Adapting to Flow: Dynamic Quoting
Static quotes lose money. Fair value moves, volatility changes, inventory accumulates, and a quote that was correct a minute ago is now an invitation. Dynamic quoting is the practice of continuously adjusting all four controls in response to what the market is doing.
The controls, from earlier lessons: centre (fair value), width (spread), skew (inventory), and size.
Responding to sustained buying
Suppose aggressive buying persists and you are being lifted repeatedly. The wrong response is to keep quoting the same market and keep selling into it.
The right response combines three moves:
Raise the centre. Sustained buying is evidence your fair value is too low. This is the most important adjustment and the one most often skipped, because it means admitting your estimate was wrong.
Widen a little. Directional flow is more likely to be informed, so charge more for it.
Reduce offer size, keep bid size. Slow the accumulation of a short position while staying available to buy back.
Notice you are now short from selling into the buying, so skew enters too: you want to buy, so lift both quotes, which is conveniently the same direction as the fair value update.
When flow is one-sided, move your price rather than just your size. Holding a stale centre and simply quoting smaller means losing slowly instead of quickly.
The core judgement: information or noise
Every adjustment rests on one inference. Is this flow telling me something, or is it random?
Signs of information: persistent one-sided flow, price not reverting after trades, related instruments moving the same way, timing near a news event, large or institutional-sized orders.
Signs of noise: alternating buys and sells, price reverting after each trade, no movement in correlated instruments, small and evenly distributed sizes.
Over-reacting to noise means chasing your own tail, moving your price up on a couple of buys, then back down, giving away the spread each time. Under-reacting to information means being run over. Calibrating this is the skill that most distinguishes experienced market makers.
Regimes need different settings
Quiet markets: tight spreads, larger size, minimal skew. Volume is the profit source.
Trending markets: wider spreads, faster fair value updates, aggressive skew. The danger is accumulating a position against the trend, which happens automatically if you quote symmetrically while price drifts.
Volatile markets: much wider, much smaller, and prepared to stop entirely.
Around scheduled events: pull or go very wide before the release. There is no edge in being the resting quote when a number prints.
Automating the response
Most of this is codified in production systems, because the reaction times required are beyond human capability and the rules are explicit enough to write down:
A human trader's role shifts toward setting parameters, handling situations outside the model's assumptions, and deciding when to override or shut down. That last responsibility is the important one: the model is calibrated on normal conditions, and the moments that matter most are the ones where normal conditions no longer hold.
Practise the full loop in the market games: quote, watch the flow, decide whether it is information, and adjust all four controls. The P&L breakdown afterwards will show whether you were reacting or chasing.