To be fair, the correlation threshold they used was 0.7 for confusable, and 0.3-0.7 for contextually confusable. But I definitely would have liked to see some examples of glyph pairs at around 0.5 correlation. And at small font sizes realistic in actual threat scenarios.