How Market Makers Choose Quotes

A quote is four decisions, not one: where the centre is, how wide, whether it is symmetric, and for how much.

Bid=Fair value12Spread+SkewAsk=Fair value+12Spread+Skew\begin{aligned} \text{Bid} &= \text{Fair value} - \tfrac{1}{2}\text{Spread} + \text{Skew} \\ \text{Ask} &= \text{Fair value} + \tfrac{1}{2}\text{Spread} + \text{Skew} \end{aligned}

1. Fair value: where the centre sits

Your best estimate of the instrument's true price. It starts from the mid but rarely stays there: order book imbalance, recent trades, correlated instruments and any model you trust all move it. This is the hardest of the four and has its own lesson.

2. Spread: how much you charge

Wide enough to cover adverse selection, inventory risk and costs; tight enough to win flow against competitors. The main drivers:

  • Volatility. More uncertainty means more risk per fill, so wider.
  • Event risk. Around earnings or data releases, much wider or no quote at all.
  • Competition. More market makers on the instrument compresses it.
  • Your own uncertainty. If your fair value estimate is shaky, widen. Charging for your own ignorance is legitimate and necessary.

3. Skew: shifting both sides together

If you are long and want to sell, move both quotes down. Your offer becomes more attractive, so you are more likely to sell; your bid becomes less attractive, so you are less likely to buy more.

Note this is different from widening. Widening changes the distance between the quotes; skewing moves the whole thing while keeping the width. They solve different problems, and confusing them is a common mistake in market making games.

4. Size: how much you show

Larger size wins more flow and accumulates inventory faster. Most market makers quote smaller in volatile conditions and in instruments where they are less confident.

Worked example: one quote, three adjustments

Fair value 100.00, normal conditions, and you want a four-cent spread. Base quote: 99.98 at 100.02.

Now you get filled on the bid, twice, and are long 2,000. You skew down by one cent: 99.97 at 100.01. The offer is now more likely to trade and the bid less so, which pulls inventory back toward flat.

Then a news headline lands and volatility spikes. You widen to eight cents while keeping the skew: 99.95 at 100.03. Fewer fills, but each one now compensates you for the elevated risk.

Key takeaway

Widening changes the spread; skewing shifts the centre. Volatility widens, inventory skews. Keeping these two responses distinct is the core mechanical skill of quoting.

The tension underneath

Every quote is a bet on the composition of the flow it attracts. Tighter quotes win more trades, but they win disproportionately from the traders who most want to trade, and those are the ones most likely to be informed.

That is the uncomfortable structure of market making: the flow you attract most easily is the flow you least want. Handling it well means being tight enough to capture the uninformed volume while reacting fast enough to avoid the informed. See hit ratio vs spread for the arithmetic, and adverse selection and the spread for where the width comes from in the first place.

Tip

In an interview market-making game, say what you are doing and why: "I'm long, so I'm skewing down a tick" demonstrates the reasoning even when the number is debatable. Silence gets no credit for good instincts.

Test your knowledge

A market maker puts fair value at $100. The market has turned volatile and they are long more inventory than they want. Which quote adjustment reflects sound principles?
Widening and skewing are often confused in market making games. What is the difference, and what does each respond to?