Quant Modelling [HFT] All modules
1 Module 01

Introduction to Quant Research & Workflow

How modern trading firms replace intuition and news with data, probability and disciplined testing — and the five-stage loop every quant researcher runs.

Exam: Quant Research Workflow Source: Lecture 1 Foundational

What is quantitative trading?

Quantitative ("quant") trading means using data to make trading decisions. Instead of reacting to a news headline or a gut feeling, a quant builds algorithms from measurable market patterns and lets a systematic process decide when to buy and sell.

Core idea

Quant trading replaces emotion with systematic research: data, probabilities, and testing drive every decision, so the same rule is applied consistently across thousands of opportunities.

The same data-driven logic powers systems you use daily — Netflix and YouTube recommendations, Google Maps traffic prediction. A trading system is just a prediction engine pointed at markets.

Traditional vs quant trading

DimensionTraditional tradingQuant trading
Decision basisIntuition, news, experienceData, probabilities, statistics
ConsistencyVaries with mood / fatigueIdentical rule every time
Validation"It felt right"Backtested on history
ScaleA few positionsHundreds of assets, automated

Modern firms increasingly automate trading decisions. Why do markets exist at all? To let participants invest for future growth, speculate on prices, manage and hedge risk, and provide liquidity.

What does a quant researcher do?

📥

Collect data

Gather clean market data — prices, volume, quotes.

🔬

Study patterns

Find statistical regularities that may predict returns.

⚙️

Build signals

Turn patterns into explicit buy/sell rules.

🧪

Backtest

Test the rules on historical data before risking money.

📈

Deploy

Run live, then monitor and improve continuously.

🏛️

Where

Jane Street, Citadel, Two Sigma, Jump Trading, WorldQuant.

The quant research workflow

Research is a loop, not a straight line. Most ideas fail validation and send you back to the drawing board — that filtering is the job.

flowchart LR
  A([Collect Data]) --> B([Analyze Patterns])
  B --> C([Build Signals])
  C --> D([Backtest])
  D -->|Edge confirmed| E([Deploy & Monitor])
  D -.->|Fails validation| B
  E -.->|Performance decays| B
  classDef s fill:#eef2ff,stroke:#4f46e5,stroke-width:1px,color:#3730a3;
  classDef d fill:#ecfdf5,stroke:#0d9488,color:#0d9488;
  class A,B,C,D s
  class E d
      
The five-stage research loop. Dashed arrows show the feedback that makes research iterative.

Alpha: a statistical edge

Definition — Alpha

Alpha is a statistical edge: a signal that is right more often than chance, producing returns beyond the benchmark. No signal works 100% of the time — quant trading is fundamentally probabilistic.

If a signal is correct on 55% of trades with positive average profit, that small edge compounds across thousands of trades. We can frame the expected profit per trade as:

Expected value per trade $$ \E[\text{profit}] = p \cdot W - (1-p)\cdot L $$

where $p$ is the win probability, $W$ the average win and $L$ the average loss. A positive expectation, applied with discipline over many trades, is the entire game.

Worked example Does a 55% signal make money?
  1. A momentum signal wins on $p = 0.55$ of trades. Average win $W = \$120$, average loss $L = \$100$.
  2. Plug into the expectation: $\E[\text{profit}] = 0.55 \times 120 - 0.45 \times 100$.
  3. Compute: $\E[\text{profit}] = 66 - 45 = 21$.
Expected profit ≈ $21 per trade. A modest edge, but over 1,000 trades that is ~$21,000 — this is why quants think in probabilities, not single bets.
Two classic edges

Momentum — assets rising strongly may keep rising as trends attract buying pressure. Mean reversion — markets sometimes overreact, so prices may snap back toward normal levels as emotions cool.

Why quant trading is hard

The four headwinds
  • Markets evolve — yesterday's pattern can vanish.
  • Competition is intense — other quants chase the same edges.
  • Signals decay — alpha weakens as the market adapts to it.
  • Transaction costs — fees, spread and slippage eat into profits.
Self-check: A strategy wins only 45% of the time. Can it still be profitable?
Yes. Profitability depends on expectation, not win rate. If the average win is much larger than the average loss (e.g. $W=\$300$, $L=\$100$), then $0.45\times300 - 0.55\times100 = 135 - 55 = \$80$ per trade — clearly profitable despite a sub-50% hit rate.
Self-check: What makes the research workflow a "loop" rather than a line?
Backtests that fail validation send you back to analyze new patterns, and even deployed strategies decay over time and must be re-researched. The feedback arrows make research a continuous filtering process.

Key takeaways

Remember
  • Quant Trading = Data + Research + Probability + Discipline.
  • Good quant researchers think scientifically — hypothesis, test, iterate.
  • The goal is systematic decision-making, not perfect prediction.
  • Alpha is a probabilistic edge, never a guarantee.