6.3 Normal Distribution (Lap Times)

Normal Distribution and Lap Times

The normal distribution gives you a way to calculate how likely specific lap times are, compare individual performances to the group, and check whether your data is well-behaved enough to analyze with standard tools. This section uses racing lap times as the running example, but the techniques apply to any normally distributed data.

Probability Calculations with Normal Distribution

A normal distribution is a continuous, symmetric, bell-shaped probability distribution. Two parameters define it completely:

• Mean ($\mu$): the center of the distribution (the average lap time)
• Standard deviation ($\sigma$): how spread out the lap times are around that average

To find the probability of observing a particular lap time (or range of lap times), you convert the raw value into a z-score and then look up the corresponding area under the curve.

Steps to calculate a probability:

1. Standardize the lap time using the z-score formula: $z = \frac{x - \mu}{\sigma}$

where $x$ is the lap time, $\mu$ is the mean, and $\sigma$ is the standard deviation.

1. Look up the z-score in a standard normal table or use a calculator. The table gives you the area to the left of that z-score, which is the probability of a lap time being less than $x$.

2. Interpret the result. You can find:

   • $P(X < x)$: read directly from the table
   • $P(X > x)$: subtract the table value from 1
   • $P(a < X < b)$: find the area for both z-scores and subtract

Example: Suppose lap times have $\mu = 58$ seconds and $\sigma = 2$ seconds. To find the probability of a lap under 60 seconds: $z = \frac{60 - 58}{2} = 1.0$

A z-score of 1.0 corresponds to about 0.8413, so roughly 84% of laps would be under 60 seconds.

The Empirical Rule is a quick shortcut for normal distributions:

• About 68% of data falls within $\pm 1\sigma$ of the mean
• About 95% falls within $\pm 2\sigma$
• About 99.7% falls within $\pm 3\sigma$

Interpretation of Percentiles and Z-Scores

Percentiles tell you what percentage of lap times fall below a given value. A lap time at the 75th percentile means 75% of all lap times were faster (lower). A lap time at the 90th percentile is slower than 90% of the other times.

Be careful with the direction here: in racing, lower times are better. So a lap time at a high percentile is actually a slow time relative to the group.

Z-scores tell you how many standard deviations a lap time sits from the mean:

• Positive z-score → above the mean → slower than average
• Negative z-score → below the mean → faster than average
• Z-score of 0 → exactly at the mean

Z-scores make it easy to compare across different contexts. A lap time with $z = -2.0$ is unusually fast no matter what track or car you're looking at, because it's 2 standard deviations below the mean.

Example: If a racer posts a lap with $z = 1.5$, that lap is 1.5 standard deviations slower than average. If another racer on a different track posts $z = -0.5$, their lap is half a standard deviation faster than average for their distribution.

Analysis of Normal Probability Plots

A normal probability plot (also called a normal Q-Q plot) is a graph that helps you check whether your data actually follows a normal distribution. This matters because many statistical methods assume normality, and if that assumption is wrong, your probability calculations won't be reliable.

The plot compares the observed data values to what you'd expect to see if the data were perfectly normal. If the data is normal, the points line up roughly along a straight line.

How to construct one:

1. Order the lap times from smallest to largest.

2. Calculate the percentile rank for each lap time: $\frac{i - 0.5}{n} \times 100$, where $i$ is the rank and $n$ is the total number of lap times.

3. Find the z-score that corresponds to each percentile (these are the "expected" normal values).

4. Plot the observed lap times against the expected normal values.

How to read the plot:

• Points forming a roughly straight line → the data is approximately normal
• A curved pattern → the distribution is skewed (lap times bunched toward the fast or slow end)
• An S-shaped pattern → the tails are heavier or lighter than a normal distribution would predict (more extreme lap times than expected, or fewer)

Statistical Inference and Sampling

These concepts extend the normal distribution into making conclusions about a whole population from a sample.

The central limit theorem (CLT) says that if you take many samples of the same size and compute the mean of each, those sample means will follow an approximately normal distribution, even if the original data isn't normal. This works as long as the sample size is large enough (commonly $n \geq 30$ as a rough guideline).

A sampling distribution is the distribution of a statistic (like the sample mean lap time) across many repeated samples. It's not the distribution of individual lap times; it's the distribution of averages of lap times.

• Confidence intervals use sample data to give a range of plausible values for a population parameter. For example, "we're 95% confident the true mean lap time is between 57.2 and 59.1 seconds."
• Hypothesis testing uses sample data to evaluate claims about the population, such as whether one group of racers has a significantly different mean lap time than another.