“The Mechanical Impact of Mario Kart’s Blue Shell”

A dissertation submitted in partial fulfilment of the requirements for the degree of MASTER OF
ENGINEERING in Computer Science

In

The Queen’s University of Belfast

By

Alex McMillan

11/05/2018
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Table of Contents

Acknowledgements.............................................................................................................................................................. 2
Abstract ................................................................................................................................................................................ 3
Introduction ......................................................................................................................................................................... 4
Original Implementation...................................................................................................................................................... 7
Technical Section.................................................................................................................................................................. 8
Implementation ................................................................................................................................................................. 11
Results ................................................................................................................................................................................ 14
Analysis ............................................................................................................................................................................... 18
Conclusion .......................................................................................................................................................................... 20
References .......................................................................................................................................................................... 24
Appendices ......................................................................................................................................................................... 25

1
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Acknowledgements

I would like to thank my supervisor John Busch for supporting me throughout the writing of this report and
invoking my sense of curiosity in Game Design. It is only due to his input over the years that this report can
be both interesting and relevant.

I would like to thank the many individuals who helped me gather data for use in this report: Ryan Pepper,
Kelvin McClean, Scott Sinnamon, Rosanna McBrien, Margaret Carlin, Clark Armstrong, Owen Mooney, Aimée
McMurray, Simon Campbell, Robert Finn, Aaron McCamley, Stuart Mathieson, Maria McQuillan, Simon
Gibson, Matthew Yardley, Laurie Bailey, Lee Smallwood and Matthew McComb, along with the 32 other
people who graciously took time out of their day to play Mario Kart for scientific purposes. Their help was
invaluable and greatly appreciated.

I’d like to also thank the community of ROM hackers at Origami64 for the huge amount of advice and
clarification they were able to give me as I took on the daunting task of learning how to ROM Hack from
scratch. I would like to give special thanks to user’s queueRAM, orbitaldecay, vexiant and shygoo.
QueueRAM for welcoming me to the Origami64 community and pointing me in the right direction, vexiant
for their prior work on the memory addresses of race drivers, orbitaldecay for their online tutorials and for
talking to me directly about properly using the ROM Hacking tools, and shygoo for building the software
needed to make this experiment feasible, project64d. I would also like to thank the team behind the
emulator Project64, without which project64d wouldn’t exist.

I would like to thank the nine people who comprise Easy Allies, for discussing the topic of my report with a
larger audience than I ever could have on my own and for providing me with a number of interesting talking
points which help colour this report. I would also like to thank them for providing me with love and respect,
and countless hours of entertainment during the writing of this report.

I would like to thank my significant other Laura, for providing emotional support and companionship during
a stressful period of my life. Without her, this report may have never been written.

Finally, and most importantly, I would like to thank my family for giving me the love and support I needed to
get me to this point. My grandmother Kathleen for letting me play my father’s old NES at her house, my
grandmother Elizabeth for letting me know every time we talk how proud she is of what I do, my sister Esmé
for sharing my lifelong passion for games, my late father Robert for showing interest in my interests, even if
he didn’t understand them, and my mother, Maureen, for keeping me fed, sheltered, happy, loved, and
always letting me pursue my dreams and helping me achieve them. I hope to make each and every one of
them proud with this report and in everything I do.

2
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Abstract

This report analyses the effects of the Blue Shell power up item in Mario Kart 64. The report opens by
providing context as to what the item does in game, and its intended purpose as a counteraction to the
phenomenon of snowballing. The report then discusses the item’s implementation in game and the potential
issues that arise as a result of this implementation.
The report then postulates that the item does not have the intended effect on the outcome of a race. The
goal of the report is to provide a better alternative to the item’s current implementation, or if none exists,
explain why the Blue Shell functions as it does.
The report then introduces the measurement of “competitive proximity”, and how this will be used to
quantify the mechanical impact of the item. The report then states that a Wilcoxon Sign-Rank Test can be
used to determine the significance of the results of the experiment, and why this is an appropriate measure
to use. The report then states the hypothesis, as well as the null hypothesis the data will be used to try and
refute. The report also states the alternative hypothesis, which will be supported if the null hypothesis
cannot be refuted. The report then outlines the experiment that will be undertaken to gather data, giving
details on all three races performed within the experiment and how they differ. The report then
acknowledges some experimental difficulties, such as the shape of the track and the co-ordinate system
used by the game. The report then shows the results of the experiments, both as a full data set and as a
filtered data set, discarding unaffected data points. The report then states that these results cannot be used
to reject the null hypothesis, supporting the alternative hypothesis. Finally, the report evaluates what
qualities the Blue Shell can bring to the game if it has no mechanical worth and discusses possible emotional
and psychological merits which may be of value to the player.

3
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Introduction

In August 1992, Super Mario Kart was released in Japan for the Super Nintendo Entertainment System. The
game was met with international critical acclaim, and great commercial success,[1] introducing the global
phenomenon of the Mario Kart franchise. However, it wasn’t until the release of Mario Kart 64 (MK64) in
1996, that one of the most iconic elements of the series was first seen; the Blue Shell. Then known as the
“Spiny Shell”, the Blue Shell is a power-up item which can be obtained by racers via item boxes. It acts as a
homing missile, targeting the racer in first place and temporarily immobilising them, allowing other racers to
catch up to them.
The Blue Shell is one of the most ubiquitous elements of the Mario Kart series, appearing in every title since
MK64. While there are some techniques that can be used to avoid the item (such as drifting or last-second
speed boosts), these vary from game to game and are often either randomised or require impeccable timing,
making contact almost inevitable. Because of this, the implementation of the Blue Shell can be described as
a negative feedback loop, wherein players who have achieved the goal of getting into first place are more
likely to be hit by the shell and subsequently fall out of first place.[2] The inclusion of a negative feedback
loop is not indicative of poor game design, as Marc LeBlanc states that negative feedback can stabilise and
prolong gameplay, and magnify late game successes. In the case of Mario Kart, none of these factors are
more apparent than the magnification of successes late in a race.[3]
The Blue Shell is a mechanism used by the designers to avoid ‘snowballing’: the idea that players in the lead
have more opportunities to accrue advantages, in turn pushing them even further into the lead. Snowballing
is considered to be an indication poor game design, as it forces players to continue playing even when their
chances of winning have greatly diminished.[4] LeBlanc refers to phenomena such as snowballing as “positive
feedback design”, as they magnify success early in the game.[3] The intended function of the Blue Shell as a
catch-up feature is made clear by Hideki Konno, the director of Super Mario Kart and MK64 (also the
producer of Mario Kart DS and Mario Kart Wii). Konno says “With Mario Kart 64, we wanted to [have
everyone in play until the end], but some of the processing problems occurred that didn’t allow us to do
that... we were trying to push them back together with 64, having eight racers on the screen all the time,
didn’t work all that well.”[5]
It can be deduced from these comments that the Blue Shell was a concession of sorts, implemented to allow
races to remain artificially tight (known as ‘rubber banding’) without having to render every racer on-screen
simultaneously. As the technology used by Nintendo has advanced, the Blue Shell has remained a constant,
with its functionality changing very little over the last 20 years. With processing power no longer a concern,
for what purpose does the Blue Shell serve now? It is not the only item counteracting snowballing present in
the series, racers further from first place are more likely to receive more powerful items (such as Golden

4
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Mushrooms or Bullet Bills), which should keep the race relatively tight. Although there now exist several
items fulfilling the same purpose, the Blue Shell remains, and is the only item to display such blatant
negative feedback. LeBlanc states that feedback systems, positive or negative, take control away from the
players.[3] It is this reduction of agency that leaves players feeling cheated after losing a race due to a Blue
Shell they had little or no chance of avoiding.
The makers of Mario Kart have considered the removal of the Blue shell. Kosuke Yabuki, the director of
Mario Kart 7 and Mario Kart 8, has said that “We have tried … to see what the game’s like without the blue
shell. When we’ve experimented without the blue shell, actually it feels like … there’s something not quite
enough in the game.”[6] This statement by Yabuki suggests that the Blue Shell provides some qualia beyond
the rubber-banding element previously discussed.
There is also the issue of player perception to consider. Yabuki says that the judgements made by the Mario
Kart team ultimately come down to whether they “feel fun”[3], so any mechanical impact (or lack thereof)
which the Blue Shell provides could be entirely discounted if it makes the game more enjoyable for those
playing. But how can an item that depends entirely on negative feedback make a game more enjoyable
overall?

Throughout this report, the following terms will be used:

Participant – A human performing this specific experiment.
Player – A human playing Mario Kart in a general situation.
Racer – The in-game character being controlled by the player.

This report will speculate the qualia which the Blue Shell has upon gameplay, to better understand how it is
that an item that on first appearance seems so unfair can make the game feel more fun.
The desired effect of the Blue Shell is to keep the race close, by helping those far behind and hindering those
far in front. Consider the following example: The racer in eighth place fires a Blue Shell at the racer in first
place. The desired outcome is that first and last place will have less distance between them after the item is
used. The racer who has used the item did so to in an attempt to move closer to first place, but by hitting the
racer currently in first, they have merely pushed another racer (likely the person currently in second) into
pole position. As the new leading racer has been unharmed by the Blue Shell, it is unlikely that the racer who
used the item has made up any significant distance on first place, or that a new racer occupies last place.

5
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

8th 1st

Fig 1. A visual representation of a Blue Shell being fired at Toad by Mario and the following change to positions.

The Blue Shell has not narrowed the distance between first and last, rather it has cycled individual racers
through the rankings. The user of the item has seen very little direct benefit, especially compared to the
distance they could have made up using another item such as a Golden Mushroom. The target of the Blue
Shell has been punished for performing well. The beneficiaries of the item’s use are those who were not
directly involved; those who were nearest to the target when they were hit, but who would be unlikely to
use the Blue Shell themselves. If the item’s primary effect is simply to push a new racer into first place and
leave the player in the lower rankings, why should that player wish to use it at all? What is it that makes an
item so mechanically obtrusive “feel fun”[7], and what does it add to the gameplay that could not be
replicated by other items that counteract snowballing?

Good game design requires every included mechanic to have a reasoning behind its implementation, and
after 20 years, the Blue Shell remains a constant element of the most popular racing game on the planet,[8]
despite the series’ director being unable to explain what it adds to the game. Is the Blue Shell a relic of dated
game design that remains simply because it is familiar or nostalgic, or does it serve some specific gameplay
purpose no other item has been able to replicate?

This report questions whether the item can be shown to have any measurable effect upon gameplay, and if
it’s inclusion in the series can be justified. By gathering real-time in-game data, the impact which the Blue
Shell has on race outcomes will be observed to interpret any mechanical effects.

6
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Original Implementation
The Mario Kart team are trying to marry two conflicting ideas; the outcome of a race should be
unpredictable, but not random. The game must give every racer a chance at winning, regardless of their level
of skill, but needs to ensure that the outcome is not arbitrary. Analyst Michael Mauboussin asserts that it is
possible to tell if a game is entirely based on chance by whether or not it is possible to purposefully lose (for
example, it is impossible to intentionally lose a game of roulette).[7] It is possible to deliberately come last in
Mario Kart by simply doing nothing, so therefore it can be established that it does indeed have an element
of skill.
As MK64 saw the first implementation of the Blue Shell, it is important to remember that its introduction
was to circumvent technical issues. As there has been very little alteration to the Blue Shell’s mechanics
since then, it is clear that it plays some key role in the skill/luck balance which has become so prevalent in
Mario Kart. This report will consider what mechanical impact the Blue Shell had when it was first
implemented in MK64, to better understand how this alters gameplay, and whether those alterations
change the outcome of races in any meaningful way.
One notable difference between the Blue Shell’s implementation in MK64 and in future Mario Kart titles is
the potential for the MK64 Blue Shell to hit several players on its way up to first place. In future iterations,
the Blue Shell flies over the head of the racers, it therefore leaves all other racers unharmed until it dives
onto the racer in first place, also producing a small blast radius around the first place racer that can
incapacitate other racers who pass through it. However, in MK64, the Blue Shell trails along the track like any
other shell, travelling along the most direct path regardless of other racers in the way. If the Blue Shell
collides with another player, it will incapacitate them in just the same way as it would incapacitate the racer
in first place. This implementation is not the standard seen in the series going forward, and the Blue Shell’s
ability to hit other racers is entirely random and inconsistent, so for the sake of this report it shall be treated
as anomalous behaviour.

7
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Technical Section

The experiment carried out for this report will consist of 50 short gameplay sessions. In each gameplay
session, a different individual will participate in three races, all of which will be at the 150cc difficulty setting
on the Luigi Raceway track (this track contains no environmental hazards which could serve as confounding
variables). Each participant will play as Mario, starting the race in eighth position. During these races,
quantitative data will be collected by tracking their positions throughout the race, as well as recording the
final placement of competitors at the end of a race. The game used for the experiment will be MK64, and the
game will be running on a special version of the Nintendo 64 Emulator ‘Project64’, which contains a
debugger. This allows for the injection of JavaScript at run-time, which in turn allows for real-time gameplay
alteration and data recording. Each of the three races the player will participate in will serve a bespoke
purpose:
● Race 1: Control – Played under standard conditions [appendix 1], this race will allow observations to be
made of the individual’s usual performance and skill in the game, and data to be collected on racer
positions throughout a regular race.
● Race 2: Zeroed Blue Shell Frequency – Played under the same conditions as Race 1, with the
exception that the Blue Shell’s probability is altered to 0% in every position [appendix 2]. The results from
this race will be compared against the results from Race 1, to determine any significant changes to
the race outcome.
● Race 3: Increased Blue Shell Frequency – Played under the same conditions as Race 1, with the
exception that the Blue Shell’s probability within the item rotation is tripled in every position [appendix
3]
. The results from this race will be compared against the results from Race 1, to determine any
significant changes to the race outcome.

The individual will not be informed of the removal of the Blue Shell in Race 2, nor the increased probability
of the Blue Shell in Race 3. The exact nature of the experiment will not be disclosed to participants in order
to remove any potential bias.
The experiment will not be concerned with the individual’s final positions in any of the races, as this relies on
too many contributing factors. Instead, the experiment will be using these races to track the changes of
position for competitors throughout the race, as well as the distance between all competitors at the end of
the race. The measurement used to determine the ‘closeness’ of any particular race will be an average of the
total Euclidean distance[9] between each pair of consecutive racers, known as a race’s “competitive
proximity”.

8
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Fig 2. The above equation calculates the ‘competitive proximity’ of any race between n participants in a three-
dimensional Euclidean space

The competitive proximity of a race is calculated by taking the Euclidean distance between first and second,
second and third, third and fourth etc, up to and including the bottom two racers. These distances are then
weighted accordingly, based on how much impact they will have on the final result (for instance, the
distance between the top two racers is more significant to the race being close than the distance between
the bottom two racers). This value is then divided by the number of distances calculated to give the average
weighted distance between any two racers at the end of the race - the race’s competitive proximity. As the
track being used in the experiment is relatively flat, the difference in the y co-ordinate between racers will
be nominal compared to the x and z co-ordinates, and so for the sake of this experiment, it will be ignored.

If the competitive proximity for a race is low, this means that the racers were close together when the
winner crossed the finish line; conversely, if competitive proximity is high, this means that the racers were
well-dispersed. It should be noted that the general equation for competitive proximity [figure 2] will work for
any distance based race, both real-world and virtual. As such, this equation could be used to compare
different Mario Kart games to one another, compare the competitiveness of racing games against one
another, or even predicting the outcomes of events such as the Grand National or Tour De France (however,
in real-world situations, exact location data would be much more difficult to gather).

Keeping in mind Konno’s prior comments that the Mario Kart team want races “…where everyone [is] in it
until the end…” , races which result in a low competitive proximity are more desirable. If these comments are
interpreted to mean that no player should be ruled out from winning until the race finishes, it would suggest
that the designers want races where all the racers are relatively close to one another as soon as somebody
crosses the finish line. This can be represented by a relatively low competitive proximity.

By comparing the three sample sets of competitive proximity values gained from the races across all
gameplay sessions, a Wilcoxon Signed-Rank Test may be performed to see if the removal of the Blue Shell
makes any significant difference to the competitive proximity [10]. The Wilcoxon signed-rank test is a non-

9
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

parametric test used to determine if two related data samples have the same distribution. A Wilcoxon
Signed-Rank Test is appropriate because in each sample there is a comparison being made between two
paired sets of values (Race 1 vs. Race 2, Race 1 vs. Race 3) to establish whether there is a significant
difference in the values after a treatment has been applied [10] (respectively, the removal of the Blue Shell
and the increased frequency of the blue shell):
● Hypothesis (H): The competitive proximity of a race in Mario Kart is affected by the presence of the
Blue Shell in the game’s item rotation.
● Null Hypothesis (H0): There is no significant difference in the competitive proximities of races in
Mario Kart with or without the Blue Shell in the game’s item rotation.
● Alternative Hypothesis (H1): The competitive proximity of a race in Mario Kart is unaffected by the
presence of the Blue Shell in the game’s item rotation.
Should the Wilcoxon Signed-Rank Test find a significant difference between the two samples in either pair of
experiments, this will allow for the rejection of the null hypothesis and suggest that the Blue Shell does have
a mechanical effect. Otherwise, this will support the alternative hypothesis, and suggest that the Blue Shell
does not have a mechanical effect.

10
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Implementation

The item rotation probabilities will be altered by modifying the hex data in various ROM addresses during
runtime, using a piece of JavaScript run through Project64d. These ROM addresses will be located by
isolating changing variables using a memory address viewer known as Cheat Engine. This involves filtering
every memory address in the emulated game by pausing the game at short intervals, where the only thing
changed by the player during active gameplay was some element which would be affected by that variable
(for instance, to find a racer’s position on the track, find every address holding a changed value while the
racer is in motion, then filter that list of addresses to find any address holding an unchanged value while the
racer is stopped. This process must be repeated several times.).

In MK64, an item is randomly selected when a racer passes through an item box by accessing a ROM address
which stores the first of 100 item codes and offsetting this address by some random value between 0 and
99. There exists a different table of 100 item codes for each racing position (first through to eighth) that
ensures better items are more likely for competitors who are further behind. In order to modify the item
rotation, three separate JavaScript files were created (race1.js, race2.js, race3.js). As Race 1 is a control, this
script does nothing of consequence; however, race2.js and race3.js each access the eight relevant item
tables and alter the hex values contained within them, such that any address containing the item code for a
Blue Shell, ‘07’, is removed and replaced with another item (race2.js) or tripled in frequency by removing
other items from the table (race3.js).
It should be noted that these eight item tables are applicable only to the human player in each race, and that
the CPU competitors have their own set of eight item tables. This distinction is important as it shows there to
be a subset of items in MK64 which the CPU is never given access to, and as such, is not programmed to be
able to use. This subset of ‘unfireable items’ (Multi-bananas, Red Shell, 3 Red Shells, Blue Shell, 3
Mushrooms, Gold Mushroom) is important to take into consideration, as it shows there is a conscious
imbalance between the item usage of CPU players and human players. Alterations made to item rotation will
not affect the CPU; nor should they, since the CPU cannot utilise Blue Shells.
In order to track racer position throughout the race and distance from first place at the end of the race, a
JavaScript file is used at run-time (trackRace.js). This outputs each competitor’s in-game co-ordinates to a
text file after a constant and arbitrary number of cycles. Upon the data output cycle which corresponds to
the end of the race, these final co-ordinate outputs are placed into a separate data file for use in a Java
program which can calculate the competitive proximity from these values.

11
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Fig 3. An overhead layout of Luigi Raceway

Under optimal experimental conditions, the race track being used would either be a straight line or a perfect
circle. Of all the tracks available in MK64, Luigi Raceway has the most geometrically simple track layout, and
is therefore the most suitable track available for this report. The ‘collapsed oblong’ shape of the track does
lead to some potential issues with calculating the distance between competitors: the points at which the
strait of the track runs parallel with the track segment within a tunnel. Should two consecutively positioned
racers be located concurrently in these two areas of the track (displayed as two red dots[figure 3]), it would be
clear to anyone observing that there was some significant distance between them on the track (displayed as
an orange line[figure 3]). Yet, their positions on the track would be relatively nearby as the crow flies (displayed
as a blue line[figure 3]), and as such, this would lead to some very misleading data. This potential anomaly can
be accounted for by checking if racer’s co-ordinate values lie within these sections of the track when the race
concludes, and then either voiding this data-set, or adding to their distance an approximate length of the
loop of track which separates them. ‘Competitive proximity’ has to take into account the distance between
each pair of consecutive racers rather than each racer’s distance from the finish line due to shape of the
track, as it is less likely for gaps of this size to occur between consecutive racers than it is for all racers to
have passed this pinch-point when the race ends.

12
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

The co-ordinate values used are approximated due to the game calculating distance using a minuscule,
arbitrary unit, which needs to be equated to a value of much smaller magnitude for computation.
Furthermore, it appears that the racetrack is stored in several different memory locations, as there exists a
point on the track (displayed as a yellow line[figure 3]) where a z co-ordinate of ~25’000’000 is immediately
followed by a significantly lower z co-ordinate of ~1’500’000, and not the value ~24’999’999 as would be
expected. This means a Euclidean equation is now being used to measure a non-Euclidean space, such, any z
co-ordinate below ~1’500’000 must be given a value equivalent to its distance from that point as opposed to
its actual value.

To limit the independent variables that may influence the results, the experiment will be performed upon a
set of subjects with a relatively similar amount of prior experience with MK64. The experiment will aim to
only use subjects who have played Mario Kart previously and can do so with confidence. Other than this
criteria, the subjects will not be limited by any other attribute, such as age or gender. While this is not
necessarily reflective of the game’s audience, it will remove the possibility of anomalous results caused by
test subjects who are unfamiliar with the game, and as such would perform disproportionately poorly.

13
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Results

Appendix 4 shows a table containing all 50 data samples for each of the three races, alongside the initials of
the participant. The cells of this table are coloured such that smaller competitive proximities are green,
larger competitive proximities are red, and middling values are yellow. When the race data sets are placed in
ascending order, it becomes apparent that the gradation from green to yellow occurs further down the
column in Race 1 than for Race 2, and likewise the gradation in column 2 occurs further down the column
than Race 3. This would suggest that there is a larger quantity of small competitive proximities in Race 1 than
in Race 2, and likewise for Race 2 and Race 3.

The Mean (and Median) values for Appendix 4 are:

Race 1 – 943.04 (1042.05)
Race 2 – 947.71 (1022.74)
Race 3 – 1029.19 (1125.68)

To take into account the possibility of outlying data among the 50 samples, the Mean value of the Median 30
data points in each race is taken. The Mean values for the Median 30 data points are:

Race 1 - 958.48
Race 2 - 980.67
Race 3 - 1062.75

The data set does not consider races which may not have been impacted by the treatment, as in races in
which the participant did not receive an item whilst in any position lower than third. These redundant data
points are struck out in Appendix 4, and produce the following new Mean (and Median) values:

Race 1 – 987.58 (1051.62)

Race 2 – 1009.11 (1126.54)
Race 3 – 1067.15 (1125.68)

Appendices 11, 12, 13, 14, 15 display a sample of race placement changes in real-time. There is an
accumulative total of 188 place changes among the top four position, as opposed to 136 place changes
among the bottom four positions. Place changes among the first four positions are 32% more frequent than
among the back four positions.

14
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendices 16 and 17 display the Box and Whisker diagrams for the full 50 data point and the filtered data
set respectively. The values derived from these diagrams are displayed below:

Race 1 Race 2 Race 3

Highest Value 1501.36 1371.49 1680.67

Upper Quartile 1193.50 1201.38 1275.87

Median 1015.70 1034.09 1125.46

Mean 936.77 953.48 1030.81

Lower Quartile 663.83 637.04 703.75

Lowest Value 362.96 372.85 386.44

Range 1138.4 998.64 1294.23

Interquartile Range 529.67 564.34 572.12

Fig 4. The corresponding values to Appendix 16, showing data referring to the full set of 50 data points

Race 1 Race 2 Race 3

Highest Value 1501.36 1371.49 1680.67

Upper Quartile 1196.95 1250.73 1292.02

Median 1051.62 1126.54 1125.68

Mean 987.58 1009.11 1048.77

Lower Quartile 708.46 678.08 711.48

Lowest Value 369.53 372.85 534.23

Range 1131.83 998.64 1146.44

Interquartile Range 488.49 572.65 580.54

Fig 5. The corresponding values to Appendix 17, showing data referring to the filtered data set

15
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

When a Wilcoxon Sign-Ranked Test is performed* upon the full set of 50 data points, the following results
are given:

Race 1 v. Race 2 (Full set)

The Z-value is -0.2172. The p-value is 0.82588. The result is not significant at p≤ 0.05.
The W-value is 615. The distribution is approximately normal. Therefore, the Z-value above should be used.

Race 1 v. Race 3 (Full set)

The Z-value is -1.6652. The p-value is 0.09492. The result is not significant at p≤ 0.05.
The W-value is 465. The distribution is approximately normal. Therefore, the Z-value above should be used.

*Wilcoxon sign-ranked test performed using the online calculator:

http://www.socscistatistics.com/tests/signedranks/

16
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

In order to perform a Wilcoxon Sign-Ranked Test upon the filtered data set, some further data points will
have to be removed, as each pair of data points must come from the same population. This means that any
Race 2 or Race 3 data points which have a filtered Race 1 equivalent data point must be ignored, as this data
cannot be adequately paired. After further filtering the data, 19 pairs of data points remain for Race 1 v.
Race 2, and 22 pairs of data points remain for Race 1 v. Race 3.

When a Wilcoxon Sign-Ranked Test is performed* upon the newly filtered data points, the following results
are given:

Race 1 v. Race 2 (Filtered set)

The Z-value is -0.2415. The p-value is 0.81034. The result is not significant at p≤ 0.05.
The W-value is 89. The critical value of W for N = 19 at p≤ 0.05 is 46. Therefore, the result is not significant at
p≤ 0.05.

Race 1 v. Race 3 (Filtered set)

The Z-value is -1.2824. The p-value is 0.20054. The result is not significant at p≤ 0.05.
The W-value is 87. The critical value of W for N = 22 at p≤ 0.05 is 65. Therefore, the result is not significant at
p≤ 0.05.

The margin of error for this experiment is 13.86%, taking into account the sample size of 50 and the global
population of 7’600’000’000.

17
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Analysis

For the three races in the full data set, the mean is a smaller value than the median, meaning the data is
negatively skewed. As the data is skewed, the median value will be better suited to compare the races with.
Race 1 yielded a higher competitive proximity than Race 2, and a lower competitive proximity than Race 3.
This suggests that each participant’s second race was closer than their first race, and that their third race
was less close than their first race.

For all three races in the filtered data set, the mean is a larger value than the median, meaning the data is
positively skewed. As the data is skewed, the median value will be better suited to compare the races. Race
1 yielded a lower competitive proximity than Race 2 or Race 3, with Race 3 returning a larger competitive
proximity than Race 2. This suggests that during races impacted by the experimental changes, each
participant’s Race 1 was closer than their Race 2, and their Race 2 was closer than their Race 3.

Appendices 11, 12, 13, 14, 15 show a random sample of 5 races among the 150 races run within the full data
set, providing a chart of position movements throughout each race. The 32% difference between the 188
position changes between the top four racers in these five races and the 136 position changes between the
bottom four racers suggests that there is an unequal distribution of position movement, and that racers in
the top four are more likely to change position than racers in the bottom four.

With the full data set of 50 data points, Race 2 produced a lower median competitive proximity and a slightly
higher mean competitive proximity than Race 1. This suggests that among the data given, an average Race 2
is more likely to have a lower competitive proximity than an average Race 1, but that those Race 1 data
points which are below average are more deviant from the average than in Race 2 (or conversely, those data
points which are above average in Race 2 are more deviant from the average than Race 1). Race 3 produced
a higher median and mean competitive proximity than Race 1, meaning an average Race 3 would have a
higher competitive proximity than Race 1, and that data points from Race 3 which are above average are
more deviant from the average than Race 1.

Appendix 16 and Figure 4 both show that the range of competitive proximities was smaller in Race 2 than it
was in Race 1, but that Race 2 did possess a larger Interquartile Range. A similar relationship of smaller range
but larger Interquartile Range between Race 2 and Race 1 in the filtered data set, shown by Appendix 17 and
Figure 5. This suggests that Race 1 produced more outlying data points than Race 2, but that those non-
outlying data points were overall more consistent in Race 1.

18
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 16 and Figure 4 both show that the range of competitive proximities was larger in Race 3 than it
was in Race 1, and that Race 3 also possessed a larger Interquartile Range. This trend can also be seen in
Appendix 17 and Figure 5, where Race 3 has a larger range and Interquartile Range than Race 1. This
suggests that the data points for Race 3 were altogether less consistent than the data points for Race 1.

The results from the Wilcoxon Sign-Rank test upon the filtered data set show an insignificant W-Value and Z-
Value for both Race 1 v. Race 2 and Race 1 v. Race 3. This suggests that in both cases, the results returned
were insignificant.

The results from the Wilcoxon Sign-Rank test upon the full data set show W-Values for both Race 1 v. Race 2
and Race 1 v. Race 3 which are below the critical value, meaning the Z-value can be used in both cases. In
both cases, the Z-Values suggested that the results were not significant.

With all tests returning insignificant results, the null hypothesis has not been rejected, supporting the
alternative hypothesis that “The competitive proximity of a race in Mario Kart is unaffected by the presence
of the Blue Shell in the game’s item rotation”.[H1]

19
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Conclusion

The results have shown that there is no significant difference between the competitive proximities of the
control data (Race 1) compared to either of the two races with altered conditions. This suggests that there is
no meaningful change to the ‘closeness’ of a race should the Blue Shell be increased in frequency or
removed entirely, supporting the alternative hypothesis, and demonstrating that it has no mechanical
impact.

If the idea that the Blue Shell exists as a catch-up mechanism to keep races close can be discounted, what
does that leave as a potential mechanical reason for the Blue Shell? The commonly held belief is that the
Blue Shell is some sort of “Great Equalizer” [11], enhancing the struggle for those players attempting to hold
onto the lead and giving racers at the back a much needed chance of recovery. Yet, as can be seen from the
results, the majority of position changes happen in the top four regardless of the frequency of Blue Shells,
and races are just as close when the Blue Shell is removed. What, then, does the Blue Shell achieve?

Taking Appendix 14 as an example, in a race where the player fired five Blue Shells, three of these Shells lead
to the player moving up in the rankings (by four positions, one position and one position respectively) and
two of the Shells lead to the player moving down in the rankings (by two positions and three positions
respectively). This means that on average, a Blue Shell fired during the race moved the player up by 0.2
positions. The five Blue Shells fired during the race targeted the three different racers who held first place at
these times, and while the shells clearly did cause some movement between these higher positions, the
player never overtook any of the three racers targeted by their Blue Shells. All three of the targeted racers
ended up finishing within the top three positions. It is clear that the Blue Shell neither helps the person who
fired it progress, nor does it make the race significantly more difficult for those targeted - the Blue Shell
serves no mechanical purpose whatsoever.

If the mechanical impact of the Blue Shell is nominal, this means that the qualia [6] the Blue Shell adds to
Mario Kart is most likely psychological. The Blue Shell engages the player by giving them something other
than a mechanical benefit in the race, in that it still “feels fun”,[7] and feels relevant to the player despite not
being relevant to the race.

While LeBlanc states that feedback systems take control away from the player, from a player’s perspective,
the Blue Shell allows for agency from a position which would not otherwise provide any. The Blue Shell

20
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

allows a player who may not have the time or distance left to move out of a lower position to alter the
outcome in those higher positions without needing to work their way up there. While the enjoyment here
could be derived from a sense of schadenfreude or contributing to the fragile alliance of ‘first place vs.
everyone else’, it is still apparent that the enjoyment comes from a sense of control, even if that control is
not impactful or useful to the player.

The report also assumes the player’s intended goal remains constant (i.e. to win the race), yet this may not
be the case. Many games establish a hierarchy of winners vs. losers and use that to allow for gameplay to
continue even after the competition itself is over. Examples of this include recent games such as Titanfall 2,
where at the end of a match the losing team is tasked with successfully boarding an airship to escape the
battleground, and the winning team are tasked with defeating them before they can do this.[12] This
phenomenon can even be seen in other parts of MK64, such as when a player is defeated in the Balloon
Battle game mode, and can continue to interact with the competition by throwing bombs to hit the
competitors that are still in play.[12] In both of these instances, the outcome has been decided for the losing
players, but they are still being provided with some superficial objective to give them the illusion of agency
(and in the case of MK64, unearned agency). This illusion of agency may be the undefined qualia that makes
the Blue Shell so appealing, as it allows players who are in an unrecoverable position to still feel like they are
affecting the race, by specifically targeting the race leaders and disrupting the racer positioning which will
likely determine the winner. The Blue Shell assumes that a change in the racer’s mind-set has occurred, and
that they no longer view winning as achievable, or even their primary goal.

The Blue Shell could be viewed simply as an in-game expression of catharsis, where the players at the back
experience a shared sense of positive emotion that the winning racer is about to be impeded due to last
place’s poor performance, at the expense of first place’s sense of fear and anticipation. The strong emotion
felt by the person who fired the Blue Shell could be that of frustration or anger at their poor performance,
and this emotion is then conveyed and experienced by first place when they are hit by the Blue Shell,
allowing for some sort of shared empathy. This effect is more prominent in Mario Kart titles in which the
Blue Shell will not hit racers on its journey towards first place, such as Mario Kart Wii. It is this amalgamation
of excitement, frustration and panic which leads to a degree of ‘purging’, or a release of built-up emotions
by most participants.

By removing the Blue Shell from the game, no mechanical improvements are lost, but what is lost is this
emotional journey that it brings. While racers at the bottom more often than not stay at the bottom, the
Blue Shell allows those racers to feel that they have made some contribution to the race, and provokes some

21
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

positive emotional response. Conversely, it keeps racers at the front in a state of unease, keeping all racers
present in their mind, not just those directly behind them, and provokes an emotional response, which could
be positive or negative. The Blue Shell allows racers to remain relevant to all other racers, even when they
are not direct competitors during play. It allows all players to remain an active participant of the experience
whether or not they can win, or even if they are no longer actively attempting to win.

Mechanically, there are items which perform the expected function of the Blue Shell already present in the
series. In MK64, the Golden Mushroom allows a player to gain several consecutive speed boosts, pushing
them closer to the front, and the Lightning Bolt allows a player to briefly stun and shrink all other races,
pushing the user closer to the front by mutually slowing all other racers down. In future iterations such as
Mario Kart Wii and Mario Kart 8, the Bullet Bill gives a trailing racer an immediate boost directly up to the
other racers, directly improving on what the Golden Mushroom can do. All of these items still exhibit some
form of negative feedback loop, but they do so by giving the losing player some advantage (or
disadvantaging all other racers, an advantage to the user by proxy). They do not elicit the same emotional
response of the Blue Shell however, as none of these items mutually benefit every racer but one (first place).

The experiment run for this report had several unavoidable issues, which could be improved in future
research. As was mentioned in the Original Implementation section, the Blue Shell in MK64 has some
anomalous behaviour, as it can potentially hit every racer between the racer firing and first place. While this
effect was not observed frequently enough during experimentation to be taken into account, future
research may wish to alter the behaviour of the Blue Shell, or use a title in which the Blue Shell does not
possess this behaviour. The CPU racers are unable to fire Blue Shells or share item rotation parity with the
player, therefore the conclusions drawn here are more applicable to races with only one human player, and
not a group of human players. Future research could improve on this by finding some way to make the MK64
CPU fire these unfireable items, or by running the same experiment with several humans at once. Using
more current Mario Kart titles than MK64 would have been ideal, however ROM Hacking becomes
increasingly more complex as the titles become more modern, and as such, MK64 is currently the most
modern Mario Kart title with an adequate tool-set to perform this experiment.

The shape of Luigi Raceway also meant that the results gained from the experiments were approximated;
more exact results could be gained from running the experiment on a track which is a perfect circle or a
straight line (however doing so may mean the experiment is not being performed adequately, as participants
would never be playing Mario Kart on a track of this shape). The Rubber Band AI[13] is also a factor which
must be considered, as this means there is the potential that some CPU racers may speed up more than

22
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

should be possible in order to keep pace with the player, this is another factor which could be alleviated with
the introduction of more human players, or by altering the code to circumvent this (however in doing so, this
may invalidate the experiment, as the CPU would no longer be acting as expected).

This report has supported the argument that the Blue Shell does not have any mechanical impact in the
Mario Kart series. Originally conceived as a workaround for technical limitations, it seems likely the item has
maintained its place in the series due to familiarity, it is improbable that an item of its nature would be
introduced to the series today. The Blue Shell’s contribution to gameplay lies instead in the emotional
journey it provides, bringing joy to poorly performing players and angst to well performing players. The Blue
Shell provides a sense of equilibrium, much like a rollercoaster, by helping to ensure that the game provides
all players with the same set of emotional responses. By doing this, it allows the designers to keep the game
balanced, just not in the expected sense. The balance that is gained is that of a distribution of player agency,
reducing the agency of the players in front (as discussed in the Introduction) to increase the agency of the
players at the back, and ensuring that everyone feels they are still a part of the race up until the very end, as
was Konno’s intention[1].

While the designers may have initially implemented the item in an attempt to keep races tight, they
unintentionally added a mechanic that gives Mario Kart its unique, universal appeal. The Blue Shell may be a
functionless mechanic for racing games in general, but it has shown itself to be a meaningful mechanic
specifically in Mario Kart. The Blue Shell shows that with game design it can sometimes be more important
for a game to feel balanced than it actually is for that game to be balanced.

23
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

References

[1] The 30 Best Selling Super Mario games of all time - http://www.gizmodo.co.uk/2015/09/30-best-selling-
super-mario-games-of-all-time-on-the-plumbers-30th-birthday/
[2] Wendy Despain et al., “100 Principles of Game Design”, 2012
[3] Katie Salen and Eric Zimmerman, “Rules of Play”, 2004
[4] George Skaff Elias et al., “Characteristics of Games”, 2012
[5] “The Maker Of Mario Kart Justifies The Blue Shell”, Stephen Totilo, Kotaku, 2011
[6] “Love it or hate it, it sounds like Mario Kart’s Blue Shell is here to stay”, Martin Robinson, Eurogamer,
2017
[7] “Can Sports Data Analysis Figure Out Exactly How Much Mario Kart Screws You?”, Phil Iwaniuk,
Eurogamer, 2017
[8] Guinness World Record: Best Selling Racing Video Game - http://www.guinnessworldrecords.com/world-
records/best-selling-racing-video-game/
[9] Wolfram MathWorld: Distance - http://mathworld.wolfram.com/Distance.html
[10] Jean D. Gibbons, “Nonparametric Statistics: An Introduction, Issue 90”, 1993
[11] Mario Kart’s Most Hated Item Is Actually Its Best – Nathan Grayson, Kotaku, 2014
[12] Easy Allies Podcast #81 - October 11th 2017 - https://youtu.be/SJjBk0h_qTw
[13] Giant Bomb: Rubber Band AI - https://www.giantbomb.com/rubber-band-ai/3015-35/

24
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendices

Appendix 1 - Table showing the probabilities of human player receiving certain items whilst in various positions during Race 1.

Item Name Code 1st 2nd 3rd 4th 5th 6th 7th 8th Avg.

Single Banana 01 30% 4%

Multi Bananas 02 5% 5% 1%

Single Green Shell 03 30% 5% 4%

3 Green Shells 04 5% 10% 10% 3%

Single Red Shell 05 5% 15% 20% 15% 10% 8%

3 Red Shells 06 20% 20% 20% 20% 20% 20% 20% 18%

Blue Shell 07 5% 5% 10% 10% 15% 6%

Lightning Bolt 08 5% 5% 10% 10% 15% 20% 20% 11%

Upside Down '?' 09 10% 5% 2%

Star 0A 5% 10% 15% 15% 20% 30% 30% 16%

Ghost 0B 5% 5% 1%

Single Mushroom 0C 10% 5% 5% 5% 5% 4%

3 Mushrooms 0E 15% 20% 20% 25% 25% 10% 5% 15%

Gold Mushroom 0F 5% 10% 10% 10% 10% 10% 10% 8%

Total 100% 100% 100% 100% 100% 100% 100% 100% 100%

25
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 2 - Table showing the probabilities of human player receiving certain items whilst in various positions during Race 2, with
changes from Appendix 1 highlighted in red.

Item Name Code 1st 2nd 3rd 4th 5th 6th 7th 8th Avg.

Single Banana 01 30% 4%

Multi Bananas 02 5% 5% 1%

Single Green Shell 03 30% 5% 4%

3 Green Shells 04 5% 10% 10% 3%

Single Red Shell 05 5% 15% 20% 16% 11% 8.5%

3 Red Shells 06 20% 20% 20% 20% 22% 22% 23% 18.5%

Blue Shell 07 0% 0% 0% 0% 0% 0%

Lightning Bolt 08 5% 5% 11% 11% 17% 22% 23% 12%

Upside Down '?' 09 10% 5% 2%

Star 0A 5% 10% 16% 16% 22% 32% 33% 17%

Ghost 0B 5% 5% 1%

Single Mushroom 0C 10% 5% 5% 6% 6% 4%

3 Mushrooms 0E 15% 20% 20% 25% 27% 12% 8% 16%

Gold Mushroom 0F 5% 10% 11% 11% 12% 12% 13% 9%

Total 100% 100% 100% 100% 100% 100% 100% 100% 100%

26
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 3 - Table showing the probabilities of human player receiving certain items whilst in various positions during Race 3, with
changes from Appendix 1 highlighted in blue.

Item Name Code 1st 2nd 3rd 4th 5th 6th 7th 8th Avg.

Single Banana 01 30% 4%

Multi Bananas 02 5% 5% 1%

Single Green Shell 03 30% 5% 4%

3 Green Shells 04 5% 10% 10% 3%

Single Red Shell 05 5% 15% 20% 13% 9% 8%

3 Red Shells 06 20% 20% 18% 19% 16% 16% 13% 15%

Blue Shell 07 15% 15% 30% 30% 45% 17%

Lightning Bolt 08 5% 5% 9% 9% 11% 16% 13% 9%

Upside Down '?' 09 10% 5% 2%

Star 0A 5% 10% 13% 14% 16% 26% 23% 13%

Ghost 0B 5% 5% 1%

Single Mushroom 0C 10% 5% 5% 5% 4% 4%

3 Mushrooms 0E 15% 20% 18% 21% 21% 6% 3% 13%

Gold Mushroom 0F 5% 10% 9% 9% 6% 6% 3% 6%

Total 100% 100% 100% 100% 100% 100% 100% 100% 100%

27
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 4 – Competitive Proximity of all 150 races run across the 50 samples, alongside an ordered list or each races’ data. The
values which were excluded from the filtered data-set (due to the player not dropping below third) are struck out.

Race 1 Race 2 Race 3 Race 1 Race 2 Race 3

0 - EMcM 1236.041803 587.6775129 1310.206508 362.956842 372.8529304 386.4353283
1 - LC 616.3754913 1018.214217 1318.567969 369.532212 384.4879298 412.944186
2 - SL 1203.843392 623.6334063 1278.989442 414.2934991 423.4430151 534.2343631
3 - KMcC 362.956842 1319.84208 412.944186 418.9925219 575.2774953 572.0233606
4 - MC 418.9925219 1022.735933 1324.333707 461.5969098 585.3583833 609.0147767
5 - SS 989.3603589 1225.101972 955.0259982 509.9641797 587.6775129 623.6775482
6 - MMcM 1222.09085 1172.429497 1253.771749 564.716221 593.2274343 635.6807167
7 - OM 700.2179641 1045.448302 680.7632972 565.6839336 604.209628 647.3197369
8 - RF 775.5430314 1248.914728 623.6775482 613.1278001 608.7816339 654.8609308
9 - AAC 369.532212 1259.610294 1205.073337 616.3754913 615.1373372 680.7632972
10 - TRG 694.5385369 384.4879298 1219.21736 629.5696911 623.6334063 695.0967825
11 - AC 1222.473483 372.8529304 1197.079615 661.7502003 632.3910861 701.5452361
12 -NH 1220.536179 638.5951009 1314.40322 664.5179149 638.5951009 704.4915574
13 - AMcC 461.5969098 1371.489358 1014.673422 694.5385369 639.6401167 718.4603608
14 - SC 743.4096338 1257.567187 1143.753748 700.2179641 659.2440692 738.7331328
15 - RMcB 1090.96374 1018.680392 1154.774089 722.3853735 672.3155816 874.2124847
16 - CSA 1154.774089 593.2274343 1326.190497 740.8562145 695.3578042 937.9864198
17 - AR 664.5179149 750.5769401 1209.658611 743.4096338 750.5769401 950.3705042
18 - SY 1247.408397 639.6401167 534.2343631 775.5430314 807.6630768 955.0259982
19 - SG 1044.15766 1060.397264 1036.818028 801.6503172 991.6056919 1014.673422
20 - MY 661.7502003 807.6630768 1281.91202 843.3895461 1005.495641 1036.818028
21 - MMcQ 1405.074364 575.2774953 1305.046698 882.7290615 1011.926347 1040.73543
22 - LB 1459.102077 1357.256833 1040.73543 965.9353923 1018.214217 1043.655466
23 - ST 740.8562145 991.6056919 704.4915574 983.0633784 1018.680392 1110.50197
24 - EL 1190.058689 1011.926347 1224.381652 989.3603589 1022.735933 1125.231147
25 - JWB 983.0633784 1255.684125 937.9864198 1042.045302 1045.448302 1125.679464
26 - JC 1287.46733 1291.249097 738.7331328 1044.15766 1060.397264 1143.753748
27 - LS 1042.045302 1167.163766 701.5452361 1048.633259 1075.655439 1154.774089
28 - AMcM 509.9641797 423.4430151 718.4603608 1051.619377 1087.955776 1181.899266
29 - RP 1178.871586 695.3578042 1125.231147 1061.790033 1121.405231 1197.079615
30 - ED 1048.633259 632.3910861 1624.373712 1090.96374 1156.53021 1205.073337
31 - MC 1293.426859 608.7816339 1376.319638 1148.230161 1165.133182 1209.658611
32 - JM 1180.626047 1005.495641 1274.827105 1154.774089 1167.163766 1211.765102
33 - PM 613.1278001 615.1373372 609.0147767 1160.049961 1172.429497 1219.21736
34 - SM 565.6839336 1207.221345 647.3197369 1178.871586 1179.809346 1224.381652
35 - NP 882.7290615 1305.825155 635.6807167 1179.378888 1181.847826 1253.771749
36 - CF 564.716221 1179.809346 1272.282035 1180.626047 1186.497238 1272.282035
37 - SB 965.9353923 1165.133182 654.8609308 1190.058689 1199.430829 1274.827105
38 - CD 1312.239464 1199.430829 1321.047954 1203.843392 1207.221345 1278.989442
39 - AM 1160.049961 1075.655439 572.0233606 1220.536179 1225.101972 1281.91202
40 - MN 843.3895461 585.3583833 950.3705042 1222.09085 1235.857775 1305.046698
41 - CW 1501.362796 1186.497238 1125.679464 1222.473483 1248.914728 1310.206508
42 - EF 1051.619377 672.3155816 386.4353283 1236.041803 1255.684125 1314.40322
43 - CB 1061.790033 1156.53021 874.2124847 1247.408397 1257.567187 1318.567969
44 - MMcM 801.6503172 604.209628 1043.655466 1287.46733 1259.610294 1321.047954
45 - SM 722.3853735 1121.405231 695.0967825 1293.426859 1291.249097 1324.333707
46 - KG 414.2934991 1087.955776 1181.899266 1312.239464 1305.825155 1326.190497
47 - IH 1179.378888 1181.847826 1211.765102 1405.074364 1319.84208 1376.319638
48 - PMcM 1148.230161 659.2440692 1680.674896 1459.102077 1357.256833 1624.373712
49 - KC 629.5696911 1235.857775 1110.50197 1501.362796 1371.489358 1680.674896

28
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 5

Appendix 6

29
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 7

Appendix 8

30
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 9

Appendix 10

31
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 11

Appendix 12

Appendix 13

Appendix 14

32
 The Mechanical Impact of Mario Kart’s Blue Shell - Alex McMillan - 40104210

Appendix 15

Appendix 16

Appendix 17

33