- Driven by the Engle-Granger cointegration methodology.
- Scan any user-chosen group within and across geographies, sectors & industries.
- Optimizes the regression process. Valid Engle-Granger regression results depend on the variable order in which they are regressed. ArbMaker manages this.
- Automatic (or manual) lag selection. Ignoring lag selection distorts results.
- Filter and refine scan results by R², speed-of-mean-reversion, half-life, cointegrating coefficient (beta), t-values, proprietary data normality measures, residual axis crossings and 95% / 99% confidence intervals.
- Full residual analysis to establish the tradability of cointegrated pairs – knowing cointegration is present is only half the job.
- Graphical analysis tools include plots of residuals, residual Z-scores, Q-Q Normality, Variance to Model, Auto-correlation & Partial Autocorrelation Functions and Indexed Prices & Spread of the underlying securities.
- Selection of technical indicators and overlays available during residual analysis including Bollinger Bands (4 bands), Stochastic RSI, MACD and Hilbert SineWave studies.
- Choice of free Yahoo! historical and delayed data; or the real-time IQFeed subscription service.
