Backtests run over longer time periods, and graphed on the typical arithmetic scale, do not provide a satisfactorily visualization of the results. Significant moves in the early portion of the graph are "cramped" because the scale range must be increased to show later results.

Example: Start w/ $100,000. A simple increase of 25% per year over 10 yrs requires an arithmetic range from $100,000 to $1,000,000. A significant increase of 10% ($10,000), early on, is only 10,000/900,000 = 1.1% of the scale -- totally undistinguishable from a visual standpoint. On a log scale the same 10% move from $100,000 to $110,000 looks the same (as it should) as a 10% move from $900,000 to $990,000.