Name of Quantlet: Onchain Insights - Parametric quantile function characterisation
Published in: Onchain Insights - Parametric quantile function characterisation
Description: We evaluate the distribution of stablecoin depeg basis points. ARIMA/GARCH residuals show poor fit against Gaussian/Student's t innovations. In order to circumvent this issu we train a Neural Network to regress the quantile function in a non-distributional manner.
Keywords: Cryptocurrency, Blockchain, Stablecoins, Decentralized Finance, Liquidity, Depeg risk, quantile regression, Full distribution probabilistic forecasting, Extreme event forecast
Author: Owen Chaffard
Submitted: 29.04.2026
Datafile: USDC_USDT_hourly_metrics.parquet
