adventures in game design

open worlds without meaningful player choices

Assassin's Creed has taught me yet something else: The more open-world you are creating, the greater need you have for varied and interesting systems gameplay.

I don't have a definition handy, but if I had to define systems gameplay on the spot, I would call it "the set of rules and actions that the player interacts with in a game".

Let's quickly sift through a few examples, and rate them at Low, Medium, and High for "Openness" and "Systems Density"

GTA 3 Series
It really doesn't get more open than any of the GTA titles. The systems include an elaborate and varied vehicle system, a varied combat system, dozens of mission types, an economy, and an escalating warrant star system, with underlying systems like stunt scores, and even gang territories to create positive feedback loops that all tie the game together.
Openness: High
Systems Density: High

Assassin's Creed
There's an awesome openness to this world, and it's probably the best game that features a player who is primarily on foot in an open world. Most of the systems work is visible in the movement of Altair, while impressive, doesn't offer much meaningful choice to the player. Additional systems include a great combat system that sadly creates a negative feedback loop, and a few mission types.
Openness: High
Systems Density: Low

Bioshock
Bioshock is not as linear as a side-scroller, but as for level progression and layout, it falls into the low category for me. The underlying systems are not too numerous, but they offer a lot of diversity of choice, and they all feedback into each other. Main systems include a deep combat system, tonics, plasmids, camera research, and hacking.
Openness: Low
Systems Density: Medium

Tony Hawk's Pro-Skater Series
Openness: Medium
Systems Density: Low
THPS has open levels, but the size of them keeps me from placing this in the high category. Systems used are limited to about 4 trick types, a good scoring system, and a skill-development system.

All in all, it's really the 'bad' example of Assassin's Creed that ends up helping my argument the most. I strongly believe that the addition of an economy, and a positive feedback loop to combat (or hiding) could have harvested a metacritic score of 90 or more for Assassin's Creed as opposed 81 (metacritic.com).

Knob Twiddling. We all love to do it.

Sometimes, I catch myself tweaking values just for the sake of tweaking. Surely, most systems designers probably do this - and in most cases it's probably harmless. One of the pitfalls I've come across with this addiction to tweaking parameters is this:

I tune parameters because they are there, without considering if they fit the game in the first place.

It's so easy to get caught up in the simple fun of changing values and often I forget that maybe - just maybe - I might be over-designing. Let's look at one of the most common examples - guns.

When tuning a pistol for an FPS, a typical list of tuning parameters might include:

• damage
• firing speed
• reload time
• clip size
• ammo type
• accuracy
• recoil effect
• damage fall off
• bullet drop
• loudness

Reason 1: The tuning parameter isn't appropriate
Before diving into the tuning, consider whether some of these categories should even be used. For instance, if the shooter you are making is for a casual audience, then you probably don't need bullet drop, as accuracy tuning and damage fall off gets a similar effect without complicating the tuning.

Reason 2: The tuning parameter conflicts with others
Does one parameter directly conflict with another? Some games include a reticle 'kick' where the effect is that the camera angle is nudged up slightly with each shot. This feature can amplify the effect of any accuracy tuning. It might be better for your game to only use one of these variables. If not, you may want to use a graphing tool so you can quickly see how adjusting each value affects accuracy over time.

Reason 3: The tuning parameter adds 'bad' ambiguity.
Many games have weapons that randomize damage values between a min and max range - why? I can understand why a D&D game uses dice rolls: to simulate accuracy, recoil, etc. But if your game already includes accuracy, recoil, and every other possible tweak, what is the point of random damage?

Another example of 'bad' unpredictability is a damage ramp for explosive weapons. Grenades may have a max damage at the center of a radius, and a min damage at the end. But in practice, this gives a sliding scale of damage where it's difficult for both you and the player to determine at which distance a grenade will kill an enemy. I argue in these cases that it's better to have damage steps rather than ramps - even that a single damage value for the sphere is better.

I wonder if games that embrace ambiguity are using it as a sort of 'blur effect' to cover up potential oversights and poor attention to detail.

Will collect your bad ideas for achievements

After playing Assassin's Creed, I had to spend some considerable time working out why I didn't like the game, because I was also very impressed by it. One reason that surfaced was how much I hated collecting flags for nothing else but Achievements. Now, I haven't played the PC version, but I'm guessing there probably aren't any achievements - so are we expected just to collect flags? Where's the gameplay in that? It's like an Easter Egg hunt, without the candy.

Let me look a step upward at a game that incorporated collecting in a slightly improved manner: GTA. To be fair, I hated collecting in GTA as well, but at least the reward was in-game, and was meaningful - The gameplay: Collect 100 dubious brown bags, and you would unlock a weapon or armor for all time availability.

After all this thinking about collecting, I realized this simple truth:
 
Collecting stuff gets harder as you progress.

Take the real world example of the Easter egg hunt. Say that Mom hides 10 eggs around the yard, and you have to find them. The first 3 might be really easy to spot, the next 3 take some more time, and the next 3 take even more time. But it's the last egg that gets you to stop and wonder if Mom really only hid 9 eggs, but told you there were 10 as some cruel joke to get you to play outside longer. The point is that as each egg gets found, and subtracted from the world, the chances of finding the next egg diminish (and therefore, the time it takes to find the next egg increases).

So how might a good collecting system be implemented? To challenge myself, I'll use Assassin's Creed for my examples to prove my point that something better could have been done using the existing tech (with maybe a few days of extra work).

• A) Unlocks at regular intervals of increasing value.
  
  • At 25 flags, concept art is unlocked
  • At 50 flags, hiding from enemies takes 50% time.
  • At 75 flags, Altair can hold 10 additional throwing knives.
  • At 100 flags, Altair's sword damage is increased by 25%.

• B) One big unlock, but introducing a system to 'detect' hidden items at regular intervals.
  • At 40 flags, flags in a small radius are added to the map when Altair climbs a perch point.
  • At 70 flags, flags are simply added to the map.
  • At 100 flags, Altair's throw damage is increased by 100%

I wonder if the function for finding items in a world is linear, or exponential...

As stone sharpens stone

So before I get straight into posting, I should probably explain the point of this. After a roundabout career path, I've finally landed a job in game design. I really can't shut my brain off, and most of the time I'm up until 3 am thinking about game design anyway, so I might as well have an outlet for all those thoughts. Lately, I've been keeping a notebook for most of the thoughts that hit me while I'm playing a game, but writing tends to be too slow and hard to share.

So the point. This is a medium where I can throw out game design problems, solutions, and conversations and give anyone the chance to comment, reply, argue. As stone sharpens stone.