Execution transparency thread
A market‑making shop shared practical execution‑ops details—sub‑second data sync, deduped logs and real‑time monitoring—to prevent quote drift and manage inventory risk in fast markets. The thread lays out operational controls relevant to desks that handle CEX/DEX or sub‑second quoting. (x.com)
Market making is the business of posting buy and sell prices all day, then surviving the gap between what the market is doing and what your systems think it is doing. A thread from Fibonacci HFT focused on that gap, describing the controls it uses to keep quotes, fills and inventory aligned in sub-second markets. (x.com) The firm’s examples centered on three operational checks: synchronizing market data across venues in under a second, deduplicating logs so one event is recorded once, and watching positions and quote health in real time. Those checks are meant to catch “quote drift,” when a posted price lags the live market, and to stop inventory from building unnoticed after fast fills. (x.com) In plain language, a market maker earns the spread between its bid and ask, but it also accumulates inventory when one side trades more than the other. Academic and industry work describe that inventory risk as the core tradeoff of market making: earn small spreads repeatedly while avoiding a large position that moves against you. (faculty.haas.berkeley.edu) (rootstone.io) Latency is the other half of the problem. Research on market making under exchange delay says even small lags between data, decision and order arrival can turn a sensible quote into a stale one, leaving the firm to buy too high or sell too low after the market has already moved. (arxiv.org) (quantstrategy.io) That is why the thread spent so much time on plumbing instead of strategy. If a centralized exchange feed, a decentralized exchange fill, and an internal risk engine are even briefly out of sync, a desk can think it is flat when it is long, or think an order is live when it has already been canceled. (x.com) (liquidview.app) Deduplicated logs matter because ultra-fast systems generate multiple messages for one trading event: market data updates, order acknowledgments, cancel confirmations and fills. A clean event trail lets operators reconstruct what happened during a burst of activity instead of chasing duplicate records that make a position or latency spike look worse than it is. (x.com) (ddn.com) Real-time monitoring is not just an engineering preference. In U.S. equities, Securities and Exchange Commission Rule 15c3-5 requires brokers with market access to maintain risk management controls and supervisory procedures designed to manage financial and regulatory risk from automated trading. (law.cornell.edu) (sec.gov) Europe has moved in the same direction. On February 26, 2026, the European Securities and Markets Authority published a supervisory briefing on algorithmic trading under MiFID II that pointed regulators and firms toward tighter expectations on governance, stress testing, outsourcing, pre-trade controls and monitoring. (esma.europa.eu) (hoganlovells.com) The thread’s practical value is that it translated those broad obligations into desk-level habits: keep clocks aligned, make one source of truth for events, and alert on inventory and quote anomalies before they become losses. For firms quoting across centralized and decentralized venues, those controls are the difference between a fast market and a blind one. (x.com) (esma.europa.eu)
