A good idea at the time

One of the questions I get asked a lot is "What's you favourite game that you've ever worked on?", to which my instant reply is always "Syndicate". I also acknowledge the fact that I've been very lucky when it comes to the titles I've actually had the opportunity to work on. Back in those days, it seemed like we could do no wrong and the games were all pretty awesome.

But that got me thinking - it can't all have been brilliant, right?

And it's true - there were certainly things we did that, looking back, I kinda wished we didn't. Just little things mind. Things that, at the time, we thought "Hey, wouldn't it be cool if..." but that actually ended up being not such a cool idea after all and may, indeed, have broken the game somewhat.

These are the Good Idea At The Time features.


That guy on the right? Pointless
This was my first gig in the industry. It was pretty damn innovative too - living, breathing world, polygonal landscape, that sort of thing. Actually, just saying "living, breathing world" does it a bit of a disservice - the simulation layer underneath the AI villages was pretty ridiculous. It pretty much covered everything from birth to death and the entire working life that took place in between.

The gameplay was to revolve around a bunch of Captains who could be used to recruit and command your armies. Your main dude was in charge and he'd issue orders to the others. The GIATT was the carrier pigeon. Taking the idea of simulation just a bit too far, orders issued to subordinate Captains were issued via carrier pigeons. These poor little buggers had to fly across the landscape to their target before the order that you had given could be put into action.

In effect, all this did was add upwards of 30 seconds of lag to any command you gave to anyone other than your main dude. As this rendered them largely useless, the game simply devolved into making your main army as large as possible and steamrollering the land in one go, rather than anything that could be described as strategic placement of forces.

Of course, it remains an interesting idea and we could probably have balanced it out a bit better, if only by making the poor pigeons fly faster.


It's like Risk... with miniguns!
What? Really? Surely nothing was wrong with this? Well, there wasn't much wrong with the game itself, but the metagame had a bit of a problem.

The original idea had us replaying territories several times. The first time was to take the territory over, the second was to quell any revolt if you taxed the territory too much and the third was to re-take the territory if it was ever claimed by a rival syndicate. That idea fell by the wayside as there were 50 territories and only me building levels and missions.

The GIATT that made it through into the final game though was the research. Now, research on its own was no real problem. The issue was the realtime nature of it, meaning that a player could just sit on the research screen until he had got everything done. Even if he didn't do that, the free-form nature of the mission structure meant that the player could potentially enter any mission with any amount of progression (or lack thereof) through the research tree. This made balancing a bit of a nightmare.

In hindsight, it would have been a better idea to link research progress to successful mission completions. Coupling that with a more controlled mission progression spread and you'd have a much better (albeit, more scripted) experience.

Magic Carpet

That building could have been miles away
Glenn's major bone of contention here was the fact that we started using the landscape geometry to create buildings. This seemed like a GIATT as we didn't have any modelling tools to construct buildings and we certainly didn't want to use sprites like our previous games as they just wouldn't work for structures viewed from first person. A neat side effect was the fact that they would also be tied in to the morphing landscape - allowing us to spawn them in with a neat effect.

Unfortunately, by doing this and thanks to the texture resolution, we managed to completely mess the scale of the world up. Instead of being able to see for miles over smooth land, you could only see a couple of hundred yards in front of your face. For that reason, it suddenly felt like you were playing in a foggy mess.

Sadly, even with hindsight it's tricky to come up with a solution that would have not required tech that we just didn't have. It's either that or make the only buildings in the game cylindrical towers, render them like the creatures and hope for the best. Or have no buildings at all.

Dungeon Keeper

Waste of a perfectly good combat system design
For me, the appeal of Dungeon Keeper was very much in the observation of the world going about its business. Actually, to be fair, this was a common theme amongst many of Bullfrog's games and one that Peter often championed. As such, the interface was often purposefully obtuse to encourage a more removed gameplay style.

Dungeon Keeper's hand was a key component. Instead of issuing commands to creatures like a more traditional RTS, you'd simply drop them where you wanted and they'd try and work out what it was that you wanted them to do from the context. If they started doing something you weren't happy about, you'd give them a quick slap to make them re-evaluate. The hand itself could also hold multiple creatures at once.

The GIATT thing was to streamline the interface a bit. In the original, we refined it by allowing you to select your creatures from a panel as well as simply plucking them from the world. In the sequel, this was expanded to include a "grab all" button, which would allow you to instantly summon your entire army to your hand in a single click.

And, in that one, simple click, you'd completely negate any strategy or tactics normally associated with an RTS in so much as you could instantly place your army pretty much wherever you wanted. Battles would devolve into slugfests with each player just dumping their armies with nary a thought to composition or positioning.

Thankfully, this one's a simple fix - just don't have that button. Or, by way of compromise, limit the number (or total size) of creatures that your hand can hold. Hell, Hand Size could even be a thing that you could upgrade.

Battle Engine Aquila

See as everything shoots everything else, emergently
Lost Toys' finest hour, Battle Engine had a lot of Bullfrog DNA in it. Hardly surprising really, given the people responsible for it. Anyway, the main feature was the living, breathing world thing again, with huge battles taking place around you.

These battles were essentially an RTS being played out by two AI commanders. Units would be built and sent to the front lines, where they'd find bad guys to fight. The player had no control over this at all and, instead, was free to roam the battle and interact with (read: shoot at) whatever took his fancy.

The GIATT was to introduce scripting to enable us to sculpt these battles a bit more. On the one hand, this enabled us to orchestrate beach landings and other complex sequences that the AI would have been unable to replicate on its own.

On the other, we ended up using it far too much and moving away from one of our major USPs and towards an average, linear title. Again, a simple problem to fix - don't use it so much - but one that very much went unnoticed until it was too late.


I guess what I'm trying to say is to be wary that every design decision you make will have consequences. Some, you'll forsee and can deal with. Others will sneak up on you.

That's not to say it's always a negative thing - some brilliant gameplay mechanics have emerged from the strangest combinations. I mean, who could forget the game-changer that was the Rocket Jump?

But other times, it can break almost everything else. Is it a good idea to be able to save and reload at any time during an FPS? Well, it certainly seems like it and, once you've done that, it's a short and logical leap towards streamlining the process with a Quick Save and Quick Load. But that is the genesis of save crawling (the act of saving after each tiny event - even to the extent of firing a single round - and reloading if it doesn't work out quite how you'd like) which, in turn, plays merry hell with balancing. Either the game becomes ridiculously simple for the save crawler or obscenely hard for those choosing to play it without the crawl.

So how do you ensure that stuff like this doesn't happen? Well, there's no real shortcut I'm afraid. You're never really going to be able to work out all of the possible ramifications ahead of time, especially if the idea is a new one.

Of course, brand new ideas are pretty hard to come by these days, so this is where your extensive knowledge of what's gone before comes into play. See how you did it before. See how someone else did it before. Learn from it and ensure you don't make the same mistakes.

In short, play a lot of games! That's not too much to ask, is it?


  1. BladeII by Muckyfoot. Someone (maybe Disky?) put some smarts into the AI - when they were down to 10% health they'd run away, find some help, and bring some friends. Sounds clever, right?

    Two problems. First, you can't see them doing this. All you see is someone running off at high speed, and you don't know why (we didn't have health bars), and then later some more arrive, and unless you're paying very close attention the two are unlikely to get linked in your head.

    Second problem was much worse. You'd go into the first room, fight three baddies, one of them would get hurt and run off. Most of the time you wouldn't even notice, and you'd keep fighting the other two, maybe kill one, then the friends arrive, you'd fight all of them, kill a few, someone else would get down to 10%, they'd go find more friends, etc.

    Because we never let you fight more than about six people at once (because that's no fun), what would happen is you'd go into the first room, fight for about twenty minutes, and kill fifty vampires. Just in that one room. And then the rest of the level was completely empty because you'd killed everyone.

    Fortunately we caught that one before ship and turned it off. GIATT.

    1. What you described Tom was the usual process for clearing dungeons in Everquest, let a mob flee and bring back the rest of the dungeon with it then let fly with Area of Effect spells (or a thorn shield and good enough armour / health or a healer that you can survive getting hit by 50 mobs at once). It was probably more of an emergent thing in EQ but glad it stayed in even if it was used to grief people as well.

      Great article Alex!

  2. Hey, just found your blog, impressed so far!

    "Thankfully, this one's a simple fix - just don't have that button."

    This seems like a non-solution. The real problem is the fact that just massing a ball of death in most RTSes is optimal. Tying one hand behind the player's back and making it harder for them to select all their units is as much a solution as if you were trying to make Mario harder by making "jump" require a complex input.

    1. Yes, if you like your RTS sans 'S'. I'm still waiting for someone to come up with an RTS where the optimal strategy isn't to concentrate fire on each enemy unit in turn until they vanish.


Post a Comment

Popular Posts