Near misses are ONLY "okay" to the extent that the outcomes are benign.
I couldn't agree more with the idea that near misses are an excellent predictor of serious accidents.
The main point was that speed limit infringement is NEITHER a good predictor of a near miss nor a serious outcome.
OTOH, when we can detect near misses we have an excellent chance of correcting the dangerous behaviour before the serious outcome - and not least because near misses outnumber accidents (on the road) by something in the range 5:1 to 15:1.