Understanding How Results Stabilize Over Time

Welcome to Understanding How Results Stabilize Over Time! You have reached lesson three of four, so the finish line for this course is in sight. So far, you have covered important ground: the first lesson drew a clear line between theoretical probability and observed frequency, and the second lesson showed how small samples can paint a wildly misleading picture. Both lessons focused on why real results deviate from what probability predicts. Now let's flip the perspective and ask: what pulls those results back?

The central idea is that observed results do not stay wild forever. As the number of observations grows, the observed proportion gradually settles closer to the theoretical probability. In this lesson, you will learn to:

• Describe why observed proportions gradually settle closer to the theoretical probability as the number of observations grows.
• Identify the Law of Large Numbers and explain what it promises about large samples.
• Distinguish between the proportion getting closer to the true rate and the raw count matching it exactly, avoiding the common expectation that large samples produce a perfect match.

In the previous lesson, you compared a small sample to a rough sketch of reality. Let's extend that image. Imagine assembling a jigsaw puzzle: with only a few pieces in place, the picture is mostly gaps and guesswork. Each new piece adds clarity, and after enough pieces the image becomes unmistakable, even though a few pieces may still be missing.

Data works the same way. Early observations hint at the overall pattern but leave plenty of room for wrong impressions. As we add more observations, the picture sharpens. This gradual sharpening is exactly what we mean when we say results stabilize over time, and it is the focus of everything we cover today.

A natural question is: how many observations do you need before you can trust the results? There is no single magic number, because the answer depends on the situation and on how much precision is needed. However, a few practical benchmarks can help build this intuition:

1. Tens of observations are usually too few for confident conclusions. The call center manager's 10-call result of 40% was off by 20 percentage points.
2. Hundreds of observations typically bring the proportion within a few percentage points of the truth. Their 200-call result of 19% was only 1 point away from 20%.
3. Thousands of observations tend to narrow the gap to less than a percentage point, which is reliable enough for most everyday decisions.

These are rough guidelines, not strict cutoffs. Reliability is a gradual process: we do not flip a switch from "unreliable" to "reliable" at some fixed sample size. Instead, our confidence grows steadily as data accumulates.

Consider a real-world example. If a new medication is tested on 15 patients and helps 80% of them, that figure could easily be off by 15 or 20 percentage points. If it is tested on 5,000 patients and helps 62%, the true rate is very likely within a point or two of that number. The larger study gives us a much tighter window around reality, making it far more useful for decisions.

The central lesson is that observed proportions stabilize as you collect more data. Each new observation has a smaller effect on the overall proportion, gradually pulling the result toward the theoretical probability. The formal name for this idea is the Law of Large Numbers, and it is one of the most practically useful facts in all of probability. At the same time, "stabilize" does not mean "match perfectly" — even very large samples leave a small gap between the observed proportion and the true rate, but that gap shrinks to a size that rarely matters for real decisions.

Now it is time to put these ideas into action. In the upcoming practice exercises, you will compare small and large samples side by side, estimate plausible ranges for real scenarios, and explain in your own words when probability predictions start earning our trust. Let's see how sharp your stabilization instincts have become!