They work for me, probably because I spent the time investigating how they could be used effectively in my own workflow rather than endlessly arguing on internet forums that they don't.
They work for you like the fellow who’s smoked 5 packs of cigarettes a day, then admits that having his left lung removed has worked well for him.
The questions you should be asking yourself is what is assigning the profile doing, why is the image blocked up in the first place. Assigning a profile has a role; its to provide the numbers a scale and definition. It doesn’t alter the values at all. If the original image has an incorrect profile, or its untagged, assign profile plays an important role. Its not a color correction move as Chris (and I) correctly point out. The failure of Dan and others to understand exactly what is happening is key here! Like Dan, polishing a turd is far less useful than avoiding turds in the first place by either properly capturing or handling images correctly. You are welcome to believe in magic underwear or “false profiles” (which again, is a made up term that has nothing to do with profiles or color management, or being false).
Dan and his followers would end up spending far less work and end up with far better data if they simply strived to produce good data from the get-go. I fully understand that in some workflows, workflows from the 90’s when Dan thought up a lot of his techniques, getting lemons demanded one make lemonade. Its useful when there is no other option. No one does it better but it isn’t something we should hope for nor expect.
I would expect
you have control over the images you create but
maybe you work in a service bureau and have to polish images provided to you. If so, stick with Dan’s techniques. If not, ask yourself
why you have a turd to polish in the first place! Its like the photographer who doesn’t understand how to properly expose his film, so the ‘fix’ is push it in the process. Or someone who doesn’t understand how to correctly develop their negs, so the ‘fix’ is to leave the print in the developer a lot longer, hoping for an acceptable image. Don’t you think proper exposure and development is a better approach? I do. But if like Dan, your lifes blood is teaching people how to fix crap images, its not in your best interest to get good data in the first place. That’s why he suggest silliness like setting ACR controls all to zero, then ‘fixing’ the resulting turd in Photoshop with a lame excuse that the curves in ACR are broken (despite the fact that it was designed this way from some very smart people at Adobe). If all you know is a hammer, everything looks like a long, extended Photoshop exercise to fix the sloppiness at capture.
So why are your shadows plugged up and why do think that assigning a profile that presumably isn’t the correct descriptor of the data is useful? Or is the original embedded profile just flat out wrong? I’m not interested in arguing, I’m interested in YOU looking a bit further under the hood here. Do it, don’t do it, makes little difference to me. But it may make a difference to you.