CPAlgorithms: Use SIMPLIFIED systematic model for Run 2 electron isolation

For whatever reason, the electron isolation SF ROOT files that are linked from the recommended Run 2 R21 map

ElectronEfficiencyCorrection/2015_2018/rel21.2/Precision_Summer2020_v1/map4.txt

have bin errors equal to their bin content. Since this is how the isolation SF uncertainty is determined when using the "TOTAL" correlation model, this results in 100% uncertainties. While we wait for proper R25 recommendations, I'm switching it to the "SIMPLIFIED" model (which is actually a better representation of the relevant sources of systematic uncertainties, and leads to more systematic terms). This is not needed for the reconstruction and ID SFs, so I'm not putting it there. It's also currently not available for Run 3.