It's not an Anecdote. It's a fact.
I didn't mean to come across as insulting, or as questioning your truthfulness. To me, an anecdote may well be true, but is not independently verified and compared with similar stories. What I meant is that for every person who has problems with Lexar, there are as many who have problems with Sandisk, Transcend and other brands. The only way to determine whether a specific brand has a higher failure rate is to do some kind of scientific test.
My main intent was to help you figure out what the problem is. As I say, not many people seem to have the failure rate you report, not for any major brand.
And exactly how do you test a CF card thoroughly?
From my work in information technology, I encountered a program called DBAN, which is a free tool to securely wipe hard drives of private data before you get rid of them. I run 20 passes on new cards. You need to be careful with this program, though, it can easily wipe your main hard drive if you're not careful.
Prior to that, I had written a program that simulated storing and erasing files on the device until it was full. I also ran 20 passes of that.
If you want to do it manually, you can do a full format, copy some files until it fills the drive, then format and repeat. That would be tedious for 20 reps. I'm told there are programs out there designed to test the speed and error condition of flash memory and USB flash drives.