Quoting Around Fair Value
The basic structure of a quote is simple:
Buy below what you think it is worth, sell above. Everything interesting is in how far, and in what happens when your fair value estimate is wrong.
Three lines of arithmetic, and it is what a market maker does all day before any instrument is named. The edge you ask for is the cushion on every fill, and twice it is what the rest of the market sees as your spread.
The three ways a quote goes wrong
Wrong centre. Fair value is 100.00 but you quote around 99.95. Now your offer at 99.97 is below true value, so informed traders lift it all day and you are short at bad prices. A centring error is the most expensive mistake because it loses on every fill, not just some.
Too tight. Correct centre, but a spread narrower than your uncertainty. You trade constantly and lose slightly on average, with the damage concentrated in the informed flow you attracted.
Too wide. Correct centre, safe spread, no fills. Nothing lost directly, but no revenue and lost queue position and market share.
Ranked by cost: a wrong centre is worst, too tight is next, too wide is merely unproductive. Spend your attention on fair value before spread optimisation.
Tick constraints
You cannot quote arbitrary prices. Instruments trade in ticks, the minimum price increment, and this binds more often than newcomers expect.
If the tick is one cent and your ideal half-spread is 0.4 cents, you cannot express it. Your options are a one-cent spread (tighter than you want) or two cents (wider). In tick-constrained instruments, the spread is frequently pinned at one tick and competition moves to queue position instead, which is why price-time priority matters so much in liquid names.
Where the tick is small relative to the instrument's value, spread is the competitive dimension. Where it is large, queue position is.
Quoting inside the market
If the current market is 99.98 at 100.02 and you quote 99.99 at 100.01, you have improved both sides and now sit at the front of both queues. You will trade a lot.
Whether that is good depends on why the existing market is wide. If it is wide because competitors are lazy or slow, you have found free flow. If it is wide because the instrument is genuinely risky right now, you have just volunteered to take on risk that better-informed participants declined.
Before improving on the best price, ask why nobody else is there. In a competitive market, an unusually wide spread is usually information rather than opportunity.
Keeping the quote current
Fair value moves continuously, so quotes must follow. The engineering problem is that updating costs queue position, so you cannot re-quote on every tiny change.
The usual resolution is a tolerance band: leave the quote alone while fair value stays within some distance of the quote's centre, and re-quote when it drifts beyond. Wide bands mean stale quotes and adverse selection; narrow bands mean constant cancellation and poor queue position. Where you set that band is a real parameter with real P&L consequences.
Two-sided obligation
Formal market makers often have exchange agreements requiring continuous two-sided quotes of a minimum size and maximum width, in exchange for lower fees or rebates.
This constrains the obvious response to danger. You cannot simply pull quotes when conditions get bad, so you widen to the maximum permitted and reduce size instead. Understanding those obligations is part of understanding why spreads behave the way they do during stress.