Developing and investigating a quantitative method

to assess the difficulty of Formula 1 circuits

May 2024
 1. Introduction

As a huge fan of Formula 1 races, I am extremely familiar with many of the circuits

commonly used in tournaments. If one thing I noticed, it is that sudden and sharp curves are

the worst nightmare of drivers when it comes to safety, but at the same time a great

opportunity for overtaking. They are where most of the accidents occur, and that is because

they require the driver to slow down, something nobody wants to do in a race. From my

experience of watching the cars run and from basic intuition, it seems that the sharper the

curve is, the more one has to slow down in order to complete it safely, which makes it harder

for a circuit to be run the sharper curves it has. This made me wonder: what is the closest I

can get to measuring the difficulty of a circuit quantitatively?

This investigation aims to assess the difficulty of the Monza Formula 1 circuit.

However, since difficulty is subjective and does not rely on a single factor, the FIA

(Fédération Internationale de l’Automobile), association that manages Formula 1 races, does

not offer any kind of data that enables a concrete comparison between the results of this

exploration and a reliable parameter.

The method of the investigation will be to first derive the radius of curvature formula,

then to map the curves using functions graphed on Desmos Graphing Calculator. After that, a

point in each curve will be chosen and I will calculate the radius of curvature of each point.

In order to assess the difficulty of the chosen Formula 1 track, Monza, using this parameter, a

sequence of criteria will be developed and explained in further sections (3. a. more

specifically).

1
Background Information

a. Method for derivation

There are several ways to derive the formula of the radius of curvature. However,

applying it to the context of the aim can be difficult and confusing considering the concepts

involved in the process, and making it intuitive can be a challenge. Using the simplest and

most concise approach to the derivation of the formula might not feel connected to the

objective of this research. Therefore, for the sake of the logical follow-through, a

combination of different methods to deduce the radius of curvature will be combined.

b. Radius of curvature

Any curve can be modeled mathematically using functions or relationships that lead

to graphs. Each point of that graph (if the function is continuous and differentiable in all

points) has a gradient, a rate of change, which can be associated with a circle, and therefore,

has a radius, as shown in figure 1:

Figure 1. A circle inscribed in curve

Source: Graph made on Desmos, screenshot edited on GoodNotes by the candidate.

2
 However, although useful, the information about it might be hard to derive, requiring

more complex mathematical models. One useful property of circles inscribed in graphs is its

radius (radius of curvature), which has applications such as in determining if a curve in a lane

is safe enough for use. The relationship between radius and safety comes down to physics

concepts such as friction and centripetal force, which are not in the scope of this investigation

and hence will not be explored. Still, mathematical models to determine the radius of a

curvature are and hence will be explored throughout this research. However, the formula that

calculates them is not exactly on the purview of the Mathematics Analysis and Approaches

syllabus, and therefore will be derived before being applied.

a. Deriving the formula and explaining

The first concept that needs to be brought before getting to the radius of a curvature,

is actually what the curvature is. Although it can be approached in a pretty simple way when

discussing circles, it goes far beyond its basic properties when studying differential geometry

and generalized curves, which is the case for this exploration. The derivation is adapted from

MATHalino website1, Robert Ferguson’s article2, and Less Boring Lecture’s youtube video3.

What I judged to be the best approach to the radius of curvature for this exploration

would be its relationship with the change in direction of a curve, since car race tracks are

being discussed. This can be explained using the concept of an arc, a fraction of a

circumference, modeled by the equation 𝑠 = 𝑟 × θ, where s = arc length (or the

displacement of the driver), r = radius, and θ = angle covered by s in radians. It is important

1
“Curvature and Radius of Curvature | Differential Calculus Review at MATHalino.” MATHalino,
mathalino.com/reviewer/differential-calculus/curvature-and-radius-curvature. Accessed 24 Feb. 2024.
2
Ferguson, Robert. “An Easier Derivation of the Curvature Formula from First Principles.” Australian Senior
Mathematics Journal, vol. 32, files.eric.ed.gov/fulltext/EJ1231087.pdf. Accessed 24 Feb. 2024.
3
“Radius of Curvature Equation Derivation.” Www.youtube.com, Less Boring Lessons, 2021,
youtu.be/r3iRNeXdYlc?si=r6GED6zecQLdnJPo. Accessed 24 Feb. 2024.

3
to understand that as the radius gets bigger, the opposite happens to the angle for a fixed arc

length, meaning a smaller fraction of the circle will be covered and less change of direction

will happen. This means that for a fixed radius, the change in arc length will result in a

change in angle, and vice-versa, which can be expressed as Δs = 𝑟 × Δθ. However, we are

looking for small changes in s, in fact, infinitesimally small changes. which brings us to

express them as d, as done in differential calculus, leading to the equation 𝑑𝑠 = 𝑟 × 𝑑θ.

In the function of a curve, each small unit traveled in the x-axis can correspond to a

change in arc length, and the radius associated with that will determine the rate at which it

changes. For instance, a fixed path s along a curve can correspond to a different change in

direction (or tangent t) depending on the size of the radius r that models the curve being

traveled. The diagram below demonstrates that visually.

Figure 2. Diagram of change in curvature with radius

Source: screenshot from GoodNotes, drawing made by candidate

4
 The arc of the smallest correspondent circle inscribed in a function at any point is

tangent to the graph at a point, meaning an infinitesimally small line, which would essentially

be lim 𝑦 where y = function that models the curve for any point a on the curve.
𝑥→𝑎

However, because of the infinitesimal behavior of a limit, the arc around any point on

a function can be approached as a small straight line if it is “sufficiently” zoomed in, and be

used as the hypotenuse of a right triangle, as Figure 3 and the equation below demonstrate:

Figure 3. Diagram of a small arc around a point

Source: a screenshot of the graph from Desmos annotated in GoodNotes by the candidate

𝑑𝑠 = (𝑑𝑥)² + (𝑑𝑦)² (Pythagorean theorem)

From that equation, (dx)² can be put as a common factor and then factored out of the

square root:

𝑑𝑠 = 𝑑𝑥²(1 + ( )²
𝑑𝑦
𝑑𝑥

𝑑𝑠 = 𝑑𝑥 × 1 + ( )²
𝑑𝑦
𝑑𝑥

5
 Now, we need to find an expression for the change in angle for any displacement to

connect all the parts of the derivation.

First, consider the angle θ to form an arc that models the displacement from position 1

(p₁) to position 2 (p₂) in a curve. Notice how this angle is the same as the angle formed

between the tangent to the curve at position 1 (t₁) and position 2 (t₂):

Figure 4. Diagram demonstrating angle complementariness in sector and tangent visually

Source: Screenshot from GoodNotes, drawing made by candidate

Therefore, any change in position, ds, making an arc angle θ with the original

position, will also form the same angle θ between the initial and final direction (represented

by the tangent) being followed by the curve on the interval s. Recurring back to the

𝑑𝑦
infinitesimally small triangle from Figure 3, 𝑑𝑥
is at the same time the gradient at the point

ds develops around, and the tangent of the angle formed between dx and ds, which as already

𝑑𝑦
proved, is θ. Therefore, it can be expressed in calculus notation as tanθ = 𝑑𝑥
.

6
 Since we want to find the change in the slope related to a change in displacement, we

can differentiate both sides of the equation with respect to x, which gives:

𝑑
𝑑𝑥
𝑡𝑎𝑛θ =
𝑑
𝑑𝑥 ( )
𝑑𝑦
𝑑𝑥

Using the chain rule and implicit differentiation, where θ is the inner function of

tan(x):

𝑑θ 𝑑²𝑦
(1 + 𝑡𝑎𝑛²θ) × 𝑑𝑥
= 𝑑𝑥²

𝑑𝑦
As already stated, tanθ = 𝑑𝑥
, and therefore, we can substitute in the equation:

(1 + ( )²) ×
𝑑𝑦
𝑑𝑥
𝑑θ
𝑑𝑥
=
𝑑²𝑦
𝑑𝑥²

𝑑²𝑦
𝑑θ 𝑑𝑥²
𝑑𝑥
= 𝑑𝑦
(1+ ( 𝑑𝑥 )²)

𝑑²𝑦
𝑑𝑥²
𝑑θ = 𝑑𝑦 × 𝑑𝑥
(1+ ( 𝑑𝑥 )²)

Recurring back to the equation 𝑑𝑠 = 𝑟 × 𝑑θ, we can substitute the expressions

obtained for ds and dθ, which leads to the following equation:

7
 𝑑²𝑦
𝑑𝑦 𝑑𝑥²
𝑑𝑥 × 1 + ( 𝑑𝑥 )² = 𝑟 × 𝑑𝑦 × 𝑑𝑥
(1+ ( 𝑑𝑥 )²)

𝑑𝑦
𝑑𝑥× 1+( 𝑑𝑥 )²
𝑟 = 𝑑²𝑦
𝑑𝑥²
𝑑𝑦 ×𝑑𝑥
(1+ ( 𝑑𝑥 )²)

Simplifying like terms, we get the following formula:

3
𝑑𝑦 2
(1 + ( 𝑑𝑥 )²)
𝑟 = 𝑑²𝑦
𝑑𝑥²

However, this format does not disallow the radius to be negative, which would not be

mathematically acceptable since it represents a distance, so in order to guarantee that the

radius is always positive (or else, finding its magnitude), the modulus of the formula is taken,

leaving us with:

3
| (1 +
𝑑𝑦
( 𝑑𝑥 )²) 2 |
𝑟 = || 𝑑²𝑦
|
|
| 𝑑𝑥² |

Another reflection regarding the choice of the method is why not to use the concept of

1
curvature, often denoted as κ, instead of the radius. κ is essentially equal to 𝑟
, and

derivations that follow other paths often introduce this concept. The reason for the use of r is

because I believe that understanding the method through a familiar concept, the radius, is

simpler than introducing and explaining a new one.

8
 3. Application

a. Introduction to the circuit, method and criteria

In Figure 2, notice how for the same arc length s, the increase in radius r makes the

change in tangent t much smaller, which explains how the curvature is inversely proportional

to the radius, meaning that the smaller the radius, the faster the tangent changes. Now

applying this to a real life context, the relationship between the radius of curvature and the

difficulty of a curve can be traced as the change in the direction of the curve, for a fixed

distance gets slower with a bigger radius, meaning that the bigger the radius, the easier the

curve is to be drove on (considering the physics behind the circular motion and not any

probabilistic event). Therefore, the difficulty of the curve being analyzed in this exploration

will be assessed based on that.

In order to determine the difficulty of a circuit with the mentioned parameters, after

picking one (Figure 5), it is necessary to model the track as a function so that it is possible to

analyze its curves mathematically. For such to be done, a series of functions are going to be

matched with the image of the track using the Desmos Graphing Calculator individually.

Through that, it is possible to use the function obtained to apply the formula derived in the

previous section and, from that, assess the difficulty of the track based on criteria that will be

explained in this section

Figure 5 shows the Monza circuit, which will be analyzed and already labels each

curve from 1 to 11. These labels will be the ones used to define the curves as they are the

official labels.

9
 Figure 5. Monza Formula 1 circuit

Source: downloaded from the Formula 1 website

DRS detection zones are designated segments where drivers can activate the drag

reduction system (DRS), making the cars less susceptible to resistive forces from the air. This

mechanism is out of the scope of this investigation and hence will not be explored. The speed

trap is where the speed of the driver is recorded in each lap. The speed will be relevant to the

discussion in further sections but the effect of the position of the speed trap is out of scope.

The criteria for determining difficulty could simply be the sum of the radii of

curvature of each circuit. Nonetheless, it is necessary to note that a bigger number of curves

would correspond to a higher difficulty; however, it would also mean a bigger sum of the

radii of curvature, which would contradictorily mean lower difficulty. Therefore, I found it

1
best to sum the inverse of the radii of curvature, 𝑟
so that a higher value also corresponds to

a harder track. This leaves us with the formula:

1
𝑑= Σ 𝑟

10
where d is the numerical value attributed to difficulty through the method developed in this

investigation.

b. Graphing the curves

In order to graph the track, I identified that most curves could be easily approximated

using rational functions of the form.

𝑎𝑥+𝑏
𝑐𝑥+𝑑

Or, rational functions of the form

𝑎𝑥²+𝑏𝑥+𝑐
𝑐𝑥+𝑑

when needed, forming oblique asymptotes of the form 𝑝𝑥 + 𝑞, a line equation that comes

after dividing the numerator by the denominator using long division. Their vertical and

horizontal asymptotes were also moved around by changing the undefined values of the

functions (when the denominator is 0) and translating them by vertical vectors, which were

then incorporated into the nominator.

The sharpness of each curve was modified more arbitrarily through the leading

coefficients of the numerator and denominator polynomials. This was done to match the

specific scale at which the image was annexed in the software, (Desmos), which was kept

constant.

Additionally, notice that no implicit relationships were used to map any curve that

failed the vertical line test in order to ensure differentiability at any point and not to

overcomplicate calculations, as simply rotating the images about the y-axis would have no

effect on the radius of curvature at the specific points.

11
 After graphing all the curves, the following functions were obtained, as displayed in

Table 1 (notice that 4 significant figures were used for higher precision even though it

pollutes the table visually):

Table 1. Functions that model each curve.

Function Curve
−1570𝑥−1522 1
1000𝑥+970.0

2000𝑥²−3000𝑥−4649 2
5000𝑥+4750

−4.700𝑥−15.20 3
2.500𝑥+8.650

260.0𝑥+870.0 4
280.0𝑥+936.0

242.0𝑥+872.0 5
280𝑥+1008

162.8𝑥+650.0 6
70𝑥+280

90.00𝑥²+898.0𝑥+2.000 7
440.0𝑥

40.00𝑥²+1.000 8
100𝑥

− 1. 440𝑥² − 0. 2160𝑥 − 0. 3741 9

−123.0𝑥+25.60 10
200.0𝑥−40.00

− 1. 690𝑥² + 3. 302𝑥 + 2. 417 11

Source: Google documents. Made by this investigation’s author.

The superposition of the functions with the circuit in Desmos will not be shown as

some of them required the rotation or translation of the image rather than the function,

making them not clearly visible together as the image would be placed in a single position.

Most importantly, it is valid to remember that this does not affect the radius of curvature

whatsoever as the actual sharpness of the curve is not being altered.

12
 Also notice that curves 9 and 11 are actually quadratic functions. The reason for this

is that these curves, in particular, were better curved using parabolas, not rational functions as

stated formerly.

Applying the radius of curvature formula requires a point in each graph to be chosen,

and the coordinate will be visually determined as the approximated apex of each curve,

around where the driver would likely need to change speed the most. The coordinates are

presented in Appendix 1 for concision issues.

c. Calculating the radius of curvature

Now that all curves have their respective functions mapped, the radius of curvature

formula needs to be applied so that each value is obtained and the difficulty can be assessed.

For the sake of concision, because the formula requires a fairly long process to be

used, only the first curve’s radius will be calculated step-by-step in this investigation, and the

same process will be applied to the other ones, whose values will simply be stated.

Recalling, the function of curve 1 (which will from now on be denoted as 𝐶1(𝑥)) was

given by:

−1570𝑥−1522
𝐶1(𝑥) = 1000𝑥+970.0
(I)

And the formula

3
𝑑𝑦 2
(1 + ( 𝑑𝑥 )²)
𝑟 = 𝑑²𝑦 (II)
𝑑𝑥²

13
 The first step is to find the derivative and second derivative of (I), which can be most

easily done using the quotient rule:

−1570𝑥−1522
𝐶1(𝑥) = 1000𝑥+970.0

𝑑𝑦 𝑑𝑦
𝑑𝐶1 (1000𝑥+970.0)× 𝑑𝑥 (−1570𝑥−1522) −(−1570𝑥−1522)× 𝑑𝑥 (1000𝑥+970.0)
𝑑𝑥
= (1000𝑥+970.0)²

The derivatives inside the rule can be done using the power rule (which will just bring the

variable x to the 0 power, since all are just raised to the first power, and get rid of the constant

terms), and the result can be multiplied by the multiplying non-differentiated functions. The

denominator function will not have its square expanded for the coefficients are very large and it

is not necessary:

𝑑𝐶1 (1000𝑥+970.0)×(−1570)−(−1570𝑥−1522)×(1000)
𝑑𝑥
= (1000𝑥+970.0)²

𝑑𝐶1 (−1570000𝑥−1522900)−(−1570000𝑥−1522000)
𝑑𝑥
= (1000𝑥+970.0)²

𝑑𝐶1 −900.0
𝑑𝑥
= (1000𝑥+970.0)²
(III)

For the second derivative, it becomes easier to consider the denominator a negative

power function, with which we can perform the chain rule:

𝑑𝐶1 −2
𝑑𝑥
= − 900. 0 × (1000𝑥 + 970. 0)

𝑑²𝐶1 −3
𝑑𝑥²
= (− 900. 0) × 1000 × (− 2. 000) × (1000𝑥 + 970. 0)

14
 And then back to rational form. The denominator will be kept unexpanded because

the coefficients are extremely large:

𝑑²𝐶1 1800000
𝑑𝑥²
= (1000𝑥+970.0)³
(IV)

The next step is to plug all functions obtained into (II), so I will get the numerator
3

ready by calculating (1 + (𝐼𝐼𝐼)²) 2

:

(1 + ( (−900)
(1000𝑥+970.0)² )²) 2

(1 + )
3
810000 2
4
(1000𝑥+970.0)

( )
4 3
(1000𝑥+970.0) + 810000 2
4
(1000𝑥+970.0)

Now the Wolfram Alpha software will be used to find the final numerator function to

be plugged in equation (II):

4 4
((1000𝑥+970.0) )
+810000 × (1000𝑥+970.0) +810000
(V)
6
(1000𝑥+970.0)

And then plug (V) and (IV) into (II):

4 4
((1000𝑥+970.0) )
+810000 × (1000𝑥+970.0) +810000
6
(1000𝑥+970.0)
𝑟= 1800000
(1000𝑥+970.0)³

15
 𝑟=
(( 4
) 4
)
(1000𝑥+970.0) +810000 × (1000𝑥+970.0) +810000 ×(1000𝑥+970.0)³
6
(1000𝑥+970.0) × 1800000

𝑟=
(( 4
) 4
(1000𝑥+970.0) +810000 × (1000𝑥+970.0) +810000 ) (VI)
(1000𝑥+970.0)³ × 1800000

And the last step is now to plug the x-coordinate for curve 1 (found in Appendix 1)

into (VI) using again my TI-84 Graphic Display Calculator (GDC):

𝑟=
(( 4
)
(1000×(−0.9300)+970.0) +810000 × (1000×(−0.9300)+970.0) +810000
4
)
(1000×(−0.9300)+970.0)³ × 1800000

𝑟 = 0. 05370

The remaining radii of curvature are presented in table 2:

Table 2. Radii of curvature of each curve from Monza circuit

Radius of curvature Curve

0.05370 1

0.02752 2

1.027 3

0.08158 4

0.08938 5

0.2216 6

0.07332 7

0.07926 8

0.4038 9

0.1592 10

0.2959 11
Source: Google documents. Made by this investigation’s author.

16
 d. Applying values to criteria and assessing accuracy

In order to assess the accuracy of this method, there need to be forms of comparison

that allow one to say what circuits are generally seen as more difficult. However, as stated in

the introduction, there is no concrete way to determine difficulty for means of comparison.

Still, a method that is commonly aligned with how circuits tend to be perceived as difficult is

related to their average speed, which might allow us to have an idea of how much drivers

need to brake and make curves, which might enhance their challenge to win a race. The

choice of this parameter is especially suitable because, like the radius of curvature, it is also

related to how much the speed of drivers need to decrease while in curves.

The other circuit that will be used for comparison with Monza (Italian Grand Prix) is

the Interlagos (Brazilian Grand Prix), and the exact same process that was done with Monza

will be done with it, except that it will not be shown for concision issues (functions for each

curve can be found in Appendix 2 and the radius of curvature found for each curve in

Appendix 3). Since both images were downloaded from the same source (which is also an

official source), the scale used in Desmos will be the same. Interlagos is shown in figure 8:

17
 Figure 8. Interlagos circuit

Source: downloaded from the Formula 1 website

Starting with Monza, it has an average lap time of 159.892 miles per hour, while Interlagos

has one of 136.620 miles per hour (FIA), indicating that, through this criterion, Interlagos would be

considered more difficult than Monza as it requires that drivers slow down more or accelerate less.

1
Now calculating d, for Monza, by using 𝑑 = Σ 𝑟 , we obtain:

1 1 1 1 1 1 1 1 1 1 1
0.05370
+ 0.02752
+ 1.027
+ 0.08158
+ 0.08938
+ 0.2216
+ 0.07332
+ 0.07926
+ 0.4038
+ 0.1592
+ 0.2959

= 18. 62 + 36. 34 + 0. 9737 + 12. 26 + 11. 19 + 4. 512 + 13. 64 + 12. 62 + 2. 476 + 6. 281 + 3. 380

𝑑 = 121. 3

This result is still meaningless until we can compare it to another one, so, for Interlagos:

1 1 1 1 1 1 1 1 1 1 1
0.1724
+ 0.1724
+ 1.051
+ 0.1607
+ 1.311
+ 0.2326
+ 0.5078
+ 0.1025
+ 0.1919
+ 0.8333
+ 0.5556

1 1 1 1
+ 0.1379
+ 1.351
+ 1.291
+ 2.635

18
= 5. 800 + 5. 800 + 0. 9515 + 6. 223 + 4. 299 + 1. 969 + 9. 756 + 5. 211 + 1. 200 + 1. 800

+ 7. 252 + 0. 7402 + 0. 7746 + 0. 3795

𝑑 = 43. 01

From the calculations, Monza shows a significantly higher difficulty index d, almost three

times the one of Interlagos, meaning that through this method the Monza circuit should present more

difficulty to pilots than Interlagos would. However, these results go against the parameter defined for

comparison, since Interlagos actually shows a higher average speed.

4. Conclusion and Evaluation

In conclusion, the results show that the method created does not accurately reflect the

difficulty of a formula 1 circuit as the values obtained go against the parameter used for

comparison, which was carefully chosen as the most suitable one for the method of the

exploration. Therefore, it was not possible to accurately quantify a numerical value for the

difficulty through this method in comparison to the parameters thereof.

A possible reason for this is that even though suitable for this investigation, the

average speed is still not a precise indicator of difficulty either, as difficulty is extremely

subjective. Even though this exploration attempts to mathematically address it, its method is

very specific to one of the many difficulties racers face; sharp curves. Most importantly, the

average speed is paramountly increased when a curve has long straight paths, which is the

case of Monza, but not for Iterlagos, which might explain the opposingness of the methods.

This factor is not accounted for in the radius of curvature method, thus possibly giving origin

to the disparity between the measures. An accurate method to quantify difficulty would also

19
have to account for a much larger number of factors, which still would be subjective

depending on specific skills and preferences of each driver. Examples of factors are the

number of laps (hence fatigue), specific weather conditions associated with location, asphalt

conditions, and other case-specific factors that demand huge work to be mathematically

mapped.

The assessment of the accuracy of the method in comparison with another more

widely used one is also a limitation of the method since it uses only two circuits to compare.

A solution could have been simply enlarging the sample, even though that would take a lot

more work.

There are a range of other limitations faced by the method developed: regarding the

mapping of the curves, the functions used do not completely match the track of the curves. In

addition, the inclinations and elevation changes, which play a big role in attributing difficulty

to a curve, were not accounted for (which could have been explored if this work disposed

itself to go well beyond the Mathematics Analysis and Approaches syllabus, as well as to

explore aspects related to physics).

Not only that, but there are infinitely many ways in which a curve can be run, and this

goes from pilot to pilot and what they judge to be the most effective way to get through it.

This investigation did not account for this. Additionally, the method does not account for how

close to each other the curves are and how sudden a change in direction might feel to the

driver, which appears as a limitation.

Now, regarding the method rather than the results, the radius of curvature function

could have been deduced in a more concise way. However, I believe that the chosen pathway

is the one that fits the purposes of this investigation the best, working more intuitively to

connect all the steps, and was developed based on a range of sources.

20
5. Bibliography

2021 FORMULA 1 SPORTING REGULATIONS. Fédération Internationale de l’Automobile,

31 Oct. 2019.

“Curvature and Radius of Curvature | Differential Calculus Review at MATHalino.”

MATHalino,

mathalino.com/reviewer/differential-calculus/curvature-and-radius-curvature.

Accessed 24 Feb. 2024.

“Desmos | Graphic Calculator.” Desmos, www.desmos.com/calculator?lang=pt-BR. Accessed

24 Feb. 2024.

“Dutch Grand Prix - F1 Race - Zandvoort Circuit | Formula 1®.” Formula 1, Fédération

Internationale de l’Automobile,

www.formula1.com/en/information.netherlands-zandvoort-circuit-zandvoort.6XdtPTI

MZzx5wLKP9mm7Ev.html. Accessed 24 Feb. 2024.

Ferguson, Robert. “An Easier Derivation of the Curvature Formula from First Principles.”

Australian Senior Mathematics Journal, vol. 32,

files.eric.ed.gov/fulltext/EJ1231087.pdf. Accessed 24 Feb. 2024.

“Formula 1 2022 Season Tracks.” Sportmonks,

www.sportmonks.com/formula-one-api/formula-1-tracks/. Accessed 24 Feb. 2024.

“Italian Grand Prix - F1 Race - Autodromo Nazionale Monza | Formula 1®.” Formula 1,

Fédération Internationale de l’Automobile,

21
 www.formula1.com/en/information.italy-autodromo-nazionale-monza.FiJN1jnQlRLe

HqOxIt13m.html. Accessed 24 Feb. 2024.

“Radius of Curvature Equation Derivation.” Www.youtube.com, Less Boring Lessons, 2021,

youtu.be/r3iRNeXdYlc?si=r6GED6zecQLdnJPo. Accessed 24 Feb. 2024.

Wolfram Alpha. “Wolfram|Alpha: Making the World’s Knowledge Computable.”

Wolframalpha.com, 2023, www.wolframalpha.com/. Accessed 24 Feb. 2024.

22
 6. Appendices

Appendix 1 - Table with coordinates of each curve’s chosen points in Monza circuit

Table 2. Coordinates of Monza’s curves’ chosen points

Curve X-coordinate

-0.9300 1

-0.9820 2

-2.670 3

-3.392 4

-3.532 5

-3.898 6

-0.06000 7

0.09300 8

0.03800 9

0.3070 10

0.9770 11
Source: Google documents. Made by this investigation’s author.

23
Appendix 2 - Interlagos circuit functions and points used

Table 3. Function of each curve of Interlagos circuit

Function Curve

− 2. 900𝑥² + 6. 380𝑥 + 0. 8110 1

2. 900𝑥² − 3. 070𝑥 + 4. 720 2

−
𝑥²−3.000𝑥−50.00 3
99.00𝑥−9.000

−
29.00𝑥²−2.000𝑥+3.000 4
98.00𝑥+1.000

−
13.00𝑥²−𝑥−50.00 5
90.00𝑥−9.000

−
8990𝑥²−3.000𝑥−50.00 6
13000𝑥−9.000

1.000 7
8.000𝑥

−
40.00𝑥²+1.900 8
50.00𝑥+4.000

24.00𝑥²−2.000𝑥+4.000 9
40.00𝑥

-6.000x²-1.800x+2.750 10

0. 9000𝑥² + 0. 7740𝑥 + 1. 710 11

−
8.000𝑥² 12
12.00𝑥+3.000

− 0. 3700𝑥² − 0. 0860𝑥 + 3. 540 13

1.000 14
1.200𝑥

− 0. 1900𝑥² + 0. 0760𝑥 + 2. 440 15

Source: Google documents. Made by this investigation’s author.

Notice that in this circuit, I mostly chose to translate (and rotate) the image rather than

the function for it required less work in most cases.

24
Table 4. Points used to calculate the radius of curvature of each curve in Interlagos circuit

x-coordinate Curve

1.100 1

0.5290 2

0.8900 3

-0.1860 4

-0.6200 5

0.0590 6

-0.3800 7

0.0900 8

0.2830 9

-0.1500 10

-0.4300 11

-0.4060 12

-0.1160 13

1.080 14

0.2000 15
Source: Google documents. Made by this investigation’s author

25
Appendix 3 - Interlagos circuit radii of curvature

Table 4. Radii of curvature of each curve from Interlagos circuit

Radius of curvature Curve

0.1724 1

0.1724 2

1.051 3

0.1607 4

1.311 5

0.2326 6

0.5078 7

0.1025 8

0.1919 9

0.8333 10

0.5556 11

0.1379 12

1.351 13

1.291 14

2.635 15
Source: Google documents. Made by this investigation’s author

26