1.5 Where Trades Actually Happen

Understand the actual venues where your orders get filled, why "free" trading isn't free, and why prices can briefly differ across exchanges.

Layer 1: Market Plumbing — Chapter 5 Goal: Understand the actual venues where your orders get filled, why "free" trading isn't free, and why prices can briefly differ across exchanges.

The Core Idea

When you click "buy AMD" on your broker, your order doesn't magically appear on NYSE. It goes through a routing process that picks from dozens of possible venues. Each venue has different rules, fees, and participants. Understanding this matters because it affects your fill quality and explains some "weird" market behavior.

The Major Exchanges

NYSE (New York Stock Exchange)

Oldest, most prestigious
Listings: Berkshire Hathaway, JPMorgan, Disney, Coca-Cola, McDonald's
Hybrid model: floor traders + electronic
Generally large, established companies

NASDAQ

All electronic from day one
Listings: Apple, Microsoft, Amazon, Google, Tesla, Nvidia, AMD
Tech-heavy
The "growth" exchange

CBOE (Chicago Board Options Exchange)

Major options exchange
Owns BATS (now part of CBOE Global Markets)
Multiple equity exchanges under one roof

Smaller Exchanges

IEX (slower by design — fights HFT advantages)
MEMX, MIAX, LTSE — newer entrants

Reality: A stock listed on NASDAQ can be traded on dozens of venues. The listing exchange doesn't restrict where trading happens.

ECNs (Electronic Communication Networks)

What they are: Electronic systems that match buyer and seller orders directly, bypassing traditional exchanges.

Examples: ARCA (NYSE-owned), Direct Edge, Instinet.

Why they exist: Faster, cheaper, more anonymous than traditional exchanges.

Who uses them: Institutions, HFTs, sophisticated retail (through certain brokers).

Dark Pools

This is one of the most important and least-understood parts of the market.

What They Are

Private exchanges where institutions trade large blocks of stock without showing the orders to the public market.

Why They Exist

If a hedge fund wants to buy 1 million shares of AMD, posting that order on NASDAQ would:

Move the price against them immediately
Tip off HFTs and other traders
Cost them millions in slippage

So they use dark pools to find counterparties privately.

Examples

Sigma X (Goldman)
UBS ATS
Credit Suisse Crossfinder
Liquidnet
Citadel Connect

How Big?

~40% of US equity volume now happens in dark pools or off-exchange venues.

That means almost half the trading happening doesn't show up in the public order book you see.

Why You Should Care

The "real" supply/demand picture is partially hidden
Big institutional moves often happen in dark pools BEFORE the public chart shows them
Services like dark pool prints track large block trades and can be leading indicators
When you see a sudden price move with no apparent catalyst — it may have been a dark pool transaction settling

PFOF (Payment for Order Flow)

Why "free" brokers aren't actually free.

How It Works

You place a market buy order on Robinhood
Robinhood doesn't send your order to NYSE
Robinhood sells your order to Citadel Securities (a market maker)
Citadel fills your order from its own inventory
Citadel pays Robinhood a small fee per share for the right to do this
Citadel makes money on the spread + information about retail flow

The Numbers

Citadel pays brokers roughly $0.0010 - $0.0020 per share
Robinhood made over $1 billion from PFOF in some years
That's where your "free" trading comes from

Is This Bad?

Arguments it's fine:

Retail gets tighter prices than they'd get on exchanges (in theory)
Brokers verify "best execution"
It enables zero-commission trading

Arguments it's bad:

Citadel sees enormous retail flow data
Conflict of interest — broker incentivized to route to highest payer, not best executor
Some studies suggest retail gets slightly worse fills than they'd think

Order received by Robinhood's systems
Risk check — do you have buying power? Is the order valid?
Routing decision — Robinhood's smart order router decides where to send it
Sent to Citadel Securities (PFOF partner) in this case
Citadel fills the order from inventory at the current best price
Confirmation sent back to Robinhood → to you
Trade reported to the consolidated tape (~milliseconds later)
Settlement scheduled for T+1 (next business day)

Total elapsed time: milliseconds to a couple seconds.

For limit orders or larger orders, the routing is more complex and may go to multiple venues.

Why Prices Can Briefly Differ Across Venues

The "NBBO" (National Best Bid and Offer) is the consolidated best price across all exchanges. In theory, it's the same everywhere. In practice:

Different exchanges have microsecond-level differences
HFTs profit from these tiny gaps (latency arbitrage)
For retail purposes, prices are effectively identical
But during fast moves, gaps can widen briefly

This is why HFT exists — and why retail can't compete on it.

Best Execution

Brokers are legally required to provide "best execution" — getting you the best available price.

How "Best" Is Measured

Price improvement (filling better than the displayed NBBO)
Speed of execution
Likelihood of execution
Size of execution

Reality

Brokers prove this through statistical reports (Rule 605/606)
Most retail orders get filled at NBBO or slightly better
The "improvement" is usually fractions of a penny

You generally don't need to worry about this for small swing trades. It matters for size.

Choosing a Broker

A practical guide for your situation:

For Beginners / Swing Traders ($9K-50K account)

Robinhood — simplest UI, free, fine for buy-and-hold and swing
Fidelity — better research, no PFOF on some products, more professional
Schwab (incl. ThinkorSwim) — best charting platform (TOS), good research
Webull — better charting than Robinhood, free

For Serious Traders ($25K+)

Interactive Brokers (IBKR) — best execution, lowest fees, professional tools, steep learning curve
Tradier — good for active swing/options traders
Tradestation — strong for algorithmic, good platform

For Algorithmic / API Trading

Alpaca — best free API for retail algo trading
IBKR — most powerful API, more complex
Tradier — middle ground

What Listing Exchange Tells You

When you see "AMD: NASDAQ" or "JPM: NYSE", what does it mean?

Listing requirements differ slightly between exchanges
NYSE has stricter rules (minimum income, market cap)
NASDAQ is more flexible — more tech-friendly
The exchange also benefits from the company's brand association

Trading-wise, it makes little difference. Both are deeply liquid for major names.

Practical Takeaways

Your order goes through many hops before getting filled. Each adds (tiny) latency and complexity.
Dark pools hide ~40% of volume. The chart you see is not the whole story. This is why "unusual options activity" and "dark pool prints" services exist — they try to surface the hidden flow.
PFOF is real but mostly invisible for swing trading. Don't worry about it.
Best execution is a regulatory requirement. Your broker is held to standards, but they're statistical, not per-trade.
Listing exchange barely matters for trading decisions.

A Mental Model

Think of the market as an iceberg:

Above water (10%): the public order book — bids, asks, trades on exchanges
Just below (40%): ECNs, alternative trading systems
Deep below (40%): dark pools, hidden orders, private liquidity
Underground (10%): OTC, foreign markets, after-hours private deals

You see only the tip. Most price movement starts beneath the surface and bubbles up.

Quick Self-Check

Before moving to 1.6, you should be able to answer:

What's the difference between NYSE and NASDAQ from a trader's perspective?
What is an ECN?
What is a dark pool and why do institutions use them?
What is PFOF and how do "free" brokers make money?
What's the rough sequence of events when you click "buy"?
What does "best execution" mean?
What percentage of US equity volume happens off-exchange?

Previous: 1.4 Market Participants Next: 1.6 Settlement and Clearing