Part 4 of 7
🎯 Learning Objective
By the end of this section, you will understand the essential quantum computing concepts needed for financial applications — without deep physics — and see exactly how these concepts map to finance problems.
| Classical Bit | Qubit |
|---|---|
| Can be 0 or 1 at any time | Can be in a superposition of 0 and 1 simultaneously |
| Like a coin lying flat: heads OR tails | Like a coin spinning in the air: heads AND tails until you look |
| Deterministic state | Probabilistic state until measured |
| Represented by a single number | Represented by two complex numbers (amplitudes) |
Where: $|\alpha|^2$ = probability of measuring 0, $|\beta|^2$ = probability of measuring 1, and $|\alpha|^2 + |\beta|^2 = 1$.
💰 Financial Intuition
Think of a qubit as representing a market outcome. Instead of simulating one scenario at a time (classical), a qubit holds information about multiple scenarios simultaneously.
Problem: A qubit is in state $|\psi\rangle = \frac{\sqrt{3}}{2}|0\rangle + \frac{1}{2}|1\rangle$. What is the probability of measuring each outcome?
Solution:
Finance analogy: This qubit encodes a biased market view — 75% chance of scenario A (e.g., market up) and 25% chance of scenario B (e.g., market down). The quantum algorithm would use this to weight expected payoffs.
Classical approach (Monte Carlo):
Quantum approach:
With $n$ qubits, you can represent $2^n$ scenarios simultaneously:
| Number of Qubits | Simultaneous States | Classical Equivalent |
|---|---|---|
| 10 | 1,024 | 1 thousand simulations |
| 20 | ~1 million | 1 million simulations |
| 30 | ~1 billion | 1 billion simulations |
| 50 | ~1015 | 1 quadrillion simulations |
⚠️ Important Caution
Superposition doesn't mean we get all answers at once. Measurement collapses the quantum state. The art of quantum algorithms is in structuring the computation so that the right answer has the highest probability of being measured.
When qubits are entangled, measuring one instantly determines information about the other — regardless of distance.
💰 Financial Intuition
This is like correlation between assets, but much stronger. In a classical portfolio, two correlated stocks tend to move together. In quantum computing, entangled qubits have a mathematically guaranteed relationship. This property allows quantum computers to efficiently represent and manipulate correlated probability distributions — exactly what we need for multi-asset portfolio simulation.
When we measure a qubit, its superposition collapses to either |0⟩ or |1⟩ with probabilities determined by the amplitudes. This is analogous to observing a financial outcome:
Key insight: Quantum algorithms are designed so that the desired answer has the highest measurement probability.
Quantum computations are built using quantum gates:
| Gate | Symbol | Action | Finance Analogy |
|---|---|---|---|
| Hadamard (H) | H | Creates superposition | "Consider all possible market outcomes" |
| CNOT | CX | Entangles two qubits | "Link two correlated assets" |
| Rotation (Ry) | Ry(θ) | Adjusts probabilities | "Weight the probability of a scenario" |
| Grover diffusion | G | Amplifies target state | "Zoom in on the answer" |
Match each quantum concept to its best finance application:
| Quantum Concept | Best Finance Match | Why? |
|---|---|---|
| Superposition | Monte Carlo scenarios | Explore many market scenarios simultaneously |
| Entanglement | Correlated asset returns | Encode asset correlations in qubit relationships |
| Amplitude estimation | Expected value computation | Extract mean payoff with quadratic speedup |
| Ry rotation gates | Probability distributions | Encode specific probabilities into qubit states |
The critical bridge between finance problems and quantum solutions:
| Finance Problem | Classical Method | Quantum Method | Quantum Advantage |
|---|---|---|---|
| Monte Carlo | Random sampling, $O(1/\sqrt{N})$ | Quantum Amplitude Estimation | Quadratic: $O(1/N)$ |
| Portfolio Optimization | Mixed-integer programming | QAOA / VQE | Explores combinatorial space via superposition |
| Option Pricing | Black-Scholes, binomial, MC | Quantum Simulation + QAE | Faster pricing of exotic derivatives |
| Risk Analysis | Historical simulation, parametric | Quantum Sampling + QAE | Faster risk metric computation |
⭐ Most Important Concept in the Quantum Section
This is the single most important comparison to understand.
| Desired Accuracy | Classical Simulations | Quantum Oracle Calls | Speedup |
|---|---|---|---|
| 1% error | 10,000 | 100 | 100× |
| 0.1% error | 1,000,000 | 1,000 | 1,000× |
| 0.01% error | 100,000,000 | 10,000 | 10,000× |
| 0.001% error | 10,000,000,000 | 100,000 | 100,000× |
💡 The Quadratic Speedup
To achieve the same accuracy, quantum methods need the square root of the number of classical simulations. This was proven by Brassard et al. (2002) and has been extensively studied by Rebentrost et al. (2018), Woerner & Egger (2019), and Carrera Vazquez & Woerner (2021).
Classical Monte Carlo
$O(1/\sqrt{N})$
Error decreases slowly
100× more work → 10× less error
Quantum Amplitude Estimation
$O(1/N)$
Error decreases fast
10× more work → 10× less error
Scenario: Your bank currently runs 1,000,000 Monte Carlo simulations for overnight VaR. The computation takes 4 hours.
Questions:
Answers:
"This is why JPMorgan, Goldman Sachs, and every major bank has a quantum computing research team."
💡 Key Insight
The quantum algorithm doesn't generate random samples like classical Monte Carlo. Instead, it encodes the entire probability distribution into quantum amplitudes and uses quantum interference to extract the answer directly.
In Part 3, our Python code did this classically:
| Step | Classical (Part 3) | Quantum (QAE) |
|---|---|---|
| Distribution | np.random.standard_normal(n) | Ry rotation gates encode $p(x)$ in amplitudes |
| Payoff | np.maximum(ST - K, 0) | Controlled rotation encodes payoff |
| Estimation | np.mean(payoffs) over N samples | Grover amplification extracts $E[f(X)]$ |
| Accuracy | $O(1/\sqrt{N})$ — need 10,000 sims | $O(1/N)$ — need only 100 calls |
"Remember our Python simulation? We ran 10,000 paths. With quantum amplitude estimation, we could potentially achieve the same accuracy with just 100 quantum oracle calls."