That's only speculation, not based on evidence. The cymbal that started this whole discussion was a blue label. Maybe Zenstat could do a statistical analysis of this phenoma and then get back to us in ablout a decade.
Just so everybody is clear on the boundaries and nature of my pricing study, it won't ever contribute anything to our understanding of breakage. It doesn't provide any information on rates of breakage either, which is why you won't find me commenting on rates of breakage.
The purpose of my study is to get a data set which I can use to examine some particular estimation methods in pricing studies. I needed a data set which I can make freely available to other analysts, so they can replicate my results or apply different estimation methods. It doesn't matter whether the data is about cymbals or not. I just happened to look at cymbals because I was interested in them.
The reason the study has gone on for a number of years is that one aspect of the estimation methods includes time series data -- for which you need enough years of data. It is a happy coincidence that Paiste re-issued the 602s along the way so that we can also look at what effect this had on the prices of the older ones. It is a further happy spinoff that because I collected weights (as one input to the predictive price modeling) I have been able to supply lots of weights for almost all of the different models of 602s and Sound Creations. And finally, it is great for me that I've been able to supply well documented information to people for insurance purposes, pricing purposes, and even to the estates of famous drummers so they know what to sell the collection for. I take great pleasure in making data useful.
A year or two into my data collection I noticed that severely damaged cymbals (beyond your flea bite, moderate keyhole, and the like) usually fetch $30 - $50. In order to simplify my data entry task I quit keeping track of them. They are excluded from the survey. You can read this exclusion in the text of the tables. Thus no useful information. Outside the scope of the research.
Getting data from eBay completed sales (and other sales channels) would be a very poor way to try and estimate the proportion of cracked and severely damaged cymbals. Thus no useful information. Outside the scope of the research.
The study of why cymbals crack is an entirely different one from price modeling. It would take place in the materials science lab, not on eBay. That is outside my area, although I'm always happy to consult on matters of experimental design and data analysis.
The suggestion of a non destructive method for estimation of element (copper, tin, gold, silver, phosphorus, etc) proportions is an interesting prospect. If somebody with access to the equipment does the lab work I'd be delighted.