I always use Blitz to try out new openings, for fun short games when I'm bored of slow games and often to fish for tactical positions I typically screw up (that I add to my personal database for review + practice later on).
Though I will never trust blitz as an improvement tool.
You are quite right to think that rapid + repeated exposure should help but it actually doesn't.
A simple analogy from personal experience:
Case 1:
As a kid, I could play 10 piano pieces during practice session in 20 mins and if I make mistakes, no big deal, I'll just re-play the bar that I messed up and keep going. If the mistake was small enough, I'd not go back and just keep playing the rest of the piece. After 20 mins are up, I think I practiced great and pat myself on the back.
The good: Covered a lot of ground. More material. Ready to learn new stuff. Fast. Fun. What's not to like?
The bad: Uhhh.. you know those mistakes I used to make at key moments in each piece? Well by rapidly practicing over them .. I'm actually re-inforcing bad technique and the only improvement I'm showing (over the long run) is that I'm making the same mistakes BETTER and BETTER!
Case 2:
I play one piece for 20 mins. Each mistake makes me slam the brakes. I start over and play it SLOWLY and keep trying until I get it right. Then move onto the next section. I strive for an insane level of perfection until I make no mistakes discernible for my level. You might burn the entire 20 mins and not even finish 20% of the piece.
The bad: What the hell is this? This is boring / less fun. I didn't even finish one piece. I feel that I accomplished nothing.
The Good: This is a variant of what is called "Deliberate Or Mindful Practice" where you force yourself to be an OCD monster who feels pain and suffering with each mistake you make. By analyzing every aspect of the challenging part of the music you are playing SLOWLY and PATIENTLY (in chess, this would be a position you are analyzing) and attempting to solve it ... you are creating a beautiful "instantaneous feedback loop" where if you don't like the result, you will use go back instantly and re-try to correct yourself.
As torturous as it is to some people, this "slow burns" the patterns over the board into your brain permanently as well as creating a sense of mindfulness where you become patient enough to not miss anything ... perhaps this is the true muscle memory that skilled atheletes/musicians call upon to perform at the highest level.
To summarize: by forcing your brain to patiently/slowly grind into positions and observe every nuance, you are really transitioning the patterns to long-term memory where they can be recalled faster and more accurately.
Ever wondered why it is easier to remember facets of a slow game you played last week than one out of 10 fast blitz games you played a hour ago?
For the same amount of time, I'd rather have 5 patterns burn into long-term memory than have 50 patterns make a fleeting stay in my short-term memory.
I've been a Case #1 person a while back (quantity over quality as they say) and a few years ago, I've started transitioning into deliberate/mindful practice. It has already helped me grow a bit in strength as well as help me become a more patient player!
I am a very low-level player (1200) and would like to improve at least to the 1500 club level. I'm spending most of my time on the tactics trainer and really working to understand each puzzle that I miss. The rest of my chess time is split between the chess mentor program and online games with virtually no time limit at three days per move. In general, it feels like my chess time is strongly focused on slow and thoughtful chess.
There seems to be no end to the posts in forums advocating the "quality over quantity" approach to learning that I've taken, but I don't think it's that simple. At my level, what I think I need most is tactics and lowering my rate of stupid mistakes, and a lot of that comes down to pattern recognition. So the question of quality vs. quantity seems rather spurious. What I want to know is, how to optimize my progression in pattern recognition? Let's consider the possibilities of a little blitz chess, though I have yet to try it.
It seems that playing blitz chess would increase the number of times I am exposed to the patterns I am working to recognize, and if repeated exposure increases recognition then it stands to reason that blitz chess could be a useful learning tool. I'm guessing that the key to actually learning from it would be analyzing my blitz games for blunders and missed opportunities so that they don't get swept under my mental carpet.
Any thoughts on the matter?