Game Design Deep Dive is an ongoing Gamasutra series with the goal of shedding light on specific design features or mechanics within a video game, in order to show how seemingly simple, fundamental design decisions aren't really that simple at all.

Check out earlier installments on the music-reactive Paranormical, the heredity system of Hero Generations, and the traffic systems of Cities: Skylines.

Also, dig into our ever-growing Deep Dive archive for developer-minded features on everything from rocket jumping in Rocket League to the one-hit kills in Titan Souls.

Who: David Perryman, Producer and Director at Caged Element

I worked in the games industry for 7 years around the turn of the century (the latter one, not that Victorian one). First by designing tracks for Rollcage and then helping to design and produce its sequel and a number of other AAA titles on PC, PS1, Dreamcast and PS2.

For the last 15 years I’ve been out of the games industry conducting research in psychology, helping people overcome their fear of flying and teaching people how to drive narrow boats. Most recently, my friends Rob Baker and Chris Mallinson persuaded me out of game retirement to help make GRIP, the spiritual successor to Rollcage.

What: Frustration within risk/reward

Rollcage was always a hard-to-get-going game: it needed some time to get the feel of the driving mechanic. We knew from the outset that this was a big challenge we were going to have to overcome with GRIP.

We’re treading a fine line between real physics and arcade playability for a five ton car doing 700km/h [435 mph], generating enough downforce to drive on walls and ceilings. It’s very easy to crash, but ultimately because of that it’s incredibly rewarding when you get it right and start pulling off perfect laps - it’s risk/reward in its most raw form.

And although crash recovery is an integral part of the players learning curve, finding that balance between a fun experience and frustration for new players is our biggest challenge. We need to find a way to help ease beginners into the game and reduce frustration, whilst maintaining that solid tense risk/reward that people love.

We had limited ability to address this difficulty with the original Rollcage. But we now have 15 years of technological and psychological progress to help us fix these issues.

Why: do some people find it frustrating and what can we do about it?

Based on feedback from Rollcage, we spent a lot of time trying to understand what made some elements frustrating for people. There were two main difficulties: Disorientation and Unfair crashing

Those were the main issues we had with Rollcage back then. Luckily, we had a chance to address them in its sequel Rollcage Stage II.

For disorientation, we examined the way the camera behaved and implemented a ‘stack-o-cam’. This identifies when you’ve crashed and pulls back from the car, pointing along the direction of the track. This aided crash recovery and got people back into the race quickly.

It worked relatively well, but it had the major disadvantage of destroying the sense of scale. These are massive tank-like cars and this camera approach makes them seem like toys. We need a better approach for GRIP.

For GRIP, we’ve looked into why some people find the game less disorientating than others and, applying some modern psychological methodologies, discovered that those players that do well are naturally good at creating an ongoing mental ‘map’ of the game, not just their position against their opponents but a situational awareness of their position in space.

They know who’s around them, what’s around them and which way they’re pointed. They know what they need to do to get back into the race if they crash. It comes from playing the game a lot, and most people get it with enough time. However, we want to encourage new players to be able to build this situational awareness quickly and overcome those early frustrations.

To do that we’ve put in a simple ‘map’ that just tells you if your competitors are in front or behind you. Keeping it minimal encourages the player to fill in the gaps and generate that situational awareness, whilst concentrating on the task at hand. Anything more, would take away from controlling a car going at break neck speeds. Other games have representational maps, and that’s great, but for GRIP they’d take too much processing to understand, taking the player away from the action and not allowing the player to build a mental situational awareness.

To encourage an awareness of orientation, we have small arrows on the HUD indicating the track direction. Contrary to other games, we want these cues to be almost subliminal in order to build that situational awareness. To encourage a sense of it, rather than a big ‘in your face’ directive arrow feeding people the answer - this is counterintuitive as a designer, normally I’d want to be conveying this information clearly and it seems to break the ‘rules’. However, In testing it seems we’re on the right track - people are finding it much easier to get back into a race after a crash without really knowing why.

Subtly lighting the half of the screen that is in the direction of travel helps too - head towards the light and all will be well…

Seemingly unfair crashing was a bugbear of Rollcage. Small bits of scenery can send you spinning or flying. This is the physics element of the game, and it’s what would happen if this were real. But it’s not fun.

Rollcage was at the limit of what was technically feasible back then, and we did everything we could to limit this. On Firebugs (aka Rollcage III), the tracks were essentially tubes and smoothed as a solution. However, this creates featureless and uninteresting tracks.

For GRIP, we want interesting tracks with the car seeming to be in contact with a real world. We have a load more options at our disposal these days. One small thing is that we have scaled the angular momentum in relation to speed so the faster you go the less likely you are to be violently thrown off course by hitting something. It has gone some way to working without destroying that tension and sense of fear you get the faster you go.

But a solution we’re most excited about is assisted landings. There is nothing better in GRIP than landing well after mistiming a jump, hitting a wall, spinning and flipping. Some of the best moments in the game come from this. We noticed that the better players pull these maneuvers off more than beginners. There is something they are doing, (releasing the throttle, dabbing the brake, rolling in the air or landing with full lock…) that is aiding it. We would love to be able to give a sense of this skill to a beginner and encourage them to learn it.

Since modern technology allows us to run the physics model ahead of time, we can essentially see into the future of how the car is going to land and we can nudge the car in small ways to achieve that perfect landing without it being obvious to the player.

However, too much of a good thing can be boring. If you know you’re always going to land well, the currency of those awesome moments is weakened. We wouldn’t want to randomize the outcome either as that defeats the purpose. But we can put the control in the players hands.

We have some limited air control, and the player can use that to ‘encourage’ the car to land in a favorable position. To do this, we grade a favorable outcome, for example: 100% is facing the right way, all four wheels on the ground, no skidding; 0% is facing the wrong way, one wheel landing and skidding all over the place. If the player does nothing in the air then they get a baseline percent outcome, say 50%-75% (depending on difficulty setting). However, whatever they do in the air adjusts that baseline percentage. If they nudge in the correct directions then we assist that up to 100%, if they get it wrong we go the other way.

This has two effects. It teaches the player how to predict the behavior of the car, and reduces the crash frustration.  They also feel like they were in control of the landing.

We can also monitor their success and adjust the baseline percent to help them out a bit if they’re not getting it. We’re yet to implement this fully, and it may well be that it doesn’t feel right - but we’re hopeful it’ll work well. Our aim is for it never to be noticed, so that the player feels they’re awesome.

There are other elements that may cause frustration in GRIP, like AI and weapon allocation, which are beyond the scope of this piece. But we’re confident that with modern techniques, we can encourage the building of situational awareness and reduce the effects of unfair crashing. All going toward reducing the frustration in GRIP that some players experienced with Rollcage.