How Rohit Discovered Linear Regression – Simply Explained

Welcome to the third article in our Simply Explained series — where we break down tough topics with everyday stories. If you're a student or someone who wants to understand complex concepts without the complexity, you're in the right place.

⚠️ Heads-up! This is just the first part of our regression series. In this story, Rohit only scratched the surface — there’s more to come! In the next part, we’ll explore how regression works with more than one variable and how you can build your own prediction models step-by-step.

The Setup

Rohit, an MBA student, came back to his hostel after a tiring cricket match. As he flipped open his book, it hit him—he had a linear regression test the next day. He rushed to his roommate, Varun, for help.

🧑‍🎓 Rohit: "Bro, all the stats lectures have just gone over my head this term. Please help me!"

🧑‍💻 Varun: "Sure, buddy! By the way, how did your match go?"

🧑‍🎓 Rohit: "Bro, we lost. And tomorrow's test might just be worse. Let’s just get started."

🧑‍💻 Varun: "Haha! Alright, let’s make it fun. Let’s play a simple guessing game."

The Guessing Game Begins

🧑‍💻 Varun: "What’s the next number in this series? 22, 23, 24, 25, 26, 27, 28, 29…"

🧑‍🎓 Rohit: "Easy. 30."

🧑‍💻 Varun: "Great! Now try this: 4, 9, 16, 25, 36, 49, 64…"

🧑‍🎓 Rohit: "Oh! These are squares of numbers. The next one is 81!"

🧑‍💻 Varun: "Perfect! Now guess this: 7, 1, 5, 8, 3, 4, 5, 1…"

🧑‍🎓 Rohit: "Wait… what?! There’s no pattern here. This is random!"

🧑‍💻 Varun: "Or is it? What if I told you, these are the runs scored by West Indies in 8 overs against India in an ODI? Now, predict the 9th over’s runs."

🧑‍🎓 Rohit: "I can only give a range, maybe within 10 runs."

🧑‍💻 Varun: "Nope! Pick a single number. Let’s make it a bet—the closer you get, the higher I'll pay. ₹1 to ₹1000."

🧑‍🎓 Rohit: "That’s not fair! You could just change the answer."

🧑‍💻 Varun: "Fine. I’ll write the answer on a slip, fold it, and give it to you. We’ll open it after you guess."

🧑‍🎓 Rohit: "Deal!"

🧑‍🎓 Thinks 🤔 — "Average runs per over are 4.25. I’ll go with 4."

📜 Rohit opens the slip. It says 5, along with: MO/Y/ECO/LSR

The Twist

🧑‍🎓 Rohit: "Hey! I was close. I should get the full prize."

🧑‍💻 Varun: "Haha, sure! But now let’s make it harder. Guess the 10th over’s score."

🧑‍🎓 Rohit: "Give me a 1–2 run margin."

🧑‍💻 Varun: "Okay. But if you get it wrong, you owe me ₹1500."

🧑‍🎓 Rohit: "Fine! I’ll go with 4 again."

📜 He opens the slip… It says 24, and below it: DO/F/EXP/HSR

🧑‍🎓 Rohit: "No way! You made that up."

🧑‍💻 Varun: "Nope. You lost because you didn’t ask enough questions. Look at the text."

🧑‍🎓 Rohit: "What do these mean?"

🧑‍💻 Varun:

• MO/Y/ECO/LSR → Middle Over, Yorker-Economy bowler, Low Strike Rate batsmen

• DO/F/EXP/HSR → Death Over, Fast-Expensive bowler, High Strike Rate batsmen

🧑‍🎓 Rohit: "So... the context changed?"

🧑‍💻 Varun: "Exactly! You assumed every over was the same. But in reality, runs depend on many things: over number, bowler type, batsman form..."

🧑‍🎓 Rohit: "So to predict properly, I need more than just past numbers. I need factors too."

🧑‍💻 Varun: "Yes! That’s the idea behind Linear Regression — using past data plus the right factors to predict something."

Understanding Linear Regression

In the first two patterns, the answer was easy: 👉 22 → 30 (Simple sequence) 👉 4 → 81 (Squares)

But in real life — like cricket — it's not that simple.

👉 Over 9 was predictable because conditions were stable.
👉 Over 10 was unpredictable because the context changed.

That’s where Linear Regression helps. It uses:

✅ Historical data
✅ Influencing factors (variables)

to predict outcomes. But only if the right variables are used.

Why Simple Averages Fail

Rohit’s mistake was using just the average. That’s like predicting monsoon day 1 rainfall by taking the average of the previous 5 days of precipitation.

But Linear Regression uses multiple variables—more context = better prediction.

It’s like:

• Over number matters (early overs = fewer runs)

• Bowler type matters (spinners vs pacers)

• Batsman-type matters (aggressive vs defensive)

Cricket to Real Life

Varun grinned: “This is why businesses use regression too!”

📈 To predict sales based on ad spend, season, and product type
🏠 To guess house price using size, location, and age
🍽️ Even restaurants use it to forecast weekend traffic

The Takeaway

🧑‍🎓 Rohit: "Wow, so Linear Regression = better guesses with better data? Now I get it! Next time I make my Dream11 team, I’ll say I used linear regression. 😆"

🧑‍💻 Varun: "Haha! But remember, Linear Regression isn’t perfect. Cricket scores depend on many changing things. Sometimes"

🧑‍🎓 Rohit: "Tell me more!"

🧑‍💻 Varun: "Sure, but first, let’s break down why it’s called linear..."

Final Thoughts

Linear Regression is a useful tool to predict things — but only when you pick the right factors. Blindly using past data can mislead you.

Next time you're predicting anything — from cricket scores to business sales — ask:

✅ What influences this?
✅ Is past data enough?
✅ Are things changing?

And just like Rohit, don’t just guess — analyze.

📘 This is part of the Simply Explained series — where we break down complex ideas with simple, relatable stories.

Check out the other stories:

• How Raju Learned About Financial Leverage

• How Suraj Got Stuck in Operational Leverage

🚀 Stay tuned for the next article, where we go one step deeper. You’ll learn:

• What happens when you add more variables to your model

• How to tell if your prediction line is working

• And how to build your first real regression model (no coding needed!)

Got a topic that confuses you? Tell us at contact@simplyexplained.in. We’ll break it down. Simple. Fun. Useful.