Answers are grounded in this paper and cite their sources.
Cointegrated time processes measuring NFL playoff game performances relative to the betting spreads are graphed in terms of candlestick charts and forecast in terms of autoregressive systems with time varying coefficients. Coefficients are modeled in terms of linear regressions on lagged shocks. Estimation is non Bayesian. Forecasts provide measures of market efficiency/inefficiency and outcome volatility. Risk assessment utilizes GARCH-type modeling in estimating volatility. Applications are presented for the New York Giants 2012 playoff games based on a data backlog of three years.
No takes yet. Share an insight, caveat, or question.
William S. Mallios (2013) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: