Background: Exchanges and Order Book Mechanics

Understanding how exchanges operate is the foundation for understanding slippage, execution, and market microstructure.

1. Types of Exchanges

1.1 Centralized Exchanges (CEX)

Examples: NYSE, NASDAQ, Shanghai Stock Exchange, Binance

Characteristics:

• Unified matching engine
• Central order book
• High liquidity
• Regulated

1.2 Decentralized Exchanges (DEX)

Examples: Uniswap, dYdX

Characteristics:

• Smart contract matching
• AMM or order book
• No trusted intermediary
• Higher latency

2. Order Book

2.1 Basic Structure

Sell Side (Ask)        Price          Buy Side (Bid)
                      $101.50
50 shares             $101.25
100 shares            $101.00
                      $100.75         200 shares
                      $100.50         150 shares
                      $100.25         80 shares

• Bid: The highest price buyers are willing to pay
• Ask: The lowest price sellers are willing to accept
• Spread: Best Ask - Best Bid

2.2 Order Book Depth

Level 1: Only best bid/ask

Best Bid: $100.75 × 200
Best Ask: $101.00 × 100

Level 2: Multiple price levels

Bid:
  $100.75 × 200
  $100.50 × 150
  $100.25 × 80

Ask:
  $101.00 × 100
  $101.25 × 50
  $101.50 × 30

Level 3: Individual order details (usually not public)

3. Order Types

3.1 Market Order

Definition: Execute immediately at the current best price

Pros: Guaranteed execution Cons: Price not guaranteed, may incur slippage

Example:

Order Book:
  Ask: $101.00 × 100, $101.25 × 50

Market buy 120 shares:
  100 shares @ $101.00
  20 shares @ $101.25

Average execution price = (100×101.00 + 20×101.25) / 120 = $101.04
Slippage = $101.04 - $101.00 = $0.04

3.2 Limit Order

Definition: Execute only at specified price or better

Pros: Price certainty Cons: May not execute

Example:

Limit buy 100 shares @ $100.75

If current Best Ask = $101.00:
  → Order enters bid queue
  → Wait for seller willing to sell at $100.75

3.3 Other Order Types

4. Matching Mechanism

4.1 Price Priority, Time Priority

Rules:

1. Better-priced orders execute first
2. At same price, earlier orders execute first

Example:

Bid queue:
  1. 09:00:01 bid @ $100.75
  2. 09:00:05 bid @ $100.75
  3. 09:00:03 bid @ $100.50

Seller market sells → First #1 executes, then #2

4.2 Continuous Matching vs Call Auction

Continuous Matching:

• Orders matched in real-time
• Used during normal trading hours

Call Auction:

• Collects orders over a period, then determines opening/closing price uniformly
• A-share opening auction (9:15-9:25):
  • 9:15-9:20: Can place and cancel orders (often fake orders to probe pressure)
  • 9:20-9:25: Can place orders, cannot cancel (true intentions revealed)
• Reduces price manipulation, increases opening liquidity

5. Market Makers

5.1 Role

• Continuously provide bid and ask quotes
• Provide liquidity
• Earn the Bid-Ask Spread

5.2 Market Making Strategy

Market maker quotes:
  Bid: $100.00 × 1000
  Ask: $100.10 × 1000

Spread = $0.10

If buy and sell orders are balanced:
  Buy 1000 shares @ $100.00
  Sell 1000 shares @ $100.10
  Profit = 1000 × $0.10 = $100

5.3 Risks

• Inventory risk: Holding too much one-sided position
• Adverse selection: Being "picked off" by informed traders

6. Market Microstructure

6.1 Determinants of Spread

6.2 Market Impact

Definition: The effect of large orders on price

Temporary impact: Price displacement during execution, recovers afterward Permanent impact: Order carries information, price changes permanently

Estimation formula (simplified):

Market Impact ≈ σ × √(Q / ADV)

σ = Daily volatility
Q = Order quantity
ADV = Average daily volume

6.3 Liquidity Measures

7. High-Frequency Trading (HFT)

7.1 Characteristics

• Extremely short holding periods (milliseconds)
• Large number of orders (thousands per second)
• Low latency (microseconds)
• High win rate, low profit per trade

7.2 Common Strategies

7.3 Technical Requirements

• Low-latency network (Co-location)
• FPGA / ASIC hardware
• Nanosecond-level clock synchronization
• Dedicated exchange interfaces

8. Importance of Latency

8.1 Latency Sources

8.2 Impact of Latency on Strategies

Strategy Type           Acceptable Latency
Core HFT Market Making    < 10 μs (microseconds)
Cross-market Arbitrage    < 5 ms (physical limit)
Intraday Quant Trend      < 10-100 ms
Daily/Swing Strategies    < 1 second

9. Practical Advice

9.1 For Non-HFT Strategies

1. Focus on liquidity: Choose liquid instruments
2. Split execution: Break large orders into smaller ones
3. Prefer limit orders: Avoid market order slippage
4. Avoid volatile periods: Opening/closing have poor liquidity

9.2 Slippage Control

# Simple slippage estimation
def estimate_slippage(order_size, avg_daily_volume, volatility):
    """
    order_size: Order quantity (shares)
    avg_daily_volume: Average daily volume
    volatility: Daily volatility
    """
    participation_rate = order_size / avg_daily_volume
    slippage = volatility * (participation_rate ** 0.5)
    return slippage

10. Characteristics of Different Markets

T+1 Settlement vs T+1 Trading: These are fundamentally different concepts. US T+1 settlement (effective May 2024) means fund and share delivery completes the next business day, but you can freely buy and sell the same stock multiple times within the same trading day (day trading). China A-share T+1 trading is a trading restriction: shares purchased today can only be sold on the next trading day, preventing same-day round trips.

Core Principle: Understanding the order book and matching mechanism is key to understanding why backtest results differ from live trading performance. Your orders don't execute in a vacuum—they compete with other orders in a complex market ecosystem.