game lengths for beginner improvement

It seems that picking one length of game (say, 10 mins each) and sticking to it probably isn't the best way to improve. I'm still re-learning the game and struggling to properly assess the position and ramifications of my moves, so I'm trying to play some longer games and take my time. 

Would you agree that longer games can help in learning to assess position and potential moves? What would be the value of shorter games? Obviously speedy games emphasize opening line memorization, and probably have value in speeding up your assessments.

I guess what I'm trying to ask is this:

what are the pros and cons of different game lengths for beginners determined to improve? what kind of mix would be best for general beginner improvement?

I'd recommend games that are at least 30 minutes for each side. 

The game length itself doesn't really matter - there are far greater factors involved. 

Ideally, games should actually be as long as 2 hours, preferable 2 hours + 1 hour after move 40, since that's how tournament chess is played. 30 minutes is considered pretty fast for tournament play. 

Quality over quantity though. 

Also, if you can't apply something you've read in a real game (and it helps you do better, rather than making you worse), then it means the knowledge is useless or too advanced. Something to keep in mind when you see stuff in a book.

If you're trying to apply knowledge, try for a little while, because usually you get worse at first, but then it starts to sink in and help you, if it's actually useful information. 

Maybe the main value of long games is that when you think hard about a move, you're more likely to remember your thought process. So when the move is a mistake, it's more clear to you what needs correcting vs a blitz game of "I just moved fast because of the clock oh well."

Sometimes the mistake teaches you about an idea or pattern you didn't know existed. Or maybe it was an analysis error like you didn't calculate clearly, you forgot to check for something, you didn't consider multiple candidate moves, you assumed the opponent would play a certain way that helps you, etc. 

More experience in your openings in a shorter span of time. Forces you to think in patterns or groups of moves. Forces you to look mainly for forcing sequences. Although there are cons too, so certainly use speed games sparingly.

 No idea what the best mix would be. Maybe the best would be not playing blitz or bullet at all, substituting quickly playing over GM games. That way you see openings and ideas, but they're high quality, and you're not practicing mistakes or shortcuts in analysis.

When I tried going over GM games as a new player it was frustrating because I wanted to know why a move was played (or why a different move doesn't work). But that's not the point of going over them quickly (5-10 minutes per game). You just want general ideas like which side of the board (queenside, center, or kingside) and how (with pieces, or pawns) each player played. You're also getting to see quality openings. Over time recurring patterns sink in. If you want to spend a long time on a game, maybe with the help of an engine, that's fine, but don't worry about the details when playing over games quickly.

But at the end of the day it has to be enjoyable. Doing something every day is better than not doing anything. So if some days you just want to have some fun playing blitz or trying out some silly opening, sure, go for it.