Under intermediate volatility and limited samples, model-free DDPG finds dynamic-arbitrage strategies more reliably than SLSQP run on noisily estimated Almgren-Chriss impact parameters, even though the latter knows the true functional form.
Universal price impact functions of individual trades in an order-driven market
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-fin.TR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Can Reinforcement Learning Efficiently Discover Price Manipulation?
Under intermediate volatility and limited samples, model-free DDPG finds dynamic-arbitrage strategies more reliably than SLSQP run on noisily estimated Almgren-Chriss impact parameters, even though the latter knows the true functional form.