So in the stats, Red is bad, yellow is good, on a row-by-row basis? But something should be done about the scaling of the #s on the stats page. It seems that

So in the stats, Red is bad, yellow is good, on a row-by-row basis? But something should be done about the scaling of the #s on the stats page. It seems that each row is a factor based on the row below. But : Look at the lines 1715f0128906 -> fca7904c77ce, where an intentional slowdown was inserted, and then reverted. The Red looks far worse than the Yellow that follows (say, +3782% -> -97.42%), whereas (I assume) that they are equivalent, but opposite. Perhaps the differences could be (t[n+1] - t[n])/t[0], so that the numbers sum up?