the very first time we quote the sigma/odds ratio (1.7\sigma) we don't say it's vs the aligned distribution, but we say in the next sentence with the +10 detection. To be crystal clear, we could write:
If binary black hole spin magnitudes extend to high values we show that the data already exhibit a 1.7σ (0.087 odds ratio1) preference for an isotropic angular distribution with respect to an aligned model. By considering the effect of 10 additional detections of systems, we show that if all observations come from a single population such an augmented data set would enable at least a 2.9σ (0.0035 odds ratio) distinction between the two models, and in most cases better than 5σ (2.9 × 10−7 odds ratio).
From Alberto: