What is the efficient frontier? (no formulas)

The idea that won Harry Markowitz a Nobel Prize, explained with everyday examples, plus how to read it in the optimizer.

Oct 22, 2026 · 4 min read

In 1952, a doctoral student named Harry Markowitz published a paper just a few pages long that changed investing forever. His central idea can be told without a single formula: you shouldn't look at each investment on its own, but at how they behave together. From that idea comes the efficient frontier, a curve you'll see every time you use the optimizer.

The underlying question

Imagine you have money to split among three assets: a US stock ETF, a Brazil ETF and a bond ETF. You can put everything in one, split it equally or choose any other mix. There are endless combinations.

Each combination has two characteristics:

Expected return: how much it might earn in a year, on average.
Risk: how much it tends to go up and down along the way. It's measured by volatility.

The efficient frontier answers a very practical question: of all those combinations, which ones are worth it?

A chart with two axes

Picture a plane. The horizontal axis shows risk (further right, a bumpier ride). The vertical axis shows return (higher up, more expected gain). Every possible combination of your assets is a point on that plane.

If you plot thousands of combinations, you'll see a cloud of points. The upper-left edge of that cloud has a characteristic curved shape. That curve is the efficient frontier.

Why that edge? Because any point that isn't on it is "worse" than one that is:

Either there's another combination with the same risk and more return (a point right above it).
Or there's one with the same return and less risk (a point to its left).

Staying below the frontier is like paying more for the same thing.

The magic: diversification lowers risk

Here's the interesting part. If you combine two assets that don't move exactly alike, the risk of the mix can be lower than that of either one on its own.

An everyday example: someone runs a stall that sells umbrellas and another that sells sunscreen. In a rainy month, umbrellas sell well and sunscreen doesn't. In a sunny month, it's the other way around. Each business on its own is erratic, but together they produce much steadier income.

Something similar happens in markets, though less extreme. Japanese and Mexican stocks don't react the same way to the same news. US Treasury bonds often rise when stocks fall. Combining assets with low correlation makes the curve "bend" to the left: more return per unit of risk.

Key points on the curve

There are some named points on the frontier, which match methods in the optimizer:

Minimum variance: the leftmost point. It's the most stable mix possible with your assets.
Maximum Sharpe: the point where you get the most extra return per unit of risk. We explain it in Sharpe ratio.
Target return: you choose a height (for example, 10% a year) and the model finds the lowest-risk point on the curve at that height.
Target risk: you choose a horizontal position (for example, 15% volatility) and the model finds the highest point.

How to read it in the optimizer

When you calculate a portfolio you'll see:

The blue curve: the efficient frontier of your asset list, respecting your constraints.
The grey dots: each asset on its own. They usually sit to the right of or below the curve: on their own they're less efficient than when well combined.
The orange dot: the portfolio you chose with the method.

Hover over the curve and you'll see how the weights change at each point. On the left, the most stable assets dominate (bonds, for example). As you move up, the more volatile assets with higher expected returns come in.

The frontier's pitfalls

The efficient frontier is a powerful tool, but it has limits worth knowing:

It uses the past. Returns and correlations are estimated from historical data. If the future is different — and it usually is — the real frontier will be different too.
It's very sensitive to expected returns. Small changes in how much an asset is expected to earn can move the weights a lot. That's why the optimizer uses a more robust estimator than the simple average, one that "pulls" every asset's return toward a common value.
It can concentrate. Without limits, the model sometimes puts a large share in one or two assets. Setting a maximum weight per asset, or regional limits, gives more sensible results.
It doesn't know your life. It doesn't know when you'll need the money or how much market drops scare you.

An exercise to understand it

Try this in the optimizer:

Add just two very different assets, for example a US stock ETF and a bond ETF.
Calculate and look at the curve: it goes from "all bonds" (bottom left) to "all stocks" (top right).
Now add a gold or Japan ETF and calculate again. Notice how the curve shifts up and to the left: that's diversification at work.

In short

The efficient frontier is the set of combinations that can't be improved: there's no way to get more return without taking more risk, or less risk without giving up return. Diversifying with assets that move differently pushes that frontier into a better zone. Use it as a map, not a prophecy.

Open the optimizer and look at the frontier of your own list.

Try it with your own assets

Build a list and see which portfolio the model calculates, in your currency.

Open the optimizer

Disclaimer: Educational and simulation tool. It is not investment advice or a personalized recommendation. Past performance does not guarantee future results.