[In this extensive article, Gamasutra takes an in-depth look at racing game track design, comparing two arcade titles -- Initial D and Maximum Tune -- and contrasting them, at important points, against the approach used in the Gran Turismo series.]
Games are about empowerment -- being able to achieve and do things that you can't ordinarily (or legally) achieve in real life. Car racing games, especially arcade racing games, are the perfect example of this. Reaching the level of skill required to drive in F1 will be out of reach of 99.9999 percent of the population.
Financial overhead also plays a significant role in being able to drive your car on a race track in real life -- not only do you need a competent car, but you also need to pay all of the fees associated with "legal" racing. The most prohibitive factor though is age and licensing laws -- every teenage boy wants to drive a high-powered car, but that just isn't going to happen in reality.
This is where the humble arcade racer comes in. A well-designed arcade racer gives non-drivers an empowering, and most importantly, accessible experience.
But how do game designers make driving "on the limit" accessible, fun, and most importantly, empowering for non-drivers? The answer to this lies in analyzing the track design of two contemporary arcade hits: Initial D: Arcade Stage and Maximum Tune.
Anyone who has visited an arcade recently would have noticed that these two games dominate the arcade market, and are nearly always played by non-drivers piloting 820-plus horsepower monsters like pros.
There are two sides to this sudden and god-like ability for non-drivers to handle these uber-cars: vehicle dynamics and track design. This article focuses on establishing a set of rational metrics for track design and then applying these metrics to the analysis of the Mt. Akina track from Initial D: Arcade Stage 4 and the Hanshin Express Line from Maximum Tune 3.
Five Essential Metrics
In order to take a rational approach to track design, it is essential that we have a set of metrics. These metrics underpin all elements of track design and can be employed in a number of different scenarios. The five metrics of rational track design are;
- Metric 1: Race Line
- Metric 2: Clipping points (and related metrics)
- Metric 3: Track Width
- Metric 4: Camber
- Metric 5: Height Variation
Before analyzing how the case studies provide an empowering experience for non-drivers, it is essential that that these metrics are explained both individually and in relation to each other.
Clipping Points & The Race Line
The most essential metrics for rational track design are clipping point and the race line. However, to understand how these two metrics work, it is important to talk about the vehicle and its system of dynamics. Using the existing Society of Automotive Engineers' standard model, a vehicle has a pre-defined set of axes which are used in all racing games. Figure 1 is an example of the standardized SAE coordinate system.
The longitudinal axis is most directly affected by acceleration and braking. The yaw and lateral axis is most directly impacted by steering and subsequent weight transition. In a racing game, we want the player to apply as much longitudinal force as possible, as this will always result in a high speed. Application of brakes and lateral force will not only slow the vehicle down but also make it less predictable. With this in mind, we can now move onto the concept of clipping points and race line.
The metrics presented in this article are all based on the effect on the vehicle, not the end user. The track will always directly impact the vehicle, but only indirectly impact the driver. The reason for this is due to the fact that driving different vehicles on the exact same road will yield different experiences for the end user.
The vehicle is a translator of sorts, and is the essential middle man between the player and the track design. As vehicle dynamics is such a broad topic in itself, it will be discussed a future article, however the SAE standard axis system is comprehensive enough to be able to understand and apply the five metrics of rational track design.
Metric 1 and 2: Clipping Points & The Race Line
Every corner in a race track will have an ideal entry point, an ideal clipping point, and then an ideal exit point. The clipping point is a target on the roads edge that the player must aim for in order to maximize the chances of taking the shortest possible route through a corner, whilst at the same time placing the least amount of lateral force onto the vehicle (Figure 1).
In some cases, a clipping point might represent the shortest possible route through a corner; however, it may require the use of too much lateral force on the vehicle. These ideal points enable the player to apply as much longitudinal force onto the vehicle as possible (acceleration) without compromising this force with the addition of any unnecessary lateral movement.
Entry, exit, and clipping points are derived partly from creating the shortest possible route through a corner and also from using the most surface area of the road. The reason surface area is important is, if a road is wider, it means that less lateral force needs to be applied to the vehicle (remembering that lateral forces compromise acceleration and also make the car less predictable).
Jumping ahead slightly, Figure 3 has a dotted green line used to denote the race line -- or most optimal route through the corner. The wider the road is, the straighter the race line becomes and less speed-compromising lateral force needs to be applied. In all cases, opportunities to increase longitudinal force are most sought after by players in any arcade racer.
It is difficult to talk about these two metrics in isolation, as they are a product of each other. For a driver, they will create a race line in their mind based on an understanding of their vehicle and the clipping points. From a designer's perspective, the race line may supersede the track layout depending on the approach taken.
No matter what approach of these approaches is used in the design of tracks, the ideal race line should result in the least amount of lateral force being applied on the vehicle. To demonstrate these two approaches, we can refer to figures Figure 4 and Figure 5 respectively.
Figure 4 is the end result of drawing a simple spline to represent your track. Based on this approach, the designer would need to plan their clipping points then create the track around this. Figure 5, on the other hand, is just as valid, and uses the opposite approach. The designer would need to draw on the clipping points, then create the ideal race line. No matter what approach is used, the end result is represented in Figure 6 as a very basic, kart-style layout.
Although this is indicative of an end result, by no means does it represent the body of knowledge that is required to make an empowering track. The next issue that needs to be addressed is corner difficulty, which is a combination of race line and clipping points.
Difficulty of a corner increases if the angle between these three points is more acute and / or if the distance between these points is reduced -- as seen in a comparison between Figure 7 and Figure 8.
A corner can have multiple clipping points within itself; however, there will always be definite entry and exit points. The player needs these definite entry and exit points as a type of "punctuation" to help memorize the circuit.
Further to this, by taking this approach, a track designer can come up with a circuit that is a combination of straights and corners.
Depending on the type of vehicle dynamics system that is being used, a combination of straights and corners can help assist in game balance by allowing equal catch-up opportunities for vehicles with different attributes.
Corners that have multiple clipping points are referred to as "compound corners" (Figure 9). The approach to ascertaining difficulty remains unchanged for compound corners. Distance between clipping points and the amount of correction required by the vehicle (as defined by extrapolating the angle between three points) are valid means for ascertaining difficulty.
It is important to note that different racing games will have different types of vehicle dynamics, which will ultimately dictate the way clipping points are used. This will be described in more detail during the case studies of Initial D and Maximum Tune.
Understanding clipping points and the how they subsequently impact on creating the race line is the most integral piece of the puzzle when taking a rational approach to track design. It is important to consider that a racing game is very much a twitch-puzzle game.
What this means is that the player needs to find elegant solutions to spatial problems by using the least amount of steering, braking and acceleration input. Although vehicle dynamics have been discussed, the type of dynamics system used by the game will have a significant impact of how designers should go about creating clipping points and race lines.
Metric 3: Track Width
Track width is one of the common-sense type metrics that is relatively easy to understand and implement. As a general rule of thumb, making the road wider will make corners easier as it creates a more obtuse angle for the clipping points and also provides some forgiveness in the track design (as seen earlier in Figure 3).
A commonly acknowledged rule for third person games is that the environment should be scaled up by a factor of around 33 percent, and the same applies for racing games and the scale of tracks. The reason this approach is used is to allow space for the camera to move through the environment, and also to create passing opportunities.
When designing the scale of a track, each lane should be 1.6 car widths wide, allowing three cars to be side by side in two lanes with a very minimal distance separating them (Figure 10). In addition to this, the road shoulder should be around 50 percent to 75 percent the width of a car (Figure 11).
This makes the player take a risk by taking this strategically powerful position -- especially around corners where the lateral force will be trying to pull the vehicle's mass against the player's will. Having a wider section of track used for a corner will allow for more margin for error. Wider corners also give experienced players more ability to apply longitudinal force as the race line can become far more smooth.
Metric 4: Road Camber
Road camber is the metric associated with how the road's pitch and roll angle will affect the car's ability to turn-in to a corner (over-steer) or alternatively the car's inability to turn into a corner (under-steer) (Figure 12). The terms under-steer and over-steer are related to describing vehicle dynamics, however on-camber and off-camber are terms associated with track design (Figure 13).
Daytona USA famously used a large amount of on-camber turns to allow for impossible corner entry and exit speeds, designed to empower non-drivers and reward them for little input. On the other hand, more realistic driving games such as Gran Turismo will use combinations of on-camber and off-camber corners to test the limits of the player and their chosen vehicle. This type of model ultimately requires more from the player in terms of practice and hence is not suited to an arcade environment.
Ideally, all corners in an arcade style racing game will be on-camber, as this allows for some bold and empowering cornering. Given the type of demographic attracted to these types of driving games, this is an ideal approach.
Metric 5: Height Variation
Height variation in track design fulfills two distinct purposes. The first is to create asymmetric balance which may favor or disadvantage certain types of vehicles and the second is to create tension, relief and the occasional "vista moment". The latter of these two purposes is a topic in itself, and one that I have previously discussed in some detail in another article. However, to concisely summarize the two uses, we can consider that the first plays a game design role, which will either advantage or disadvantage certain vehicles based on their mass and power.
To paraphrase the later use, we can consider it to be a purely emotional component of both track design and level design in general. Limited line of sight will result in the player being anxious, as they are less able to plan ahead. When the player has increased line of sight, they feel empowered, as they are able to easily plan their moves several steps in advance. The previous article expands this concept in some depth.
From an emotional perspective, height variation makes tracks interesting, so long as it is used in moderation. Height variation can also be used to create "vista moments" for the player that add that extra wow-factor to the environment and the play experience.
These vista moments are usually achieved by restricting the player's line of sight for an extended period by forcing them to travel up hill.
Once they reach the plateau and begin their subsequent downhill run, their line of sight is increased significant in comparison to the uphill run, allowing them take in a full view of the circuit and the level art. Once again, as this is such a significant topic, it is best to use the previous article as a starting point to investigate the concept further.
Putting the Metrics Together
Now that we have an understanding of the metrics of track design, it is time to apply these in the contexts of the two case studies. However as vehicle dynamics plays such a key role in understanding the design of racing games, it is necessary to differentiate Initial D and Maximum Tune by examining the types of vehicle dynamics that they use. Maximum Tune favors a type of driving called the "Scandinavian Flick", whilst Initial D is characterized by a system which can be defined as "Race Line Punishment".
The Scandinavian Flick
Figure 14 is an example of how the Scandinavian Flick works. The player hits the ideal entry point and then turns the vehicle towards the ideal clipping point.
Once the vehicle's mass is heading towards this point, the player then counter-steers and aims the vehicle towards the ideal corner exit, eventually countering the vehicle's inertia and transitioning the mass of the vehicle towards a different point in space using a combination of longitudinal force, lateral force and yaw. It is the excessive amount of longitudinal force used in the Scandinavian flick which differentiates itself from pure 'grip' forms of driving.
Earlier in this article, it was mentioned that lateral force is detrimental; however, the vehicle dynamics in Maximum Tune have been designed in such a way that small portions of lateral force are actually required for the ideal race line.
In the case of Maximum Tune, the vehicle will lose speed on the turn in, but will gain speed on the counter-steer. What this means is, in Maximum Tune, it is often more desirable to spend more time in counter-steer -- and this subsequently means that in Maximum Tune most of the ideal clipping points are less than 50 percent of the way through a corner.
This is one of ways in which Maximum Tune addresses the issue of lateral force being detrimental. By having the clipping points earlier, it means that the majority of the corner is spent apply more longitudinal force than lateral force. Maximum Tune is able to make clipping points come earlier, by making the corner exits wider than the entries.
Missed Clipping Points
Both Maximum Tune and Initial D use missed clipping points as a way to punish players and provide overtaking opportunities. Figure 15 is an example of a player who has missed the ideal clipping point. In order to maintain the race line (and block the fastest possible way through a corner) the player will need to make a further correction, hence slowing the vehicle down unnecessarily.
A vehicle does not like to be disturbed by unexpected forces once in motion. Every time a correction is required, this places some extra and unexpected force of the vehicle, subsequently slowing it down or making its behavior erratic. In all racing games, a high corner exit speed is always the most desirable outcome for the player. These examples will be further expanded upon when examining the case studies.
Race Line Punishment
Initial D uses a system of vehicle dynamics which combines slip and grip vehicles and subsequently creates an asymmetrical relationship between vehicles -- vehicles do not need plastic symmetry, as the combination of the player and the track will always create asymmetry.
Some will be more inclined to use Scandinavian Flick, whilst others will be pure grip, race-line vehicles, taking the most effective line through a corner. This asymmetric relationship means that the primary mechanic in Initial D is race-line punishment.
Figure 16 is a time-lapse example of two vehicles entering the same corner. The silver vehicle takes a less risky race line, however it opens itself up to being overtaken on the inside using the more risky, but subsequently more effective, race line. As the red vehicle is able to hit the ideal clipping point, it will yield a higher corner exit speed. The CX values of the vehicles (a measure of how much energy they absorb in the case of a collision) means that the red vehicle has the advantage in Initial D, as its mass will push the silver car off the ideal race line.
In both Initial D and Maximum Tune, the player needs to weigh up how they will approach the corners. Figure 17 is an example of player who takes less risk by allowing greater room for correction on either side of the vehicle. Although they are less likely to hit the sides of the road (and subsequently lose la significant amount of speed) they are also taking a less effective race line and subsequently allowing room to be overtaken on the ideal race line.
By taking the most risky line through the corner as seen in Figure 18, the player has the most to gain if they hit the ideal clipping point, but also the most to lose if they fail, due to the lack of correction space.
Generally speaking, when vehicles in a racing game collide with each other, they impose less inertial forces than hitting a track barricade. What this means in relation to driving games is it is more beneficial to hit another vehicle as opposed to any other obstacle, as the player will ultimately be able to maintain a higher level of vehicle stability. When the vehicle is more stable, it is more likely to accelerate quicker and corner more predictably as there are no other forces acting on the vehicle's dynamics.
Initial D vs. Maximum Tune
So far we know that Maximum Tune and Initial D both have different approaches to vehicle dynamics and track design. However, they can both be described as highly empowering experiences for non-drivers. Both games provide a different experience and mood, so by analyzing some specific examples of track design within these games, we can come up with a track design language that will be useful when taking a rational approach to track design.
Characteristics of Maximum Tune track design and vehicle dynamics
- Tracks are much wider as they are designed for four players, instead of two or three in Initial D
- Much higher track speeds, so corners are more forgiving
- Designed to have other NPC vehicles on the road during gameplay
- Lead players can choose to take a different route through the course
- Much fewer tight turns than Initial D
Characteristics of Initial D track design and vehicle dynamics
- Two player network play against either the AI or other players, which emphasizes close, focused rivalries
- Slower speed vehicles which operate on narrower tracks
- Race line punishment
- No traffic obstacles
Hanshin Express Line: Maximum Tune 3
The Hanshin Express Line is a popular track in Maximum Tune 3, and has a lot of similarity to the track design in Daytona USA -- particularly the beginner circuit. The track is wide and has lots of sweeping, on-camber corners. Just like the Mt. Akina Circuit used in Initial D, the Hanshin Express Line is an embellished version of a real-world roadway -- this one in Osaka, Japan.
As Maximum Tune is based around the illegal freeway racing scene epitomized by the manga series Wangan Midnight, there is an emphasis on creating tense yet flowing high speed maneuvering through regular traffic.
As mentioned earlier, Maximum Tune is able to implement its ideal clipping points early in the corner, hence allowing for greater corner exit speeds. The way that it achieves this is by making the corner exits wider than the entries. From an emotion point of view, this may sound boring. However, the designers of Maximum Tune have built in a clever feature to embellish the player's actions and make them look better than what they actually are. This concept is known as needle threading.
Needle Threading: Maximum Tune
The Hanshin Express Line is epitomized by these needle threading corners. Put simply, a needle threading corner is one which provides a very wide corner entry point, an ideal clipping point (just one) and then the perception of a very narrow corner exit (Figure 19).
Needle threading corners have significant emotional value, and often give the perception of mastery for the players who are able to achieve them. Just like anything in game design, it is all about making the player look better than they actually are -- especially in the highly voyeuristic environment of the arcade.
Figure 20 shows how this sense of mastery and achievement is achievement. As the player corrects towards the ideal corner exit point, the vehicle is still sideways whilst passing through this "needle thread" point, making the track walls closer to the vehicle. The positioning of the camera is purposeful and captures this impossibly accurate near-miss which would more often than not lead to disaster in real life -- especially at the speeds cars are traveling at within the game.
When implemented a needle thread corner, the initial design needs to be done without the narrow exit point. If you look at Figure 20 and remove the barricade, we can see the design methodology at play. In Figure 21, you can see how the corner exit becomes noticeably wider than the entry, meaning that the ideal clipping point can be much earlier in the corner designs. The addition of the barricade is the illusion that the player needs to embellish their actions and make them feel empowered.
Asymmetric Balance: Initial D
The combination of multiple types of vehicle dynamics used within Initial D means that track design is of the utmost importance when trying to provide a balanced and empowering experience for the player. No track epitomizes this mantra more than Mt. Akina, an embellished version of the real world Mt. Akina road in Gunma Prefecture, Japan. Figure 22 is a planar map of the circuit. Although the track can played in either direction, the normal start line in sector one is the highest point of elevation and moves downhill progressively towards sector ten.
The type of vehicle that the player chooses at the start of the game will have a large impact on what types of track it will be best suited to.
A heavy car like the Nissan GTR will be balanced in such a way that it is usually more capable on an up-hill circuit where having more power is advantageous. A smaller car like the Toyota AE86 will have more of an advantage in down-hill circuits due to its lower mass and ability to over-steer on corner exits.
The Mt. Akina circuit uses a clever combination of corners to balance the gameplay and yield close rivalries, even without significant rubber-banding. Mt. Akina uses a unique leveling section in sector eight. If you are familiar with the Initial D franchise, then you will know the three consecutive corners as shown in Figure 23.
As the road markings play such an important role in the sensation of speed, they have been left in the diagram to demonstrate the sense of urgency that is conveyed to the player.
There are a number of metrics that need to be considered before the importance of this particular corner can be explained. In Initial D (and most other racing games) a compound corner which becomes progressively tighter will be advantageous for a lighter vehicle, whilst a compound corner which becomes progressively more obtuse will benefit a heavier, more powerful vehicle (Figure 24 & Figure 25).
There is also an additional component of compound corners which needs to be considered, and that is the correction distance between clipping points. To a large extent, this metric of difficulty can be explained by the angle extrapolation method covered in the clipping point metric. However, Figure 26 demonstrates the difficulty metrics associated with the distance between clipping points.
The reason the distance between these correction points is so critical is due to vehicle dynamics and Newton's first law -- an object in a state of uniform motion tends to remain in that state of motion unless an external force is applied. Figure 27 combines vehicle dynamics, distance between clipping points, and Newton's first law. The point of critical weight transfer is used in Initial D to either advantage or disadvantage certain vehicles.
The Use of Straightaways: ID4 vs. MT3 vs. GT5
One thing that has not been covered is the use of straightaway sections in tracks. Straightaways are necessary in all forms of racing games, however if a straightaway is too long, then this will adversely affect the sense of empowerment for non-drivers.
Straightaways may seem fun on first glance: a section of road where the player can apply as much throttle as they want. However there is a significant problem if they are used incorrectly, especially in an arcade driving title. To understand the difference between more realistic straightaway use and use with arcade titles it is necessary to take a basic statistical overview of three titles: Initial D, Maximum Tune, and Gran Turismo.
Figure 28 compares speed telemetry data from experienced players in powerful cars within Maximum Tune 3's Hanshin Line, Initial D Arcade Stage 4's Mt. Akina and Gran Turismo 5's Trial Mountain. Both Maximum Tune 3 and Initial D have the lowest standard deviation, whilst Gran Turismo has the highest deviation from the car's slowest speed to its fastest speed.
The graph plots the vehicles' speed telemetry at five second intervals and compares the current speed sample to the last previous sample. A value of 1 indicates that the vehicle has not changed speed since the last sample. Positive values denote an increase in speed and negative values denote a decrease in speed.
The graph demonstrates that in more realistic driving titles such as Gran Turismo, straightaways allow the vehicle to significantly overshoot the ideal corner entry speed, requiring the player to brake for extended periods prior to corner entry. This necessitates the player needing to learn the course, and takes away an aspect of forgiveness in the track design.
Console and computer games have the privilege of being far less forgiving. On a console, amateur players are not paying per play, hence allowing more time for practice and memorization. And most importantly, the home is a safe environment, away from the eyes of other gamers -- it is safe to make mistakes, and the player can pick and choose the time when they take their skills public via online play or tournaments.
Arcade titles though need to pander to the ego. Players pay per-play and they do so under the watchful scrutiny of others. Although straightaways are used in all forms of track design, to empower the non-driver, it is essential that straightaways do not allow the vehicle to reach an excessively high speed prior to corner entry. If the straightaway does lead to a tight corner, then there should be a number of easier corners prior to this, allowing the player to shed sufficient speed before attempting corner entry.
The initial aim of this article was to examine ways in which non-drivers could feel empowered within racing games. Initial D and Maximum Tune are both excellent examples of creating an empowering experience and via the analysis applied to each, we can come up with a set of design heuristics, based on the five metrics of track design:
- The use of straightaways needs to be implemented in such a way that the player is never able to significantly overshoot the ideal corner entry speed. Any braking should be kept to a minimum.
- Corners need to be progressively tighter to ensure that players are able to have close to ideal corner entries.
- The standard deviation in vehicle speed should be quite narrow, ideally within a plus or minus 30 percent region.
- Road markings are essential to creating a sense of speed.
- Needle threading corners should be used to provide a sense of awe for the player and embellish their skills.
- All corners in an arcade racing game should be on-camber.
- Height variations should always be used in track design to create emotionally diversity. Rapid changes in height should be avoided though.
- Use straights to create "punctuation" for the player. Separate corners using landmarks or iconic features to aid this.
There are other elements of best practice which are beyond the scope of this article. There needs to be a clear distinction between actual difficulty and perceived difficulty. The best example of this is the comparison made between speed changes in Initial D and Gran Turismo. Initial D is very good at giving the player the perception of difficulty, when in fact they are doing very little in comparison to what they would need to do given the same track in Gran Turismo.
As such, the next big area to address is vehicle dynamics. It is important to note that different vehicle dynamics do not downplay the significance of the five metrics, rather it the vehicle will provide a significantly different experience for the player, especially in regards to these different approaches to building difficulty.
Special Thanks to Cliff Kamarga, Katie Murphy and Natasha Lewin for the images.