CCD's being faster than CMOS is not a myth. Saying that CMOS have caught up and are faster than CCD is the myth. Everyone has become brainwashed into thinking CMOS is more sensitive because of the beautiful images full of rich detail that they can produce, not realizing that in fact they are taking way longer to get that image. The problem is that all the images that are posted from CMOS cameras are stacks, while all the images posted from CCD cameras are single exposures. That is not a fair comparison. I have done a bucket-load of tests, with dozens of camera models, and for a single exposure CCD's still produce the best SNR...reworded another way, CCD's produce the shortest Time To Observable Image. Specifically the ICX828 CCD still blows everything else away, by more than a factor of 10. Slowly I am putting all my testing together to present to the community, it is just taking a lot of time to do correctly. Please stay tuned, I am on it!
My empirical judgement is that most sensitive CMOS takes double the time of the most sensitive CCD. I looked at what folks were doing with ASI224 a long time and compared it to times for my CCD Revolution2/LN300. The new higher resolution CMOS look much better, but requires a more accurate mount because of the doubling of required exposures. That said, I hope to upgrade to CMOS soon after being on sidelines for a couple of years.