I wouldn't really go off the roll out numbers. I just listed them since that's what some people quote, so we're just give you all the data.
I also want to mention these are not numbers from the screen. The screen seems to be optimistic compared to the actual raw data, (meaning interpolating to get the exact time at 0 MPH and 60 MPH (like PerfDB does), not the software's calculation. For example, we upgraded a few non power bits later in the day when it was warmer and cut at 3.6 0-60 as indicated by the screen. According to the actual data, this was 3.735 seconds and 3.380 seconds with 1-FT roll out. Interestingly enough if you average those two results and round up to the nearest 10th, you get 3.6.