Season 2
EP03 - Quant Edge
How quantitative trading provides economic value. Learn about liquidity provision, price discovery, market making, bid-ask spreads, and how quants make markets more efficient.
Tweet coming soon
It makes me sick, Kurumi.
What does? The latency?
The waste! This bank capital could build houses! It could build bakeries! Hospitals! Science!
Instead, it's lent to math nerds to play a zero-sum game of "Guess the Number" in microseconds.
They re-engineered the internet just to act faster than everyone else. It's pointless.
Nothing is created. No bread is baked. No cars are built. It's just money moving in a circle.
You think it's a circle?
It's a **Refinery**.
Imagine a world without Quants.
This is a market without **Liquidity**.
He wants to sell. But the buyer (The Bakery) isn't here right now. The buyer comes next Tuesday.
Since I have to wait until Tuesday to sell it... I'll pay you 50% of the value.
That's the **Risk Premium**. Take it or leave it.
Okay, that sucks. But we talked about it. We don't need *that* much liquidity.
Do we not?
The Quant says: "I will buy your Apple stock right now for $149.99."
"And I will sell it to the Bakery next Tuesday for $150.01."
The Quant made $0.02.
And the Pension Fund saved **millions**.
Instead of losing 50% to the Pawn Broker, they lost 0.01% to the Quant.
This "Minigame" you hate? It is a war to reduce the cost of doing business.
Quants fight each other to offer the *best* price.
But they borrowed money from the bank to do it!
They borrowed money to be the **Inventory**.
You can't run a supermarket with empty shelves. Quants borrow cash to fill the shelves with stock so *you* can buy whenever you want.
So they provide... service?
They provide **Price Discovery**.
The world is messy. People panic. Prices fluctuate wildly based on rumors.
The Quant takes the noise—the panic, the rumors, the glitches—and mathematically filters it.
By betting against the panic (Arbitrage), they force the price to stabilize.
They stabilize the cost of bread for your bakery. They stabilize the exchange rate for your vacation.
But the leverage, Kurumi! They borrow 10x their money!
If they are just "stabilizing," why do they need so much cash? It feels like gambling.
Gambling is betting on a coin flip (50/50).
Quants bet on **Inefficiencies**.
The profit margin for "stabilizing" the market is tiny. Maybe 0.001%.
To make a living off 0.001%, you need massive volume. You need leverage.
Banks lend to them because they are **Market Neutral**.
They don't care if the market crashes or booms. They profit from the *volume* of the trade.
It is safer than lending to your bakery, which goes bust if people stop eating gluten.
Okay. They are efficient.
But do they have to ruin the internet? They lay private cables just to be 1ms faster!
Do you pay to trade stocks on Robinhood?
No. It's free.
Why do you think it's free?
Because... they are nice?
Because the Quants **pay the broker** for the privilege of trading with you.
It's called **Payment for Order Flow**.
Their "Minigame" subsidizes your access to the market.
Fine, I get *why* they exist. But *how* is their business model possible?
It feels like magic. They borrow billions, trade instantly, and somehow print money. It has to be a scam.
It's not a scam. It's **Stochastic Control Theory**.
We process raw data into profit using four engineering pillars.
1. Inventory Control (Liquidity).
2. Cointegration (Arbitrage).
3. Covariance Matrices (Leverage).
4. Flow Segmentation (PFOF).
That sounds like a Physics textbook.
It is. Money is treated like a fluid. And fluids have rules.
Let's start with **Liquidity**. You think Market Makers just "buy and sell."
The engineering problem is: **The Hot Potato**.
The Apple tank is overflowing!
That means we bought too much Apple stock.
If the price drops now, we lose millions. We have **Inventory Risk**.
To fix this, the algorithm adjusts the **Quotes**.
Usually, Apple is $100.00 (Buy) / $100.02 (Sell).
But since our tank is full, the algo **Skews** the price down: $99.98 / $100.00.
By lowering the price, we discourage sellers (don't give us more!) and encourage buyers (please take it!).
It's a **Negative Feedback Loop**.
So... the price change isn't random?
No. The price moves because the Market Maker needs to empty their bucket. It's a control system balancing mass.
Okay, Inventory is defense. How do they attack?
One idea is to use **Cointegration**.
Most stocks are **Random Walks** (Non-Stationary). You can't predict them.
But the *difference* between two related stocks is often **Stationary**.
See this sine wave? It always returns to zero.
We don't predict the price of Coke. We predict the **tension in the spring**.
The spring is stretched! Coke is too high!
So the algorithm executes a **Pair Trade**.
Short Coke ($-$100$). Long Pepsi ($+$100$).
We are **Market Neutral**. We don't care if the market crashes. We only care that the spring snaps back.
But the spring only snaps back a tiny bit! Pennies!
How does that pay for the servers?
That's where the **Banks** come in.
Banks lend to us because we engineer away the risk using **Diversification**.
If I bet on one coin flip, it's risky.
If I bet on 1,000 coin flips that are *uncorrelated*, the variance drops to near zero.
The Law of Large Numbers protects the bank.
So you borrow 10x your money...
Because my mathematical risk is 1/10th of a normal investor.
I am not gambling. I am running a **Variance Harvesting Machine**. The leverage just scales the tiny edge into a living wage.
Payment for Order Flow. Isn't this the controversial one people complain about?
Engineering-wise, it's just **Signal Separation**.
Why do HFT pay for the smooth rocks?
Because the smooth rocks are **Uninformed Flow**.
This order comes from a dentist in Ohio buying Tesla because he likes the car.
It has **Zero Correlation** with the future price. It is random noise.
It is safe to Market Make against him because the price won't run me over.
This order comes from a Spy who knows the factory burned down.
This is **Toxic Flow**. If I trade with him, I will lose money.
PFOF is the fee we pay to trade *only* with the harmless dentist. We are buying a **Safe Space**.
Inventory Control handles the risk. Cointegration finds the profit. Covariance allows the leverage. PFOF filters the opponents.
You're learning pretty fast.
So it's not a slot machine. It's a casino operation!
It's a system designed to extract entropy from the universe and convert it into order (and cash).
Banks lend more to just "Business Ideas"
They also lend to **Engineering Architectures** that can prove their failure rate is statistically impossible.
"It does feel like a great thing to build, after all."
EP02 - HFT Trading
How high-frequency trading works. Learn about latency arbitrage, order book dynamics, FPGAs, colocation, and how HFT firms profit from microsecond speed advantages in markets.
EP04 - Alpha Decay
Why trading strategies decay over time. Learn about backtesting mistakes, overfitting, look-ahead bias, reverse engineering, and why alpha strategies become obsolete.