. To put it simply, they’re 2D puppets made of MovieClips for each part. When I was first confronted with making a bunch of zombies, it seemed like the most obvious choice. I should note thought that I’m a bit of a hack and there’s probably other more experienced devs out there screaming at their monitors right now.

It worked for me, remembering that this was an AS2 project and I didn’t have the bones system or IK at my disposal.

The benefit of building characters this way is that you can animate the “dummy” once and then have all your clothing pieces as frames within the individual piece movieclips. It also has the huge plus of being super simple to animate, provided you’re ok with a bit of frame by frame action.

---

Let the Combat Begin!

For the combat in The Last Stand: Union City I wanted to retain the feel and satisfaction from the previous games. The biggest difference between the previous games and this incarnation was immediately obvious, without the barrier between the zombies and the player, the player is always at risk. Even though the barrier was gone, things stayed much the same. The zombies now attacked the player directly rather than bashing at the wall, in effect the player became the new barrier. Same thing, different bucket in the end.

The combat only became a real issue when it came to blending it with the character levelling / progression. The formulas started getting a bit crazy in places as all of the following needed to be taken into account in some cases.

- Recoil (size of cursor)
- Weapon accuracy
- Precision attribute
- Luck attribute
- Weapon skill
- Stance
- Move state
- Jump state

All of those factors came into the equation of just the direction of the bullet. Past that it would calculate the damage and chance of missing with similarly complicated information. In the end I think I struck a decent balance between player skill and the numbers. My biggest fear was that I’d end up with something like Deus Ex (the original) where I felt like I was at the mercy of the numbers.

---

Building Union City

Grand plans. This best describes what I’d intended for the “world” of Union City. Early on I toyed with the idea of creating rooms dynamically and randomly through code, a lot of technical issues and the difficulty in making things look “right” put a stop to that. I also considered building an external editor to create the scenes with, that too ended up being more effort than it would have been worth.

In the end, I decided to create all of my environments internally within Flash. With the rooms /environments contained within MovieClips, I could quickly place and arrange scenes from a handful of props. It also allowed me to re-use props in different ways by scaling and tinting them to look like other objects.

Rooms were created from wall and floor pieces, with props such as windows, doors, containers placed over the top. That whole container was then converted to bitmap at runtime to make everything run smoother. For all of the searchable areas, doors, lights etc, I just used named movieclips. When creating the world, the engine just goes over everything in the scene, figures out what each object is and treats it accordingly.

---

Until next time where I’ll cover the loot side of things. Let me know in the comments if there’s anything else you guys want to hear about.

New version of The Last Stand: Union City is up on Armor Games now.  To get the new version, clear your browser cache. Be careful though, do NOT delete your cookies or plug-in data as that will delete your save games.

This version has a bunch of changes, including a fix for the grey / white / black screen issues people were getting. Turns out it was the encryption I was using in combination with some of the newer versions of Flash. Along with that there’s some other fixes which are listed below.

---

TLS:UC – v1.3 Changelist

Additions
- Added explanation for  “Allow storage of local files” dialog box
- Shove help
- Sprint help
- Added a check when trying to delete unique items (achievement rewards)
- Added some clearer quest descriptions

Fixes
- Fixed undefined vending machines in Uptown and Stadium
- Made the safehouse in the Whistler’s Forest township a safehouse
- Fixed stamina regeneration from being 0 when starving
- Adjusted lock picking achievements to make the do-able now
- Changed requirements for Brains Beats Brawn and Brawn Beats Brain achievements
- Stopped character from saving companions while dead
- Fixed Docks spawning you on the wrong side of the wall
- Increased quality on character creation screens

Just to get it out of the way up front, “on rock and roll”.  The following is a kind of post-mortem on the development of  The Last Stand: Union City and what I’ve learned along the way. Hopefully some of you who’ve played the game or followed the development will find it marginally interesting. If you haven’t played it yet, go now!

---

Starting Out – The Documentation Days

Way back in the ancient times of 2009 I’d just wrapped up Warfare 1944. The game was running well and was a relatively short development cycle of about 4-5 months. I was feeling more confident in my programming skills and had a backlog of ideas of where to take The Last Stand in it’s inevitable second sequel. So I dove in, alone, to what would be the longest 18 months of my life.

I knew that I wanted to make  a bigger Last Stand game, something with more of EVERYTHING. Weapons, locations, characters, zombies, survivors. It was basically a bigger is better scenario as far as I could see. Around that time I’d been playing entirely too much Fallout 3, as most of you have astutely noted from the references in the game, but I’d also been playing Torchlight and Titan Quest a fair bit as well. With three great Action RPG’s swimming around in my head, I announced to Armor Games that I wanted to build my “Zombie RPG”.

I had the “world” part of my RPG already and was amazed at how much people had projected their own stories onto the previous Last Stand games. Forum members were writing choose your own adventure style stories, people were making videos, writing their own stories. It was in people’s heads, it wasn’t something I’d intended. I just wanted to make a zombie shooter, but it seemed like fans wanted a story from me.

So in November 2009, I started writing documentation. Read the very short original design document for an idea of where my head was at.  I find documentation invaluable, not because I re-read it and check that I’m sticking to plan very often, but just to force me to get my ideas out and actually think them through. If you have to write down how to solve a problem that you think you know the answer to, you’re inevitably going to be forced to explain your workings. Aspiring game devs, if you’ve got a game idea, write it up. It will do nothing but improve your game design skills.

Around that time I also started prototyping the character creator, the skills and attributes was something that I didn’t have 100% figured out until the end of the project so I’m glad I started on them this early. Although some of the skills ended up being dropped or changed, around 80% of them were laid out here and stayed intact.

Character Creator Prototype

---

Decisions, Decisions

Early on I played with the idea of making the game top down and even made a prototype (which I subsequently lost) to try it out. It was interesting and allowed for a more movement related strategy to the game. It opened up a whole lot of issues to do with path finding for enemies though that I wasn’t capable of tackling at the time. The other major problem with it was that it took away a lot of the “cool” factor of The Last Stand series; the top down view made it incredibly hard to do nice looking / easy to read animations.  The profile view allowed for weapon and zombie animations that were more impactful. Firing the M82 side on just looks cool as the recoil makes it’s barrel fly up into the air, shells spew out and fall to the ground, zombie death animations just look better, all thanks to the 2D perspective.

In the end, I decided to stick with a 2D profile view. The choice then was to give it depth or not. There was definite precedence to, with the previous two games having it. Of course, it created a slew of issues as well. In terms of programming it needed some more complex path finding and collision code. Each scene would have needed a collision map, which would need it’s own separate editor. On top of that, it meant that I’d need actual 3D artwork for any objects that would sit in the mid-ground. Characters would need to be from a slight angle as well. It all just piled up to being too hard when I was foreseeing needing to multiply the process out 100 times.

Ultimately, I ended up with the 2D side view you see above. The world was made up of three planes with a background, mid-ground and foreground. This allowed me to quite simply create backgrounds and foreground objects from flat photographs (more on that in Part II). Also in that image is the early block out of the interface, an important step in figuring out where things belong. For most of those elements, I looked at the likes of Gears of War and a collection of first person shooters over the last 5 years or so.

In Part II - I’ll talk about the world building process, the beginning of the combat systems and getting the community involved.