Asymmetry: Percentage of forecasted values that PND-1186 are greater than the respective actual values, time frame 12/08/2011 to 01/04/2012, all horizons and all turbines.WPPTGWPPTGWPPT-AL1 StepTurbine A49.50%50.33%69.76%Turbine B53.20%57.29%79.51%Turbine C53.65%54.25%76.99%Turbine D53.90%56.90%80.87%72 StepsTurbine A55.26%56.25%71.02%Turbine B57.86%61.09%71.07%Turbine C57.09%58.01%67.33%Turbine D57.32%59.51%70.03%144 StepsTurbine A58.39%60.62%71.34%Turbine B56.58%59.43%67.08%Turbine C54.22%55.37%63.92%Turbine D55.18%56.81%66.23%216 StepsTurbine A54.85%55.38%69.92%Turbine B58.76%61.88%68.62%Turbine C55.49%56.04%64.50%Turbine D56.94%57.41%67.01%Full-size tableTable optionsView in workspaceDownload as CSV
After all, the goal of GWPPT-AL is not to produce precise forecasts, but to maximize monetary profit. Therefore, Table 8 presents projected yearly monetary gain of hypothetically switching a) from WPPT to GWPPT and b) from GWPPT to GWPPT-AL. While these gains occasionally become negative in the most imprecise case of 216 steps ahead (36 h), most of the times the increase in profit is substantial. Looking at the more reliable forecasting horizons 1 step (10 min) and 72 steps (12 h), GWPPT can increase profit over WPPT by up to 14,000 €, just by being the more precise forecasting model. However, GWPPT-AL can gain another up to 67,000 € over GWPPT per Turbine. Additionally, the gain of GWPPT-AL over GWPPT is by far greater than lipids of GWPPT over WPPT. As GWPPT provides more monetary profit than WPPT and GWPPT-AL more than GWPPT, the comparison WPPT vs. GWPPT-AL is omitted here due to predominant redundancy.