[roflwaffle]

The last EPA link has more detail.

In fuel consumption terms, the new regression line tilts relative to the previous analysis. Most label values are likely to be unaffected except those with higher MPG values, where the difference is most pronounced. Some high MPG vehicles may see a decrease in label values if they continue to use the derived 5-cycle values, but of course the actual impact will depend upon the results achieved with complete 5-cycle testing and whether a manufacturer opts for derived 5-cycle methods for determining label values. The difference relative to the previous equation is more pronounced for the highway derived 5-cycle value. Some low MPG vehicles may actually see an increase in label values, but the difference relative to current values is so small and subtle that any change would simply be the result of changes due to rounding of MPG values. The following figures illustrate the new versus old equations, both in terms of fuel consumption and MPG. Further, as a result of the equations providing a better fit with current vehicles, we expect a decrease in the volume of failures of the “litmus test” with a corresponding increase in the number of vehicles eligible to use the derived 5-cycle values.