Understanding Markov Chains

Markov Chains sound like something out of a sci-fi movie, but they are actually a very intuitive way of looking at how things change from one state to another. Let's see how they work.

The Weather Analogy

Imagine you live in a city with very predictable, but changing, weather. You notice a pattern:

• If today is Sunny, there is an 80% chance tomorrow will also be Sunny, and a 20% chance tomorrow will be Rainy.
• If today is Rainy, there is a 60% chance tomorrow will be Rainy, and a 40% chance tomorrow will be Sunny.

Notice how tomorrow's weather only depends on today's weather. It doesn't matter what the weather was like last week, or last month. The only thing that matters for predicting tomorrow is what is happening right now.

This exact concept—where the next state depends only on the current state—is called a Markov Chain.

States and Transitions

In a Markov Chain, we talk about "States" and "Transitions."

States are the possible situations you can be in. In our analogy, the states are "Sunny" and "Rainy."
Transitions are the probabilities of moving from one state to another (like the 80% chance of Sunny to Sunny).

How We Use It For the Lottery

In a perfectly fair lottery, every draw is independent. The fact that the number '12' was drawn last week shouldn't influence whether it gets drawn this week. A completely random lottery has no "memory."

However, as researchers, our job is to prove that independence, rather than just assume it. We use Markov Chains to model lottery draws to check if there are any subtle transition patterns hiding in the data.

An Example

Let's group the lottery balls into two states: Odd Numbers and Even Numbers.

We can look at a sequence of draws and ask: "If the first ball drawn today is Odd, what is the probability that the first ball drawn next week is also Odd?"

By building a Markov Chain model of the historical draw data, we can calculate these transition probabilities. If the lottery is perfectly random, the chance of transitioning from an Odd number to an Even number should match the theoretical probability perfectly (roughly 50/50, depending on the exact ball pool).

If our Markov model shows that an Odd number is followed by another Odd number 70% of the time, that would flag an anomaly! (Don't worry, lotteries are highly secure and such anomalies are incredibly rare, but our tools are designed to catch them if they exist.)

Summary

• What it is: A mathematical system that transitions from one "state" to another, where the next state depends only on the current state.
• Everyday Example: Predicting tomorrow's weather based only on today's weather.
• Our Use Case: Modeling historical lottery draws to ensure that one draw doesn't secretly influence the next draw.