Part VI · Dealers · Chapter 6A

Who is on the other side

The dealer's business

Every option you have ever traded had someone on the other side: not a mysterious "market", but a specific business with specific obligations and specific behaviors. The mechanical decisions that business is forced to make, second by second, add up to one of the largest non-fundamental forces acting on the S&P 500 on any given day with no news.

This part builds that picture from zero, assuming only what Parts I to V provide: what an option is, and what its Greeks measure. Mathematics appears when it earns its place, and every formula is followed by what it means in plain terms.

1 · What this part covers

The five chapters of this part follow one dealer through one trade:

1. Chapter 6A, this chapter: who is on the other side of your option order, and why.
2. Chapter 6B: the exposures, Greek by Greek, that the dealer acquires when it sells you the option.
3. Chapter 6C: delta hedging, how the dealer mechanically neutralizes the directional risk.
4. Chapter 6D: gamma, why the hedge does not stay neutral, and why this drives the whole mechanism.
5. Chapter 6E: how often to re-hedge, when every rebalance costs a spread.

The parts after this one build on it. Part VII derives the formal mathematics, showing why delta hedging is the only strategy consistent with the option's price. Part VIII adds up the per-position dynamics across the whole market, and Part IX applies the framework to the VIX, same-day options and reading the tape. By the end of Part IX you should understand, formally, why a rising VIX forces SPX selling and why 0DTE options have changed the intraday character of the market. You should also know what gamma exposure measures, and where the conventional framework for it breaks.

2 · How much dealer hedging matters

Much of what circulates on social media and retail trading platforms about "dealer flows" and "gamma squeezes" is qualitatively right but quantitatively overstated. The academic literature gives effect sizes in tens of basis points, not full percentage points. Where this part cites numbers, they come from peer-reviewed work or directly from Cboe's own data, and where industry sources disagree with academia, the text says so. The goal is calibration: knowing not just that dealer hedging matters but how much it matters, and when.

The two traditions this book draws on weigh this force differently. The dealer-mechanics tradition of this part treats dealer hedging as a measurable force, modest most days and large in stress. The edge-by-edge case of Part X is more skeptical.

Part X cites a 2025 Cboe study estimating 0DTE market-maker gamma hedging at about 0.2% of SPX daily liquidity. That study measures same-day hedging only, so applying it to dealer hedging in general, or to levels built mostly from longer-dated open interest, extends the measurement. The Edge Book uses the sign of dealer gamma only as low-confidence context (Chapter 10H). It also rejects trading around gamma-exposure levels in a retail account (Chapter 10I). Read the chapters that follow with both positions in mind.

With those caveats in place, the chapters of this part are some of the most important framework material in the book.

3 · The market maker is not betting against you

Begin with a single small trade: you buy one SPY call option, strike $600, expiring in 30 days. You click "buy" in your broker's app, and a millisecond later you own the contract. Somewhere, a market maker now holds the opposite of what you hold: it is short a call. Who is it, and why did it take the other side?

This is the most important point in this part. A market maker, called the dealer in options, is not a speculator on the other side of your trade. It has no view on whether SPY goes up or down. Its entire business is to quote two prices at once, a bid at which it buys and an offer at which it sells, and to capture the spread between them.

A $10 round trip is small until you multiply it by 100 contracts a second, across every option in every name. Then a franchise emerges, built on scale rather than on being right about direction.

In US-listed options, a small number of firms dominate this business. The table below summarizes the shares that Chapter 2B gives from 2025 and 2026 data, where the denominators behind them are spelled out.

The Big Three options wholesalers, Citadel, SIG and Wolverine, together execute roughly 85% of retail options order flow. Beside them sit the dominant institutional liquidity providers: Jane Street, Optiver, IMC and DRW. Jane Street's net trading revenue alone, $20.5 billion in 2024 followed by $39.6 billion in 2025, shows how thick the layer of intermediation is between you and any counterparty with an actual directional view.

4 · The problem the dealer just created

Back to the trade: you bought a call and the dealer sold it, so the dealer is now short a call. It loses money if SPY rallies and makes money if SPY falls. It holds a directional exposure that it did not want and is not paid to hold.

Carrying that exposure overnight is unthinkable, because the dealer captured a spread of about 10 cents a share while a modest move against the position costs many times that.

The figure below sells 100 such calls and runs the book through 400 simulated days, with and without the delta hedge. Compare each histogram with the thin line of the spread, then raise the day's volatility and watch the hedged book's losses on large moves grow.

day's volatility

1.00% a day

new days

spread kept = $1,000

no hedge: typical swing $32,200

hedged: typical swing $3,355

A 1.00% rally costs the unhedged book $32,019, 32 times the spread. Hedged, it leaves the book $1,120 ahead, spread included.

So the dealer neutralizes the directional exposure at once and mechanically, by delta hedging. Delta hedging, the subject of Chapter 6C, is the engine that drives nearly everything in this part and the three after it.

5 · One trade, end to end

Walk through what happens when you click "buy" on a single SPY $600 call. The chain is mechanical, fast and almost entirely invisible to you, with times measured from your click in milliseconds (ms).

1. At 0 ms, you click "buy" at the NBBO offer of $3.05, sending a marketable limit order at $3.05 for one contract.
2. At 1 to 3 ms, your broker's routing engine decides which venue gets the order. For SPY, an equity option, that is almost always a wholesaler: Citadel Securities, Susquehanna or Wolverine. SPX and XSP index options route directly to Cboe instead (Chapter 2B).
3. At 3 to 5 ms, the wholesaler's internal matching engine checks its principal book, the inventory it trades for its own account, where its own mark on the call is $3.03. It fills you at the NBBO, $3.05, which gives you no price improvement against the public offer and gives it $0.02 a share over its internal mark. That $0.02 is its actual edge.
4. At 5 to 10 ms, the wholesaler is short one SPY $600 call, and its risk system computes the delta at the current spot as 0.58. That makes it short 58 share-equivalents, so the system fires a hedge order in SPY shares, typically batched with hundreds of other flows arriving at the same moment.
5. At 10 to 50 ms, the wholesaler buys 58 SPY shares, or proportionally fewer if opposite flow already offsets part of the exposure. The hedge typically executes at or inside the NBBO, through internalization, a lit exchange or an alternative trading system (ATS). The principal book is now delta-flat with respect to your trade.
6. One day later, the trade settles. You hold one long call, with no shares delivered until exercise or expiry, while the wholesaler's offsetting short is registered at the Options Clearing Corporation (OCC). From then on your counterparty risk is on the OCC, not on the wholesaler.
7. Throughout the option's life, the wholesaler delta-hedges continuously as spot moves, with gamma, vanna and charm all feeding in (Chapter 5D). Its profit and loss (P&L) over the trade's life depends on whether realized volatility along the hedge path exceeds the implied volatility (IV) built into your $3.05 fill. That comparison is the gamma-theta arithmetic of Chapter 5C.
8. At expiry, if the call expires worthless, you lose the premium and the wholesaler keeps it, less its hedge P&L. If it expires in the money (ITM), SPX and XSP settle in cash, while SPY settles physically, with shares changing hands.

6 · The franchise at scale

One trade looks negligible: $0.02 of edge on a one-contract order is $2, so it is volume that makes the business. Citadel Securities discloses executing roughly 35% of all US retail equity and options volume. On a typical day that is billions of share-equivalents and tens of millions of options contracts running through its book.

The whole US options market traded approximately 13.8 billion contracts in 2025. If Citadel handled 35% of the retail share, and a smaller but still substantial fraction of institutional flow, its annual options volume runs to billions of contracts. The franchise's revenue, in simplified form, is the edge on each contract times the contracts traded, less what the dealer spends:

Here $e_i$ is the edge per contract on product $i$ and $n_i$ the number of contracts traded in it. Hedge slippage is what the dealer loses between the price it expects for its hedges and the price it gets.

Per-contract edges run from about $0.50 to $2.00 across the retail options book. With hundreds of millions of contracts a year, the franchise produces multi-billion-dollar revenues consistently. The arithmetic is dispiriting if you imagine taking the other side, and gratifying if you are the wholesaler. Jane Street's $20.5 billion in 2024 and $39.6 billion in 2025 are not aberrations: they are the profile a high-volume franchise with a small edge per trade tends toward at scale.

7 · Risk capital and alpha capital

Two kinds of capital are at work in any market-making operation. Telling them apart shows what dealers care about and what they do not.

Risk capital is the collateral the dealer must post against its open positions, sized to cover the worst loss it could take before it can unwind them. Regulation after 2008 (Basel III, Dodd-Frank and EMIR) made this constraint binding through stress tests. A modern dealer's biggest operational constraint is not how much it can make per trade. It is how much capital each book consumes per unit of expected revenue.

Alpha capital is what a firm uses for active directional or relative-value bets: outright views the desk holds. Market makers run negligible alpha capital relative to their risk capital, since they are nominally flat in direction. In practice every dealer book carries small residual exposures that it chooses not to hedge perfectly: to skew, to the term structure, to basis and to correlation. These residuals are where the real edge of a well-run dealer book lives, and where the franchise can lose money in stress.

If you read that "Citadel Securities lost $3 billion in March 2020", the loss almost certainly refers to the residual book, not the spread franchise. The spread franchise compounds quietly through every regime, while the residual book produces the headline losses (and the headline gains in benign years).

8 · Why the dealer takes your trade

Retail traders rarely ask why the dealer is willing to take the other side of their trade in particular, and whether it worries that they might know something.

The honest answer is that, in aggregate, retail flow is informationally uninformed. A wholesaler internalizing $50 billion of retail flow a month does not worry that the median order knows tomorrow's news. It worries about the tails: the rare retail order that is informed, through insider activity or leaked earnings data. Those orders are flagged and routed to lit exchanges rather than internalized. Routing them out transfers the adverse selection risk to whichever venue takes them.

For institutional flow the calculation is different, because a hedge fund placing a $200 million SPX options trade is presumed informed. The dealer quotes wider, or refuses, and the order goes through the order book directly. This is why "institutional spreads" on large index options trades are often 10 to 30 bps wider than the SIP NBBO that retail sees. The Securities Information Processor (SIP) is the public feed that computes that NBBO.

The structural lesson: retail order flow is profitable to wholesalers because it is uninformed in aggregate. The wholesaler is not paid for taking risk against a smart counterparty, because it hedges at once. It is paid for providing immediacy, and for capturing the spread against the noise that retail order flow represents in aggregate.

9 · Check yourself