Consider the following technique for improving one's writing:
at the end of every nth day, rip up all your work for that day and destroy it
when you replace it, try to do something better, whether in overall conception, form, content, or all of these.
I haven't tried this yet, but the idea is increasingly appealing. What are the pros and cons?
This sounds like a blunt-instrument extreme variation of "Kill your darlings."
The idea behind kill your darlings is that sometimes we as writers fall too much in love with our own voices. That perfectly-turned phrase, that exquisite image, that awesome scene, that character who's too cool for the room — we don't want to give them up.
But sometimes that perfect whatever is exactly wrong for where it is. That character needs his own book and shouldn't be sucking all the oxygen out of this one. That awesome scene is slowing down the story because it's a pointless tangent from the plot. That exquisite image isn't an accurate description of the event. That perfectly-turned phrase isn't something the character would ever say. So you have to kill them: remove them from the book. (I usually keep my dead darlings in a slush file, the better to coo over their corpses when I'm feeling unmotivated.)
Your suggested technique seems to be putting "everything you've done today" in the category of a darling, and forcing you to rewrite everything from scratch so you can work on it afresh and maybe come up with some other angle on the same problem.
Could it work? Well, maybe. Depends on a lot of variables: how much you do in a day, whether it was critical progress, whether you're stuck or in a groove. I wouldn't recommend it wholesale, but you could give it a try and see if it does anything for you.
For me, this method wouldn't work.
When I write some idea, that idea is "satisfied" or "done" for me emotionally, and I find myself unable to write it again. This might be a personal shortcoming, making rewriting impossible or tortuous for me. But I also feel that the first version of an artwork often contains some raw energy, that is lacking from the more technically refined subsequent versions.
My recommendation is to not remain stuck on some imperfect work, but rather write the next book.
The main problem I perceive in inexperienced writers is that they cannot let go of their works but keep polishing or rewriting in an attempt at perfecting it. My view is that either a book is good – there is no perfect! – in the first go, or it isn't. And if it isn't, your time and energy are better invested in another book, which, in my experience and observation, will usually be better than the previous one*.
Now, you were asking about a day's worth of work, and not a whole book, but to me a similar principle applies: if I have written a page of my book, and it works, then I cannot write something equally good again (no matter how bad it may be).
Writing for me is an organic process. As an analogy, think of dancing or swimming or riding a bike. If you tried to "undo" a particular part of this large and fluent movement, going back to a previous place and repeat (or do better), what you had previously done, you'd simply trip, or drown or fall under a car.
Of course, when you learn to dance or swim or ride a bike, you do break down that huge, unmasterable mountain of skill and pracice individual movements, and if you practice writing in the same way, doing dialog exercises for a few weeks and opening sentences for a month, then, sure, you can rip up your work and do it again, but you are doing that anyway with every new exercise every day.
When, on the other hand, you are actually trying not to drown on your way from one side of the pool to the other, there is no redoing that arm stroke – either you make it, or your dad grips you.
Writing is, of course, a bit different. All writing contains redoing parts. You write the next sentence, and it may sound a bit bumpy, so you change it, until it sounds fine. Then you move on to the next sentence or paragraph or chapter. We all redo parts of our writing constantly. But that is not "ripping it up". That is part of base writing. That is like moving your arm through the water, reacting to the waves and other swimmers. That is like holding your balance on your bike when there is a bump in the road. That is like adapting your dance to the movements of your partner. When you write, you too adapt your writing to your thoughts and the restrictions of language (or vice versa), until your sentences feel balanced, in tune, and right. That is not "ripping it up". "Ripping it up" would mean that you delete the writing when it works (or after you are done). And that is not possible without falling/drowning/losing your step.
When you write something "for real", with the intention, not to practice, but to create a work of art, then writing is like being out in the open water, out in the city traffic, out on a date with a person you desire, and when you approach this kind of writing (or living) with the idea that you are just practising, then what you create cannot become a finished work of art (or love).
If you want to create art, you have to approach it with the intention to create art and the courage to risk failure. If you don't, all you'll create is an exercise.
If you find that exercise useful, you'll have to find out for yourself.
* exception: the second volume in a trilogy
This feels very opinion based (maybe the bulk of my time spent on scifi.se has conditioned me to cringe at opinion, not fact based responses :) - however, as a software developer, there is some precedent for this practice in what is called a "spike" - which is thought of as an experiment or proof of concept.
With a true 'spike' - you always throw away what you write because it wasn't intended to be the final product, but a learning experience. From the experience gained in the spike, you can then start over and construct the code properly and using best practices.
If we take this idea to the writing world, you could 'spike' on thought experiments, writing prompts, etc... with no pressure other than to gain experience and learn. I don't know if I would recommend a set of rules though that demanded you throw out work on actual projects though.
