gp is right,

there was 8/23 wrong classifications of "Bottled vs Soma" vs. 10/23 wrong classifications of "Tap vs Soma"

In Figure 6 , the author shows there's no statistically significant difference between "Bottled vs Soma"

So like he wrote :

> Soma filtered water performed the worst, having a taste statistically indistinguishable from tap water

I could write using the same data:

Soma filtered water performed the best, having a taste statistically indistinguishable from bottled water