Arcade Fire

Arcade Fire

By Ben Gunstone

August 21st 2007 at 11:50AM

Developers love digital distribution, thatâ??s a fact. Here, Stainless Gamesâ?? Ben Gunstone offers up some key advice on making games for Xbox Live Arcadeâ?¦

Stainless Games has now got a number of games published on Xbox Live Arcade and each and every one of them has presented a new challenge for us to overcome.

After moving from Sega Europe and working on big budget retail products I did briefly allow myself to think this was going to easy. XBLA: smaller games, shorter timelines – brilliant! But you know what? Making games is never easy (it wouldn’t be fun if it was, right?) XBLA just gave us different challenges, especially when we are applying normal AAA production values to small budget games.


1) Production QA times are the same – especially multiplayer

Every developer knows this – multiplayer games are more prone to errors than single player games.

Novadrome is an eight player Live game, and just because it fits within the 50MB download limit doesn’t mean that it’s sending any less packet data around. Just because it cost 800 points rather 40 or 50 pounds doesn’t mean that lag is allowed. Testing and bug fixing this aspect of the game took as long as any other retail eight player racing game. Your QA dept needs to be geared up for this testing and you need to budget for this.


2) Trial game – it's the Holy Grail

Don’t ever underestimate the power of the trial mode in an XBLA game. Getting this right is the way to increasing that all-important conversion rate – that percentage of free demo downloads that go on to achieve a sale of the full game.

There is the perennial balancing act to achieve between letting players see too much (so they feel they don’t need to buy your game) versus letting them see too little (so they don’t get to see the more exciting aspects, and therefore don’t buy the game). For instance, when Crystal Quest was originally released on XBLA the trial game showed the first eight levels.

This was great because it showed the player how the game mechanics worked but it didn’t let them experience the frantic excitement of the game’s later levels. In fact, we’ve decided to address this issue in a forthcoming Title Update, where the trial version ‘skips’ levels in order to give the player a better taste of the full version.


3) AAA Production values on a Budget Cost

Right from the beginning Stainless have pushed the message that we are providing AAA values at a budget price. We saw what was coming out on XBLA and decided we can do better. Of course, this costs money. With a small budget game you really have to make sure that the budget reflects the end quality expected from the publisher.

If you get it wrong and have to redo large amounts of work it’s pretty unlikely you’re going to make any money. This is really the same as bigger budget games but the difference is that the amount of content is smaller by a certain degree but the amount of people working on the title and the budgets available are smaller by a larger degree.


4) Consumer and Press expectation

It’s a constant irk that XBLA games get reviewed seemingly alongside the rest of the X360 games releases. Considering a lot of XBLA games get ‘casual’ releases on PC through sites like MSN games or Steam you don’t see those PC versions getting reviewed in mainstream games press. I’d like to see Eurogamer do a casual PC review section on their site to get some direct comparisons.

If your lead SKU is XBLA, it doesn’t matter that other retail games have had millions spent on them, have a DVD of space to fill up and maybe two or three years to make with a team of hundreds. Make sure you try and pick and choose the key ‘cool’ elements and see how you can emulate them.


5) Production times – expectations and prototyping

Designing original games that play well is an art. And guess what: it takes a fair old time to get it right. If you’re making a game with a six month time frame (start to submission) and you want it to be AAA quality then it’s easy to fall into the trap of designing on the fly and making the game at the same time.

Try to get the prototyping done and work with your publisher to make sure they understand the implications of a short timeline – there’s not going to be time to do central re-designing three months in as you’re already half way through the entire dev cycle.


6) Working on multiple short timeline projects

Expect to be in ‘release mode’ most of the time. This is more for fellow producers and QA teams out there rather than whole teams. If you are working on multiple short time-lined products (like we are at Stainless) expect to be in ‘release mode’ pretty much most of the time. That means most weeks there is either a milestone to deliver or game gearing up for a submission of some sorts.

A lot of Stainless’ clients are in the US which can mean a lot of late nights for the producer. We do put this to our advantage though as it can mean my dev teams here can work all day on a build, I prep it and send it over to the US client for overnight testing, with feedback ready for our UK-based team to work on the next day. Again, it’s the same if you were doing a retail product with a US client but the difference is that we only have six months to finish a high quality game, so time is of the essence.


7) Marketing products and general product awareness

We’ve found that XBLA games get next-to-no marketing or product awareness campaigns pre-release. I’m pretty sure that some of this is Microsoft defending the XBLA marketplace and giving the one or two releases each week a chance to breathe. A lot comes down to budget as well – publishers tend to allocate marketing budget based on a percentage of dev budgets and that’s not much for an XBLA game normally.

Make sure the marketing team understand the ‘XBLA space’ as it’s obviously different to retail and try to help them as much as possible. We’ve seen some great examples recently with the Sega game Happy Tree Friends where the website is doing a weekly blog of their visit to Stainless for a big design meeting.


8) Task management solutions

You’ve got a small team – that’s good. It means communication is simpler and task control is easier. However, that shouldn’t lead you into a false sense of security. We’ve learnt that even with the smallest teams you can’t revert back to old-school practises of just doing it all on the fly and hoping it will turn out right. But we do need to keep it simple. There is no point rolling out a product management tool designed to run huge teams on complicated projects if you have a small team of 4 or 5 people all sat in the same room.

We take a practical approach to this that is also scalable to our publisher’s needs. We use a central schedule built from major tasks, broken down to a level that I know each individual on the team needs. This is created using MS Project. This gets broken down into a combination of Excel task-sheets for weekly task updates and an online bug/task tracking database called Mantis, followed up with at least weekly meetings and constant dialogue within the team. We build in feedback from publishers whenever they desire (within reason!). This is normally at four weekly milestones, but can be more or less often depending on how the publisher wants to play it.

If you have been used to working with large teams over longer time periods make sure you use a project management solution that fits the situation – my key tip is to keep it is simple and make sure your team are prepared to use the solution you devise.


9) File size

There are strict file size limits set by MS for XBLA. It used to be 50Mb and has recently been increased to 150Mb. The file size limitation is also important when considering PC download. Sites like AOL and MSN can start to charge premiums on content that tops the 50Mb limit. With this in mind you need to be careful how you pitch the game (don’t set expectation too high) and you need to be very careful with data management.

Our art creation pipeline uses an internal tool called Acolyte for level creation and texture/mesh manipulation. Once items leave Acolyte we then use our in-house tool called CRUSH to compress the data. CRUSH pre-processes the assets and strips out non-essential data to convert to the exact binary representation for the Xbox thus leaving us with smallest possible file size.


10) Access to Microsoft XBLA Team

MS have a liaison and tech team working on XBLA. If your game is being published by a third party publisher you won’t get direct access to this team. This makes some of the XBLA specific certification rules more difficult to access – it can be a trial by error scenario. Try your hardest to get access to this team as they are the experts.


11) Content Packs

One area where extra revenue can be generated on XBLA is through the use of content packs. It makes a lot of sense to build this into the code structure from the beginning. It can be a gamble as it will change the way the code is built and it may make take more time but leaving it open to create content packs (even if the initial specification doesn’t call for it) can allow you to capitalise on your game after release.

As a fall back MS do allow you to do Title Updates but beware if you do more than one Title Update MS may well ask you to cover the certification cost if it includes executable data.

When you boil it down, making games for XBLA is the same as for any other platform. You need to create a feature set and a schedule to meet the timeline and budget that is available to you. The difference is that the time from design to going Live is so much quicker so there is less room for error. You have to be quick and you have to be right and you need to be quick and right the first time round.


Ben Gunstone is production director at Stainless Games overseeing the studio’s portfolio, which includes twelve Xbox 360 Live Arcade games. He previously worked in QA for Nintendo and then SCi, moving up to become a producer before moving to a similar role at Sega Europe, where he managed external products.