Category Archives: Gameplay

Finishing up Julia

In the last weeks, we have made great progress in terms of finishing up Julia, and we’re currently in the homestretch to having her done (for now). This mean we are spending a lot of time on polishing and tweaking the framedata, hitboxes and animations. Once we are done with these final tweaks, we will finally start working on our second character, Ushah, and we couldn’t be more excited!

On the programming side, all the goals for the engine we had for this ‘Julia-phase’ are done, and our programmer is already gearing up for the next phase, where we will finally get rid of our rather dated and heavy engine.

The scripting for Julia is still being tweaked every day. Many pixels and lines of code are examined to make sure that all the moves have the proper spacing, as to not create extremely strong combos or braindead hitconfirms and to make sure that the jumps and hops jumps are perfectly spaced to set up that one awesome ambiguous cross-up.

On the artwork side, our animator has delivered a piece of Julia which portrays her in the style we like to see her in, a mixture between western comic style and a Disney-esque vibe.


Sadly, as much as this is a gorgeous drawing, having him as our main artwork guy as well as animator/spriter will spread him too thin, so we are in dire need of some new blood. Are you a talented artist in need of some portfolio work or just experience, drop an email at djono [at] team18k [dot] com. Do note, at the moment we are without funds, we hope that this is a temporary predicament, but at the moment we can’t offer any financial compensation.

Attentive followers may have noticed that Julia’s Break* uses the animation of Julia’s jumping HK as a grounded attack, we want to eventually create a completely new animation for this move. We’ve been brainstorming some different concepts, and Bob Sagat has drawn up some quick sketches for this move.


Julia now finally has fully drawn wakeup animation and flip animation. We have also tweaked here jump arcs and added a quickstand mechanic, all are shown in this video below.

In the coming weeks we will showcase some hitboxes and insights on why moves have certain properties, all explained by our framedata master Phoenix.

– Coal

*As we ourselves got confused by all those ‘shatters’ in the Shatter System we have chosen to rename the ‘Shatter Attack’ to Break.

Playtest at LLL.MBR’s Bootcamp

Last weekend we had a good opportunity to playtest our game at LLL.MBR’s bootcamp. More than a dozen people tried Shattered over 2 days and gave their input. While the game is still in a fairly bare bones state, the people trying the game weren’t hampered at all in exploring the game, developing inventive footsie games and quickly discovering new tech. People seemed to enjoy the game a lot and were eager to give suggestions and their opinions.

Seeing people enjoy our game is incredibly satisfying and one of the reasons why we are making this game, and why we will continue to push forward.

We were aware of the presence of some bugs in this build, and another few others were discovered, but overall the game is really starting to come together and become a truly playable experience.

We’ve of course made some recordings of the matches as well! In these videos you will see one of the most prominent bugs that we were aware of even before this testing session: Multiple fireballs can be on the screen at once. The players were quick to find out that this meant that throwing fireballs is very strong, as you will see in the videos below.

The first match is between Pr3y on the left and LLL.MBR on the right.


We also ran a small tournament with the guys present at the bootcamp. We’ve made some recordings of the finals between LLL.MBR (left) and EX Snorlax (right), some of the rounds of these finals can be seen in this video.

Shoutouts to Miku for managing to unintentionally crash our engine every 20 seconds.


We’ve received some questions about what we will be doing with the Netplay for Shattered. We’d like to give you guys some insight into to plans concerning online play.

Fighting games are one of the most unforgiving genres of games when it comes to online play. Even the slightest delay in your inputs can be detrimental to landing your combos and reacting in time to an opponents attack. This means there’s technically only one solution, and that is rollback netplay.

Rollback netplay ensures that there is zero delay for both players. Both players send their inputs simultaneously, and once your receive the input of the player several frames later, the move that was supposed to come out is retroactively placed into your game.

As there is no way to completely get rid of delay, this means that once the game state is corrected the move will have already begun, and you’ll miss the visual cue of some frames of the start up of a move. This is usually not a problem, when the delay is fairly low between two players, but it can, sometimes lead to the dreaded ‘rollbacks’. If you press a move that hits before the opponent received the data that says you pressed the button, it is possible that you see your move hit, and a few frames later suddenly find the opponent magically blocking it.

Some players find these rollbacks incredibly annoying, and for this reason prefer delay-based netplay over rollback netplay. But the magic of rollback netplay is that it can be both!  Rollback netplay can also let you choose to manually set an amount of frames of delay, to whatever number of frames that you still consider playable. This gives the netplay more time to figure out what is happening, and therefore increases the visual fidelity.

We have not decided on whether we will license a third party rollback netplay, or whether we will implement our own netplay. If the latter is the case, we luckily have access to excellent technical articles by Ponder (the creator of GGPO) and Mauve (the creator of Rollcaster) detailing the issues and technicalities of rollback netplay.


Throws are an essential part of fighting games. Throws are unblockable, and therefore play an essential role in breaking the defense of someone who is focused on blocking all incoming attacks. Without these unblockable techniques playing defensively would become too strong of a strategy.

Throws in Shattered are executed by pressing the two punch buttons simultaneously. Every character has a regular throw, which will be fairly quick (currently, we are prototyping with 3 frames start-up), some character also have special moves that function as throws. If a throw hits it will usually knock down the opponent, which allows you to continue pressure.

If you are being thrown, you can break the throw (but not special throws!) by pressing the two punch buttons, in a small window at the start of a throw. We have not yet decided how big this window will be, but we are currently prototyping with 12 frames, which can be reacted to if you are really on point. If this turns out to make throws too weak, we will make this window smaller.

Of course, characters are completely throw invincible during blockstun, and throws generally cannot be comboed into either. Also after hitstun and blockstun you will be invincible to throws for a few frames. This will make the strength of tick throws less dominant. We want throws to be used as a means to break the opponent’s defense, not as a really fast unblockable poke.

Finally, we would like to show you the throws working in a early build of the game, as you can see, the throws still use placeholder animations. Enjoy!

Julia Verne: Gameplay

In the past weeks, you have already had the chance to see some of the animations of Julia. Julia is a fighter that uses her bow to shoot bolts of energy, and uses it as a striking weapon to add extra punch to her attacks. In this post we will discuss her gameplay.

Julia is the most basic character of our cast, a jack-of-all-trades. She has normal and special moves that basically cover most angles of attack. This does not mean she has the perfect answer to everything, but she will not have an explicit weakness in a certain area. Julia fills the so-called ‘Shoto’-archetype from the classic fighting games. Although Julia is a fairly basic character, this does not mean she is a simplistic character. We intend to make her an interesting and deep character, with plenty of room for mastery.

Julia does not have long extended juggle combos. Instead, she mostly focuses on simple normal to special cancel combos. This makes Julia the ideal character for beginning players, and players who like to focus on the fundamentals of fighting games, rather than perfecting long difficult combos in training mode.

In previous videos, we have showed off some of Julia’s basic movement animations and special moves. As one would expect from a Shoto, she has a projectile, a flying uppercut and a horizontally moving kick attack. While special moves are the moves that make a character distinctive and recognisable, no character could function without its normal moves. The video below showcases the normal moves of Julia.

Gameplay design philosophy

We are making a ‘traditional’ 2D fighting game. Considering the long history and huge variety of 2D fighting games, that might mean different things to different people. This post will try to give an insight into what some of our design philosophies are and what kind of things motivate our choices.

One thing that we find very important is transparency. Fighting games can be complex, and the underlying mechanics can be very obscure. For example, some games have safeguards against infinite combos by making the opponent fall faster and faster as a combo progresses. These kinds of safeguards can make understanding which combos will work works and which won’t very unintuitive. We want our combo system to be transparent. We want the player to be able to figure out what will work and what won’t with minimal pains.

A second principle of our design is that the options are straightforward. We want to have a clear and elementary combo system. There are only a few general rules that influence combo possibilities. For example, light moves can usually chain into other light attacks and it is not possible to chain into heavy attacks. Some attacks can be cancelled into special moves while others cannot.

The combosystem will also have link combos. In some games links can be a source of frustration, because it is not always obvious without framedata, to see whether a move will link into another move. This is a property of links that cannot be completely avoided, but we have decided to give universal hitstun to moves that will aid in developing a feel for whether a move will link or not. For example, a heavy attack, regardless of the character, will do the same amount of hitstun.

When you try out a new character, you will quickly be able to get an idea what combos are possible, by simply trying out the different moves. This does not mean that you will be able to execute every combo instantly, but it does mean that you can start practicing combos that you know will work.

We want our game to have great freedom of movement. We want walkspeed to be high enough, so that you can use walking in and out of your opponent’s attack range with ease. This gives combat a dynamic flow, where one can use skillful movement to coax your opponent to press a button that he’ll regret pressing. Within this freedom of movement, we have decided that some control of your jumps is also important. Therefore the game both has a normal jump, and a short hop. Having two different jumps allows you to cover more angles of attacks dynamically, which ties into our freedom of movement. It also allows for more fast-paced rushdown based offense.