1. Exponential Growth
Before building a bell curve, let's look at basic exponential growth: $e^x$. As $x$ gets larger, the value shoots up to infinity. This doesn't work for probabilities, which must be finite!
Notice how the curve explodes infinitely to the right.
2. Exponential Decay
What if we make the exponent negative? The function $e^{-x}$ decays beautifully to zero on the right, but it explodes to infinity on the left! We need a function that decays on both sides.
Notice how the curve now explodes infinitely to the left.
3. Distance from the Center
To fix the infinite growth, we want a shape that is highest at the center ($x=0$) and drops off as we move away in either direction.
By raising Euler's number to a negative absolute distance ($-|x|$), we get a peak at 1 that shrinks symmetrically.
Notice how this fixes the explosions, but creates an unnatural sharp point at the top.
4. The Smooth Bell Shape
By replacing the absolute distance with a squared distance ($x^2$), the sharp peak becomes beautifully smooth and rounded—creating the foundation of the famous "bell" shape.
Through calculus, measuring the variance (spread) of this exact shape results in an awkward standard deviation of exactly $1/\sqrt{2}$ (about 0.707). This is a messy starting point for statistics!
5. The "1/2" Math Trick
To fix the awkward fraction, we add a $1/2$ into the exponent. Why? In calculus, the derivative of $\frac{1}{2}x^2$ is simply $x$.
When we run this widened curve through our calculus formulas, the $1/2$ causes the messy math to cancel out beautifully. It forces the baseline spread (Standard Deviation, $\sigma$) to be exactly 1.0, making all future statistics much easier!
6. Adding Spread ($\sigma$)
Not all data is tightly clustered. We introduce a variable $\sigma$ (sigma), known as the Standard Deviation, to control how spread out the bell curve is.
Notice how a larger $\sigma$ stretches the curve horizontally on the main graph.
7. Shifting the Center ($\mu$)
Data doesn't always center around zero. We introduce $\mu$ (mu), the Mean, to shift the entire curve left or right.
We subtract $\mu$ from $x$. Notice how the graph shifts to the right as $\mu$ becomes positive.
8. Normalization (Area = 1)
A true probability distribution must have a total area under the curve equal to 1 (representing 100% probability).
Currently, when you increase $\sigma$, the area grows. Calculus tells us the area of our previous function is exactly $\sigma\sqrt{2\pi}$.
By dividing our function by this constant, the area remains perfectly locked at 1, no matter how much you stretch it!
Total Area = 1.000