Realistic market making backtester with tick-by-tick L2/L3 replay, FIFO queue simulation (iceberg detection + cancel inference), latency modeling, maker/taker execution, multi-asset support, realistic fill simulation and detailed adverse selection metrics. Built for strategy research, focusing on real queue dynamics and fill quality.

Game theory meets actuarial science: Rothschild-Stiglitz screening, bonus-malus incentive design, and mixed-strategy fraud audit games — with closed-form equilibria, numerical solvers, and Monte-Carlo validation in Python.

Replication of Avellaneda & Stoikov's market-making model with closed-form quote formulas, inventory vs symmetric strategy comparison, and an adverse-selection extension.

LLM agents learn to lie in a dating marketplace and honest agents quit. A multi-agent simulation proving AI can exploit information asymmetry.

Picked Off — a market-making game. Quote against hidden fair value, survive informed flow. Glosten-Milgrom under the hood.

A Python simulator for studying passive quoting, inventory control, adverse selection, markouts, and risk limits in market making.

Limit Order Book Matching Engine