I think it is possible that the chart is structured so that it visually exaggerates the differences in detection rates. The chart only shows the top 40% of detection rates, i.e., from 60% to 100%. If it showed the entire 100% from bottom to top, the small differences in detection rates would appear relatively smaller than they do in the chart as it currently exists. Ehh, I could be wrong, but I think the differences between products would feel less if one got to see the entire chart starting at 0% detection and going up to 100%. Honestly, maybe that would give one a visual impression that better matched real world differences. I will see if I can make a chart like that and somehow load it up. Really, there is a 6.7% detection difference from the top products to the bottom product in this test. Does that percentage difference really add up to anything in the real world? I have no idea. Having said that, I am sticking with products near the top of these and other tests.