It’s been just over 4 weeks since we posted our initial public build of Deuce as part of the Facepunch Prototypes on Steam.  It’s great knowing that there are people out there playing and hopefully enjoying what we’ve shown so far.


Reading our initial batch of feedback has shown there is still some way to go in making the game feel responsive and fun. Feedback has pretty much fallen into two categories here; the way the player and ball moves, and general performance / frame rate.

I’ve made a few small changes to the way the shots are triggered.  One of these changes is that the player can now be pushed back a touch if you try to return a ball that is just in front of you.  This means you can return the ball right up to the point that it’s level with the player .  The hit animation plays at lightening speed but this change does seem to improve playability.  

It’s now also possible to break out of moves quicker if you want the player to run.  For example, the player will now start running the moment his feet hit the ground after a serve.  If you don’t want to run then the rest of the settle animation is played out.  I’ve added these extra transitions for exiting shots and smashes too and it seems to make a real difference, especially if you are a player that likes to run around the court a lot.

In early Deuce builds we had a lot of “auto running” where the game gave you a lot of assistance by running to the best position to return the ball.  Frankly, this turned out to be a bit rubbish… you could easily return shots but it barely felt like a game.  We replaced this system a while back for normal shots but up until last week it was still used for the smash.  Now the game will show a marker helping the player find the optimal place to smash a lobbed ball and the smash anim can be sped up to connect better with the ball.

I’ve also spent a chunk of time adding some optimizations and fixing bugs that would hamper people’s enjoyment of the game, especially on lower spec systems.  For example, there was a bug that rendered the game to a render target at a set size.  It was possible that this target could have been bigger than your actual screen resolution adding unnecessary overhead and slowdown.


Right now I’m probably the world’s best Deuce player and if I find the game hard to beat (without exploiting any known bugs) then it’s probably too difficult.  The first build we put out basically just had the AI returning the ball back to you.  I then started to make the AI attempt some form of strategy, it would play the ball back to you if the rally counter was low and then try a passing shot when there was a chance to maximise damage to the player.  This just made the AI difficult to beat and unfair…. Plus, losing all the time is no fun.  

I’ve started updating the AI to be a bit fairer.  The AI now has three components to determine how good they are:  reaction time, skill level and player type.  Lowering the reaction time means that the AI takes longer to decide how to respond to shots and the skill level decides how good the AI is at placing the ball.  The type is used to set whether the AI stays at the back of the court and is a baseline player or whether they are more of a serve and volley player.

I’m not the only guy working on Deuce so starting in this blog other team members will be giving their updates on the project.


Since the last Blog, I have started creating the 6th environment with the work-in-progress name of Temple Court. Temple Court is set in a vivid and wild forest, with ruined stones and mysterious icons throughout.

I thought I’d do a write-up on the thought process behind developing a new court, highlighting some decisions, revisions of those decisions and sometimes complete U-turns.

Here is the original design sheet for the environments. This was created back before any art had been done on Deuce, but we’ve stuck pretty close to it. One change was the introduction of the Dojo court as a test and training court, but the general sense of progression from bottom to top.

So taking Rural Env 03 into account, the general theme is Jungle, Temple, Overgrown with foliage. Creatures also, but art-wise they come later.

Next, comes thinking up a rough Court Hazard idea. This isn’t final, as it can always change at a later date (we’re still thinking which idea of a few we’ll use for the swamp court) but it can have a bearing on the court art. I thought that it might be nice to use some kind of iconographic symbols as it’s a temple, combined with jungle ruins and came up with an eclipse court hazard idea where, as the environment lighting moves into eclipse, the powerup meters are full for that duration. An eclipse has a sun and a moon, so they’d need to feature.

I came up with a mockup, again before any of the other art was created where I tried to demonstrate this idea:

The idea here being that as you’re playing, the sun and moon slowly travel across the sky and as they eclipse the lighting (and maybe textures?) change.

All good right? Should be pretty visually interesting and quite fun? There’s a problem with that. Since creating the other courts we’ve become aware of just how limiting the game camera has to be. You just do not have the height required in that concept to see the sky. You’ll never see the sun and the moon.

So time for some concept art. Howie came up with this as a first sketch. As you can see, you can’t see any sky at all. I love the closed feeling of this court, as most of the others are really quite open so you can see stuff past the court to the sides.

Here you can see a sun and moon icon that could be used in the court hazard. I can make a shadow of an eclipse pour over the court so hopefully you’d still get the effect. We did feel though that the court needed a bit of Deuce technology going on and maybe it wasn’t open enough at the back.

Much more interesting, it’s much more unique when viewed on the level select screen, and it has it’s own feel to it. We need to work out exactly how the court hazard works, but that won’t be too hard.

So now on to implementation. I broke down the court in to a number of tasks: Greybox, Ground surface, Stone stairs and props at the back, grass, roots and trees, backdrop.

Here you can see an early stone floor in, with some early grass. It’s got some very flat lighting with different colours.


After cutting in the court lines and experimenting with different colours.

Creating high poly assets. On Deuce we don’t have a lot of high poly work as the game is quite stylised, but certain things like the floor and the roots really needed this work to look their best. Most of the assets in Deuce are quite bespoke to make each court different so each main root was created individually, but I can reuse the trees in different positions when I put them in.


And that leaves me up to where I am right now. Adding all the roots and then it will be on to probably the backdrop before I start on stone work. Most things are untextured and everything’s very much a work in progress, so I’ll post updates on the next blog!




For the past few weeks I’ve been working on some new character concepts. Mainly, but not strictly, focusing on the more villainous rogues and scoundrels of the game!

As was mentioned in the last blog the world of Deuce is a place where anything goes so I’m trying to make sure that I am catering for that in the scope of the character designs.

Some of the ideas that I’ve come up with so far include; Gladiator droid Khan 2.0, Geisha Assassin Cobra Mei, Dragonslayer Morag The Bold, Martian descendent The Mesmerist, post apocalypse (he thinks!) scavenger Sergeant Prepper…….the list goes on…..horny Giant Dwarf Bang-Gor……there are more.

Here’s how some of them look. See if you can match the names to the faces!

The character that we have decided to develop next is Dréadora, The Queen of the Dead.

This is the concept and the idea for her special move.

One of the first things you may notice is that she is topless. This design decision wasn’t made for any provocative reason. In the case of this character I believe it reflects, and enforces, her strong, confident, couldn’t care less nature and made more sense to me than dressing her up for no real character led reason.

My hope is that most, if not some, will see the character in this way and not in a sensationalist view.

We shall see.

This is how she’s shaping up in ZBrush so far. Still early days but I think she will make a cool addition to the existing line-up.

The move from 2D to 3D is sometimes tricky, things that work and look good in a 2D drawing quite often don’t translate well into 3D, and of course once something is in 3D it needs to look good from all angles. I’m a firm believer that any concept needs to continue being developed once it gets into the 3D stage of the process.

I’m happy with her so far, she has an Eartha Kitt feel about her which I think is great for the character and the game. I’m looking forward to making the skeletons for her special move, roll on Jason and The Argonauts!

That’s it from me. Next time maybe she’ll be finished, maybe I’ll show some more concepts, and maybe the characters will have some finalised names!




Hi! I’ve been working on the menus recently, trying to refine the design of them to make them simple and super quick to get into a game from while also making them more compatible with the open style development thing we’re doing and of course looking nicer.


Main Menu

Having the menu scroll vertically instead of horizontally gets rid of the gaps we’d have to fill with huge character renders or something else, while still having the panel style, it also makes a lot more sense intuitively. In my mission to get rid of all unnecessary ‘Press a button to start’ screens, the logo sits on the menu and any other logos will live on the splash screens the consoles now have. The menu can be split into a 3×3 set up so additions to modes etc. can be easily made. I also changed the fonts being used throughout the menus from the rough type featured in the logo to this sleeker looking font to hint at the future tech featured on the courts and characters as well as making everything more readable and as a plus it will make translation easier in the future.


Character Select

We haven’t nailed down an exact number of characters yet, so I’ve tried to make the character select screen a bit more open so we can add characters one by one and put them in whatever shape works once they’re all in. I’ve also made some nice little indicators for player 1/2/cpu for each system to indicate what you have hooked up to your machine.


Level Select

There are pictures for all the levels now so you can see where you’re heading for your match. Just like the character select, I’ve made it easier to add levels one by one. Also like the character select, the boxes don’t have anything inside them yet, we’ll be adding character portraits and iconic landmarks from the levels eventually.

These new menus won’t be in the latest build at the time of writing, but sometime soon, maybe next time.

Since the last blog, we’ve also added the new special meter (sort of).

This hasn’t been fully implemented yet but we’ve been discussing how specials are done and how to divide them up.

The idea is you have four cells to fill, using one cell (Y) will give you a focus shot, using two (LT) gives you a random super shot, using three (RT) lets you use a pre-selected super shot and four (LT+RT) uses your character specific super shot. Right now I believe the supers are tied to Bomb, Ice and Fire respectively.

This change gives timing the usage of the supers a bit of strategy and also makes things a lot easier to interpret than before. Once the random specials are decided on and implemented I’ll be making some shiny new icons for them to go inside the empty cells.

Dylan also made an awesome effect for when a cell is active.

Next up for me is making some mock ups of the sub menus like options and the pause menu as well as making some more transitions between menus in the paint swipe style that are already in game.



Mavis used to land between shots & only fly for dives & runs. We noticed that because of the quick paced nature of anims being played this was quite jarring.  So I’ve gone through all the anims so she constantly flies. It feels much more natural now.


I also fixed the jitter issues we’ve been having with the ball on serve anims. This involved a complete re-export of the characters anim set & has now been completed on Mavis, Owl & Lion-El.


What’s Next


The focus for the next few weeks is going to be getting a range of special moves implemented and balanced for each character.  We’ve got some sweet ideas including health stealing orbs and skeleton minions for the Queen of the Dead character.  Hopefully we should get a new character in game too, the cowboy dude, complete with his six shooter special.