Legend of the Golden Robot
Rob James/ robotJAM (Artist/bad coder/ designer): I spent 10 years making Playstation games and then 5 years in advertising/advergames and am now an indie game developer. I work with Rob Donkin on an informal basis, we come together when one of us has a concept to make.
Cathy McBurney (Game design/Animator): I’ve been obsessed with games since the Spectrum era, mainly adventures and JPRGs. Rob J is a good friend from university where we met on the same Computer Visualisation & Animation degree course. I was an animator and FMV artist at Rage and Acclaim in the UK, working on Expendable, Wild Wild Racing, XGRA, Alias and Rocky. Game design is something I’ve wanted to do for a long time and it’s wonderful to be able to hassle Rob and Rob with ideas and see them brought them to life.
Rob Donkin (Coder/Game Design): I’m a UK-based developer and have been making flash games for the last 3 years. My first ever sponsorship deal (Panda:Tactical Sniper) came with the added bonus of the sponsor getting robotJAM to redo my terrible artwork. Since then we have made a whole bunch of games together including Toxers, The Dreamerz and three more in the Panda series.
The Initial Concept
I wanted to be an archaeologist when I was a kid but only to dig up shiny things and big swords, hence the idea for Golden Robot. A realistic archaeology game would need too much specialist knowledge (and probably not be much fun) whereas reckless treasure hunting satisfies an obsessive need to collect loads of stuff. The initial design was just for a digging game, more like a single player Battleships. It had the big list of treasures, the spades, terrain types and costs to give it an element of strategy and the main goal was to find all the pieces of an ancient golden robot. Many of the treasures were based on real artifacts such as the Disc of Nebra and the Gundesrup Cauldron. Some were based on myths or popular fiction. I even managed to get a Queen reference in there!
I thought this sounded pretty cool, as I’d always fancied making an RPG type game in Flash and it had the start of the basic elements, gamers love games based on chance (like Solitaire, or Boomshine- games which don’t require any skill) so I knocked up a really basic prototype. This was basically the treasure dig areas and the main digging game in action. The initial prototype was built over a period equivalent to lunchtime, but took a few days to get the data structure sorted for generating all the treasures procedurally so I didn’t have to manually say what treasure was in what area. The game does create a different set of data each time you start a new game.
I thought it was the start of something, then I found a game by Jacksmack which had been entered into the Jayisgames Sandbox competition. It was a similar game nicely presented but somehow it really didn’t grip me game play wise (it was a nice game I just didn’t want to end up doing something that was similar). At this stage of the project it really was just a hobby project so I wasn’t really bothered if it had a stupid timescale. I made a decision at that point to make it more RPG-like, so I had an idea for making it into “Dig Fighter”, where each square would be a treasure or a monster. I fired off the idea to Cathy.
The Fighting Element
Rob J suggested there should be combat to add more scope. We worked out a scissors-paper-stone system with the spades as weapons. Your stats came directly from the spades because the emphasis was on finding treasure rather than fighting. At this stage the enemies were monstrous animals that you might dig up and the hero was going to be a character from an unmade sequel to Harry Quantum: TV Go Home.
I took a break from the digging side and knocked up a fighting prototype based on a system Cathy devised which at the time seemed to work ok. I also made a start on the main character, he was basically a cliche of all the adventure heroes I could think of, basically he’s a combination of Allan Quatermain, Indiana Jones and Chuck Norris.
At this point though Cathy decided to send me a new doc called Toxers, which was a similar game, but set in a city overrun with bugs (it was kind of like “I am Legend meets the Simpsons”). I dropped Golden Robot and rather crazily me and Rob Donkin jumped straight into production of that game. He had just finished University and we really wanted to start a new project together after having success with 5 other games we had worked on, most notably the Dreamerz. I carried on with the art a bit more and worked up the map screen from its initial basic state to a more detailed version.
This was starting to take shape but I really felt it was far too bright and colorful. I really needed to find a way of toning done the saturation, so I decided to have a go at trying to make it look like an old sepia map which is animated.
This basically involved just layering over some textures and then reducing the colours a bit from the art. Its something we carried over onto several other screens and I really felt that worked as a style.
Code Hand Over
About 3 months later Toxers was released and did pretty well, great scores on Kong and Newgrounds and had some amazing playtimes. I hadn’t actually shown Golden Robot to Rob D. yet as it really was my pet project that I was doing in my spare time. Once I did, Rob liked the start but we felt it was lacking a bit of character. I also hadn’t really touched on the artwork yet, the fighting section was looking good but the digging section was really just work in progress functional art. I left Rob with the job of converting my timeline AS2 disaster into his AS3 framework. The project really had gotten too big for me to handle alone so it was great having him come on board again.
When Rob showed me his prototype it was already looking fantastic so it didn’t take much persuasion to get me on board. The title alone was enough for me to say ‘hell yeah’! I thought that the digging for treasures worked nicely especially since there seemed to be so many unique treasures to find (I believe there were 91 special treasures and 20 generic treasures in total). Cathy and Rob had clearly already put a lot of love into the game so I was very grateful to them to invite me to join the project.
Having done Toxers already I felt confident that we could handle a project of this size so I wasted no time chopping the good bits out of Rob’s code and converting them into AS3. It was great to have a rough frame to work from and the data already there in nice neat arrays so it didn’t take long to get a more robust AS3 digging prototype up and running.
Although I liked the digging mechanic I soon got the feeling that there was something lacking. There wasn’t any strategy to digging so it didn’t take long before you were just systematically digging up every square in an area in a grid-like fashion. So it seemed logical to include a way of being able to locate treasures. We had a few ideas (metal detectors etc) before deciding on a minesweeper style where the number of neighbouring treasures of a dug square were shown in the corner. This seemed to work really well because now there was a thought process involved with digging if you wanted to maximise your profits. This was definitely something that the players seemed to respond well to (although of course not everyone reads instructions so there were plenty of people who didn’t even understand what the numbers meant…).
Rob had shown me his fighting prototype as well but I think we both felt that perhaps it could use a bit more attention. Although this was never really meant to be an RPG it was clear that, with the combat, it lent itself to a leveling up style of game play and some of the feedback on Toxers indicated that that was something the players wanted as well.
With this in mind we came up with a bunch of stats that you could build up to affect your performance in battle. From there we modified Rob’s initial combat prototype in an attempt to make it a bit more engaging. We wanted it to be simple but maintain a few elements of strategy so that it would stay interesting to players – which is not an easy task with turn-based combat! We toyed around with a few concepts before settling on the final version which included 3 different attack options, a defend option and a special item option.
Art and Animation
Having Rob do the programming was great as it really meant I could concentrate on getting the rest of the artwork sorted. The main bulk left was the creatures you would fight and the animations for them.
Rob and I brainstormed some ideas for bad guys and I spent about 3 days animating up around 28 different bad guys. They are all based on about 4 templates, so it really sped the process up as I could just change the components (legs/ arms / heads etc) on the characters and the animations then just required small bits of tweaking or 1 or 2 custom animations. Usually I like to draw stuff on paper before I work on the computer but as there were so many bad guys to draw I just dug out my Wacom tablet and drew them straight in. The first lot were the generic characters. I tried to make sets of them to save some time, so there’s 3 different goblins, 2 skeletons etc. This made the job from a few weeks into about 3 days for the lot.
And then it was boss time, we really wanted to have a few non generic characters to fight, Rob also came up with the idea that bosses would have a buddy with them which had some sort of special attack. I think he was just trying to make work for me at that stage and I really was too stupid to say no. The list of bosses were mainly from Greek mythology like the minotaur, cyclops, basilisk and the ninja goblin (I might have made that one up). The buddies I drew before the bosses, so they don’t make a lot of sense, but hey its a game, so why can’t a minotaur have a pet ferret. I did pretty much no design at all for the Evil wizard, it was only a week later I noticed he looked a bit like Ming the Merciless from Flash Gordon. I often find that I draw stuff thinking its new and I’ve just seen it somewhere and my subconscious has remembered it.
One thing we learned from Toxers was that players really don’t like missing with an attack and they especially don’t like missing several times in a row. We saw a lot of comments from people complaining that the hit percentages were way off in Toxers (because they had missed 3 times in a row with a 95% hit chance which as we all know is impossible). They complained so much that I had to check that my math was sound several times before I could be certain that it wasn’t messed up in some way. So for Golden Robot we did three things to keep the players happy. Firstly we displayed the hit percentages as a bar rather than a number so that they wouldn’t know the exactly what percentage of their hits should be landing. Then we reduced the chance of missing and also added in a ‘bodge factor’ so that it was impossible to miss more than twice in a row.
So at this stage the game was looking pretty near completion. We had added in all sorts of RPG type goodies to dig up or buy in the shop and some mini-bosses that, if defeated, would hand over a fighting buddy to assist you in combat. We’d added in a range of weaponry and defensive items and a load of monsters to bash. The final part of development though on a game of this size is always the toughest (as we had already discovered on Toxers). Inputting stats for the monsters and items was really tricky and required lengthy testing sessions to try end ensure that everything was well balanced. Cathy was a huge help at this stage and played for hours on end checking for bugs and balancing issues. There doesn’t seem to be any easy approach to balancing an RPG (if there is then please let me know about it!) but after several revisions we were pretty happy with how it was playing. Choosing to use version control was a must though and we were able to make several more revisions after releasing the game in response to player’s comments.
Metrics/Stats or “Keeping them playing”
Our goal on Golden Robot was to make a a big game with a big play time. I hadn’t seen any games which have an average of over 30 mins gameplay. Even Steambirds manages only 24 mins of playtime for the average player. I really was quite worried when we came to release to see if there was even a market for Flash games which take over an hour to complete.
The stats since release really are quite stunning though. The average player spends over 40 minutes playing the game. We also tracked loads of other metrics in the game to help with testing. Here’s a few of the more crazy ones.
- Total Monsters Killed : 79 Million
- Games of Shove : 7 million
- Cheated Death: 2.7 million
- Games of Dice: 6 million
- Spades Sold in shop: 700,000
- How many players with over 5000 kills: 1
- How many girlfriends that player has: 0
Which is slightly mad, the 3 mini games combined have been played 13m times, which since they were all an after-thought is amazing as that’s more than most really good games get played.
The whole process from start to finish on Golden Robot really was hap-hazard but then I’ve always thought making games is a creative process, you can’t simply write a doc and copy it word for word into a game. Well you can but that’s not really much fun to do. We did have a wiki which we used to write designs and ideas into but the process from start to finish really was an organic one, where we tried out new ideas, saw if they worked and then incorporated. It was also a bit odd leaving Golden Robot to work on Toxers. They are similar games but from the feedback from Toxers we definitely pushed Golden Robot more towards being a proper RPG.
Golden Robot and our previous game, Toxers, were kind of based on each other. The original idea for Toxers goes back to about the year 2000. I started thinking of a Flash version just before the treasure hunting idea came along. The search mechanics in Toxers came from the treasure hunting and some of the game play in Golden Robot is an expansion of Toxers because people commented that it should have more RPG elements, although it was never meant to be an RPG!