Math AI SL Internal Assessment

Tire Degradation Impact on Lap Time in Formula 1

Yasmine Armstrong - M25

Total Page Number:

Introduction

Formula 1 (F1) is the highest class of single-seat auto racing in the world. It is one of

the most technically developed motorsports, where every single detail can influence the outcome of a

race. A variety of mathematics is used to determine these details - leading to the best optimal

outcome. One key factor which influences a car's performance is the tires, where grip is essential in

high-speed racing in various climates. Over the course of a race, tires gradually degrade due to the

combined intensity of heat, speed and track conditions, leading to reduced performance.

Being a F1 fan myself, one thing that I’ve always admired is the smooth pit-stops teams

make during the race to switch out their tires or perform mechanical repairs. Tires are always switched

out during pit-stops at incredible speed and precision, and thus, when I was given the opportunity to

explore a topic that interests me, I wanted to explore how tire degradation impacts lap time.

Throughout F1 races, teams have to decide when is the most appropriate time to switch tires,

otherwise they run the risk of slowing down and losing valuable points. These split-second decisions

play a large role in the outcome of a race, and a difference of a couple seconds can be the difference

between first place and second. Tires have to be switched out because of tire degradation, which is

where tires lose their durability and grip due to a variety of factors (heat, speed, and friction), and

switching out tires is essential for optimising lap time. F1 regulations state that teams must use two

different types of tires during dry races, adding another element to tire management. There are a

variety of different types of tires that teams can use during dry and wet conditions. During dry races,

teams can use soft, medium or hard tires - each offering unique compounds. Soft tires provide

maximum grip and fastest lap times, however they degrade quickly due to the soft rubber, medium

tires offer a balance between performance and durability and provide consistent performance, and

hard tires are the most durable, but take longer to heat up therefore offering less grip1.

Thus, my aim is to model the effect of tire degradation on lap times in the 2021 British

1
“Pirelli Motorsport: Best Products for Motor Competitions.”
Grand Prix. The race is known for its fast corners, and is abrasive and demanding on tires, making it

appropriate for investigating how tires impact lap time. I will collect data on individual drivers and

their long stints (which is a prolonged period of time where a driver used one type of tire), and the tire

compounds they used. After my analysis and evaluation, I will be able to plot the models and find the

trend of lap times as tires degrade and interpret which had the largest impact on lap time.

Exploration

The 2021 British Grand Prix took place in the iconic Silverstone Circuit, known as the

home of F1 racing. The Silverstone Circuit is made up of 18 turns, and is a fast-paced circuit with a

variety of different fast corners. The high-paced race and abrasive surface put significant stress on the

tires, making tire management a key part of the race2. Therefore, I decided that this race was an

excellent choice to use.

I never knew that tires played such an important role in the performance of a car, and

throughout my investigation, I was surprised at the importance of tire strategy in order to maximise

outcomes. In order to calculate tire degradation, I analysed the change in lap times over consecutive

laps within a tire stint, using the number of laps completed on a specific set of tires as an estimate for

the level of degradation. After this, I used scatter graphs to plot the lap times vs the number of laps on

a set of tires, for each type of tire (soft, medium and hard). I then used three different mathematical

models; linear regression, polynomial regression, and exponential model, then calculated the best fit

model, thus finding the impact of tire degradation on lap time.

Analysis

My first step is to gather data on the 2021 British Grand Prix. I used racing sources in

2
“FORMULA 1 PIRELLI BRITISH GRAND PRIX 2021.”
order to find the lap-by-lap time for three drivers: Lewis Hamilton, Lando Norris, and Carlos Sainz. I

chose these three drivers because they represent the top three teams (Mercedes, Mclaren, and Ferrari),

all with different tire strategies and cars. These drivers also finished across varied times, Hamilton in

first, Norris in fourth and Sainz in the top ten - all drivers scoring points for their teams. I used lap

times from the beginning of the stint (Laps 5-9), middle (Laps 19-21), and end (Laps 25-27).

Driver Lap Number Tire Lap time (s) Stint Number Tire Age
Compound

Hamilton 5 Medium 93.049 2 1

Hamilton 6 Medium 93.113 2 2

Hamilton 7 Medium 92.738 2 3

Hamilton 8 Medium 92.928 2 4

Hamilton 9 Medium 93.013 2 5

Hamilton 19 Medium 92.597 2 15

Hamilton 20 Medium 92.079 2 16

Hamilton 21 Medium 92.297 2 17

Hamilton 25 Medium 92.115 2 21

Hamilton 26 Medium 91.854 2 22

Hamilton 27 Medium 92.852 2 23

Table 1 - Lewis Hamilton Stints Data Table

Driver Lap Number Tire Lap time (s) Stint Number Tire Age
Compound
Norris 5 Medium 93.318 2 1
Norris 6 Medium 93.428 2 2
Norris 7 Medium 93.556 2 3
 Norris 8 Medium 93.402 2 4
Norris 9 Medium 93.373 2 5
Norris 10 Medium 93.011 2 6
Norris 17 Medium 92.586 2 13
Norris 18 Medium 92.52 2 14
Norris 19 Medium 92.586 2 15
Norris 20 Medium 92.293 2 16
Norris 21 Medium 91.418 2 17
Table 2 - Lando Norris Stints Data Table

Driver Lap Number Tire Lap time (s) Stint Number Tire Age
Compound
Sainz 5 Medium 93.782 2 1
Sainz 6 Medium 94.026 2 2
Sainz 7 Medium 93.807 2 3
Sainz 8 Medium 93.937 2 4
Sainz 9 Medium 93.915 2 5
Sainz 19 Medium 93.344 2 15
Sainz 20 Medium 92.73 2 16
Sainz 21 Medium 92.457 2 17
Sainz 25 Medium 91.855 2 21
Sainz 26 Medium 91.819 2 22
Sainz 27 Medium 92.26 2 23
Table 3 - Carlos Sainz Stints Data Table

I created three tables with all the data needed. I organised it by driver, adding the lap

number, tire compound (the type of tire which was used during the lap, which in this case is medium),

lap time in seconds for each driver, and the stint number. Additionally, I took data from all three

points of the stint (beginning, middle and end), to see the impact of tire degradation throughout an

entire stint. From this data I found the tire age, which is the number of laps completed on a certain set

of tires. When a driver switches to a new set of tires after a pit-stop, the tire age starts at 1 for the first

lap done on that set, and increases by increments of 1 for each subsequent lap.
Degradation Rate 1

Firstly, I will calculate the degradation rate of each lap time. This will help to determine how quickly

tires lose performance as they age. This is essential in identifying any patterns and modelling the

relationship between tire age and lap time. The degradation rate (𝐷) is the change in one lap time to

another:

𝐷 = 𝐿𝑎𝑝 𝑇𝑖𝑚𝑒𝑛 − 𝐿𝑎𝑝 𝑇𝑖𝑚𝑒𝑛−1

𝐿𝑎𝑝 𝑇𝑖𝑚𝑒𝑛 is the lap time for the current lap

𝐿𝑎𝑝 𝑇𝑖𝑚𝑒𝑛−1 is the lap time for the previous lap

Taking an example from Table 1 (Hamilton), I will calculate the degradation rate of lap 26 and 27.

𝐿𝑎𝑝 26: 91. 854

𝐿𝑎𝑝 27: 92. 852

𝐷 = 92. 852 − 91. 854 = 0. 998 𝑠𝑒𝑐𝑜𝑛𝑑𝑠 𝑝𝑒𝑟 𝑙𝑎𝑝

Now, I will calculate the degradation rate for each lap. To ensure clarity, I will be focusing primarily

on Hamilton’s data, then comparing it to the other two drivers. Below, I have calculated the

degradation rate for each of the laps in Table 1:

Driver Lap Number Lap Times Degradation Rate

Beginning of Stint (Medium tires)

Hamilton 5 93.049 0

Hamilton 6 93.113 0.064

 Hamilton 7 92.738 -0.372

Hamilton 8 92.928 0.19

Hamilton 9 93.013 0.085

Middle of Stint (Medium tires)

Hamilton 19 92.597 0.049

Hamilton 20 92.079 -0.518

Hamilton 21 92.297 0.218

End of Stint (Medium tires)

Hamilton 25 92.115 -0.571

Hamilton 26 91.854 -0.261

Hamilton 27 92.852 0.998

Table 4 - The degradation rate on each lap (Hamilton)

While the degradation rate was expected to remain positive, some laps had negative rates. When I

think back to the F1 races I’ve watched, I recall that other factors may lead to a faster lap then the

previous one. One of these may be drivers' performance, and Lewis Hamilton is known to be a driver

always pushing to the max and attempting to get the fastest lap record, resulting in faster lap times as

the race progresses. Other factors may also include tire warm-up (when a tire reaches optimal

temperature, and thus, performance), track evolution, and fuel weight burn off - or the impact of these

factors simultaneously. This proves that degradation isn't linear, but may be temporarily masked by

other factors.
Graph 1 - Impact of Tire Age on Lap Time for Hamilton

The next step is to plot the data in order to observe any trends. A positive correlation would indicate

degradation, a negative slope would indicate other external factors like lighter fuel loads which are

improving lap times. However, as seen above in Graph 1, there is a stable trend. This means that there

may have been external factors, such as a decrease in fuel load throughout the race - which may have

hindered the data. Thus, I have decided to do a fuel correction, and eliminate this variable from my

data.

Fuel Correction

In order to calculate the fuel correction and create a stable fuel rate throughout all the laps (which will

isolate the tire wear variable), I have to find the relationship between fuel load and lap time. The

increase in speed is seen in the table above, and we can make the assumption that this is caused due to

an increase in speed due to reduced fuel load. Formula 1 cars have a set tank throughout a race, and

are not allowed to refuel at any point - this means that as the car consumes fuel it becomes lighter,

ultimately contributing to increased lap time. Since the cars and tires are the same throughout a stint,

the only difference is the fuel load.

In order to find this, I compared the first four laps of a stint to the 5 last laps of a stint, before the

pit-stop to change tires. Four comparisons will give me enough information to find the average

difference between times. I will simply subtract the end lap time from the end beginning lap time in

order to find the difference.

No. of the Lap time - Beginning Lap time - End of Stint Difference between
laps of Stint lap times (s)

Y1 𝑡ℎ
5 & 23
𝑟𝑑
93.049 92.311 0.738

Y2 𝑡ℎ
6 & 24
𝑡ℎ
93.113 92.686 0.427

Y3 𝑡ℎ 𝑡ℎ 0.623
7 & 25 92.738 92.115

Y4 𝑡ℎ
8 & 26
𝑡ℎ
92.928 91.854 1.074
Table 5 - Difference of lap times within the same stint in different periods of the race (Hamilton)

Now, I will find the average difference - then divide by the number of laps between the beginning of

the stint and end of stint. This will give me the average time speed increases in a lap due to a lighter

fuel load.

𝐴𝑑 = 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒

𝑌1 + 𝑌2 + 𝑌3 + 𝑌4
𝐴𝑑 = 4

𝐴𝑑 = 0. 7155

𝑁𝑠 (𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑙𝑎𝑝𝑠 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑏𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 𝑎𝑛𝑑 𝑒𝑛𝑑 𝑜𝑓 𝑠𝑡𝑖𝑛𝑡) = 18

𝐴𝑑 0.7155
𝐹𝑢𝑒𝑙 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 = 𝑁𝑠
= 18

𝐹𝑢𝑒𝑙 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 = 0. 03975(𝑠) 𝑝𝑒𝑟 𝑙𝑎𝑝

The next step is to find the ‘clean’ lap times, which will effectively eliminate the fuel load variable

from my data - making it easier to calculate tire degradation. In order to find this, I will subtract the
current lap from the total number of laps in a specific stint (in this case when medium tires were

used). Then I will multiply it by the fuel correction above, finally adding it to the real lap time. This

formula will result in lap times that act as if the fuel tank was nearly empty.

𝑁 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑙𝑎𝑝

𝑁𝑡 = 𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑙𝑎𝑝𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑠𝑡𝑖𝑛𝑡

𝐶𝑙𝑒𝑎𝑛 𝑙𝑎𝑝 𝑡𝑖𝑚𝑒 = 𝑅𝑒𝑎𝑙 𝑙𝑎𝑝 𝑡𝑖𝑚𝑒 − ((𝑁𝑡 − 𝑁) × 0. 03975)

𝐸𝑥. 1 : 𝐶𝑙𝑒𝑎𝑛 𝑙𝑎𝑝 𝑡𝑖𝑚𝑒 (𝑙𝑎𝑝 5) = 93. 049 − ((27 − 5) × 0. 03975)

𝐶𝑙𝑒𝑎𝑛 𝑙𝑎𝑝 𝑡𝑖𝑚𝑒 (𝑙𝑎𝑝 5) = 92. 1745

I then calculated the clean lap times for all 10 laps I’ve extracted from the data, and then arranged

them in a table for easy analysis.

Lap Number Real lap time Clean lap time

5 93.318 92.1745

6 93.428 92.59325

7 93.556 92.761

8 93.402 92.64675

9 93.373 92.6575

19 93.011 92.693

20 92.586 92.30775

21 92.52 92.2815

25 92.586 92.5065

26 92.293 92.25325

Now that I have the clean lap times after fuel correction, I am able to proceed with graphing and

modelling the impact of tire degradation alone on lap times. Next I plotted the data with the clean lap
times to see if there was any difference in the trend observed now that I've eliminated fuel load by

stabilising it across all laps.

Graph 2 - Impact of tire age on the clean lap times

As seen above, I successfully managed to eliminate the extraneous variable from my data, leaving me

able to focus on tire degradation alone. And based on the trend above, tire degradation has a clear

impact on lap times. As the tire age increases, the lap times become longer - this positive correlation

shows that older tires do have a clear impact on slower lap times. The trend is linear for the most part,

showcasing the relationship between tire wear and slower lap times.

Linear Regression

Furthermore, I decided to quantify the impact of tire degradation on lap times. After incorporating the

clean lap times to eliminate the fuel load effect, I performed linear regression to find the relationship

between tire age and lap times.

Evaluation
One of the limitations I ran into was a faster lap time rather than a slower time, meaning

that as tire degradation increased, it seems that drivers were actually driving faster. This was

interesting, as tire degradation was expected to show slower lap times as the tire ages. I figured that

this could be attributed to other external factors such as fuel load reduction, improved tire adaptation,

or track evolution, which may have counterbalanced the effect of tire wear. In order to isolate tire

wear, I decided to take a new approach and find the relationship between fuel consumption from tire

wear, and then isolate the lap times from the effect of fuel consumption.

One assumption I made was that all cars had the same fuel load throughout all the laps in the race. I

assumed that the fuel load was nearly empty throughout, this allowed me to eliminate fuel load as a

potential extraneous variable in my calculations. However, this is not realistic - as Formula 1 cars’

fuel load decreases throughout the race and stint. While I attempted to adjust for this by adding a fuel

correction, this oversimplification may have impacted the accuracy of my results.

Moreover, another limitation was the use of only medium tires throughout my calculations. F1 rules

state that teams must pit-stop at least once during a race, switching out their tires. This was evident in

the 2021 British GP, where the drivers started on medium tires then switched over to hard. Focusing

on one tire set within a stint allowed for focus within my calculations - but is not a reflection of what

happens over the course of a race.

Works Cited

Pirelli. “Pirelli Motorsport: Best Products for Motor Competitions.” Pirelli.com, 2024,

www.pirelli.com/tyres/en-ww/motorsport/car/formula-1.

“FORMULA 1 PIRELLI BRITISH GRAND PRIX 2021.” Formula1.com, 2021,

www.formula1.com/en/racing/2021/great-britain. Accessed 15 Nov. 2024.

“2021 British Grand Prix Interactive Data: Lap Charts, Times and Tyres · RaceFans.” RaceFans, 18

July 2021,

www.racefans.net/2021/07/18/2021-british-grand-prix-interactive-data-lap-charts-times-and-t

yres/.

“2021 British GP: Race Pace.” F1bythenumbers, 19 July 2021,

f1bythenumbers.com/2021-british-gp-race-pace/. Accessed 16 Nov. 2024.