Forecasting in Financial and Sports Gambling Markets: Adaptive Drift Modeling

This book discusses cointegrated time series associated with financial and sports gambling markets are analyzed in terms of time-varying parameter models. Modeling premises are that present and past disequilibria—shocks both within and between time series—may affect subsequent changes and rates of these changes within individual series and sufficiently large shocks may disrupt/alter model structure such that resulting forecasts may be temporarily unreliable. Reduced forecasting equations are in terms of higher order ARMA models that are not limited to bilinear processes. Sports forecasting models based on public information are usually more effective—in terms of profitable trading/wagering strategies—than those for the financial sector for two reasons: insider information is less prevalent, and modeling is simplified since lagged shocks associated with the gambling lines/spreads are known—in contrast with financial modeling where there are no comparable gambling shocks, only unknown, lagged statistical shocks in terms of MA variables. Forecasting is illustrated for NFL and NBA playoff games. In financial markets, cointegration is discussed in terms of candlestick chart variants with modeling illustrations given in terms of recent Google price changes.

Chapter coverage includes candlestick charts, higher order ARMA processes in financial markets, the effects of gambling shocks in sports gambling markets, cointegrated time series with model drift, modeling volatility, and the promotion of financial and mathematical literacy.

Keywords: exploratory drift modeling, cointegrated time series, higher order ARMA processes in financial markets, gambling shocks in sports gambling markets, cointegrated time series with model drift, candlestick charts, algorithmic trading, candle stick charts, financial modeling and sports, financial economics, time series analysis, sports forecasting models