Everything in Life is Linear Regression

When I started learning ML, I was first introduced to Linear Regression. In short, it describes an algorithm where you can model a function using a linear expression:

y = wx + c

Largely similar to the equation of a straight line. Here, the value of y (dependent variable) changes with x (independent variable).

Now, if we extrapolate this to multiple independent variables:

y = w₁x₁ + w₂x₂ + … + wₙxₙ + c

In most use cases of linear regression, this is the case. An outcome or output is dependent on multiple factors.

Suppose you’re modeling the house price of a city using linear regression. You’ll find that historically, the price of a house depends on multiple factors - area, number of rooms, sq footage, parking (available or not), and so on. The Linear Regression algorithm tries to find those coefficients - w₁, w₂, … wₙ - and we get a model (or equation) on which, if we feed in new values of x₁…xₙ, we can “predict” or “estimate” the cost of the house in question.

Now, the idea of this blog is not to deep dive into LR. It’s because I seem to find a parallel between everything in life and this mathematical concept - not the linear part, but the combination part where everything is a combination of multiple things with different scaling factors associated with each of them.

For example:

Suppose you missed a train on a certain day. You become extremely angry and start blaming your mom for apparently “making you late” by asking you to eat breakfast before leaving. But this is black-and-white thinking - sure, it might have played a role. But there are other factors here as well to consider. Like the fact that you slept late last night despite knowing you have a train to catch the next day. Also, the traffic at that time was more than usual.

Multiple factors contributing to missing a train — Missing the train isn't about one factor - it's a weighted combination of breakfast delay, waking up late, traffic, and more

If I were to put it in the equation:

minutes_late = w₁(breakfast_delay) + w₂(woke_up_late) + w₃(traffic_level) + w₄(distance_to_station) + w₅(train_punctuality) + c

Where:

minutes_late = how many minutes late you arrived at the station (or how close you were to missing the train)
breakfast_delay = time spent on breakfast (in minutes)
woke_up_late = how late you woke up compared to planned time (in minutes)
traffic_level = traffic congestion factor (could be 1-10 scale, or actual delay in minutes)
distance_to_station = distance you need to travel (in km)
train_punctuality = how early/late the train typically runs (in minutes)
c = baseline constant (accounts for other unmeasured factors)

The weights (w₁, w₂, w₃, etc.) represent how much each factor contributes to the outcome. For instance:

Maybe w₂ is large because waking up late has a huge cascading effect
w₁ might be small because breakfast only added 5 minutes
w₃ could be moderate depending on how unpredictable traffic is

The more experiences I have in life, the more I resonate with this.

Now, I know real linear regression has assumptions about linearity and independence that life often violates. But as a mental model for thinking about multiple factors contributing to outcomes, it works surprisingly well.

This also helps me approach differences in opinions in a calmer and composed manner. Let’s say India wins a cricket match - some say it was because of Virat’s ton. Some say it’s because of Bumrah’s fifer. Or some say it was because of Rohit’s quickfire 25 off 10 balls.

I say it’s all of that. Just with different weights.

Another example:

Someone says, “They broke up because he was toxic.”

But reality?

relationship_strain = w₁(miscommunication) + w₂(incompatible goals) + w₃(external stress) + w₄(personality issues) + w₅(past baggage) + c

We love single-factor explanations because they’re simple. But life is multivariate, not binary.

Ever since I started viewing life through this lens - not the linear part of Linear Regression, but the weighted combination part - I’ve become less judgmental, more curious, and surprisingly more forgiving.

Because nothing “just happens.”

Life as a sum of weighted factors — Every outcome in life is the sum of multiple factors, each with its own weight - plus a bit of randomness

Outcome = Σ (all factors × their weights) + some randomness

And most of us are just bad at estimating the weights.

For example:

Another example:

More from my blog