I wonder, would it be useful if we could have a dedicated thread to compile where we think pacenotes are late - or incorrect. Maybe they could be upvoted, maybe a Google docs created or something?
For me, it's the inconsistency which means you cannot rely on them. Most are fine but it's just that one or two that are late, and a couple more are just wrong. If we just took the existing notes and made them earlier as a blanket adjustment, then most would be very early.
I guess we could look on the racenet analysis for distance markers?
Or maybe the meta data could show if there were common corners that many people crashed at which might point to a poor call?
I'd be interested to understand how they are triggered. It feels like a distance thing rather than a time thing which is why some come late especially with multiple instructions