Depth?

Hi! I'm a beginner, sub 1000 rapid, at chess. I always try to analyze my games to understands as much as possible of what I did wrong in order to improve. I'm a platinum member here on chess, thus limiting my analysis depth to 18. Of what I've understood, GMs seem to want at least depth 24 to draw conclusions from computer analyzed positions, but should this apply to a beginner, too? Or does a beginner not learn anything more from a higher depth than 18? When I'm analyzing the opening phase I tend to visit Lichess and their free cloud-analysis service giving depths of 45+ in common openings - is this for me wasted time? 

Most blunders at the sub-1000 level will be revealed by lower depth searches. Certainly , a depth of 18 is more than sufficient.

I don't know about using the higher depth searches for opening analysis.  I didn't know they were reaching depths of 45.

@notmtwain That makes sense. But if I'm looking at an opening, then isn't more depth needed? 

You're not going to be well served using an engine that's not running for hours at an opening, other than to find superficial blunders.  You're much more likely to have success looking at human-notated analyses of openings, as well as understanding general strategy ideas for specific ones.

Where can I find human-notated analyses?

Woh, guys, slow down. The assumption above about how deep a strong player looks is way too high. During a practical game, even strong players look only a few moves deep into the position. It's not going to be 9 moves deep, unless a line is very forcing or a very easy endgame. And even then it doesn't happen often. And how a strong player analyses his/her games and positions after a game is not something that you should strive to duplicate as a weaker player. 

If you're a sub 1000 player, analyzing your own games could be problematic. You don't know enough about the game yet to easily spot mistakes. This might make you rely on what a computer says too much. Some lines the computer suggests are quite impossible for a human to spot. I can discard those lines based on my own experience. A sub 1000 player does not know which lines to discard yet. Furthermore, your computer spots lines that work in a specific position. It does not care whether it's a good move in general or only right here. So you might have seen and learned a move that was very good in a specific position that you try in a new game with a slightly different position, only to see that it inexplicably fails due to that slight difference.

For a sub 1000 player I would not suggest looking over your own games on your own. If you do want to look at your own games, do it with a stronger player. It will save you time and greatly speed up on your learning. If you do want to analyse your own games, look at moves where you could have won or lost a piece or more. See if you could have prevented that. After you've done this without a computer, you could use a computer to check the whole game for tactics you might have missed. If the line is longer than 2 moves for both sides, just discard it for now. Don't worry about any positional suggestions the computer makes for now. If you want to improve, it's so much better studying tactics (not just doing puzzles, but also learning what tactics there are and how to spot them) and trying to understand some basic middlegame strategy. I would worry about analyzing my own games maybe from 1400+ OTB, but even then you could be better served by spending your chess time on a different way of studying.

I will keep in mind not to assume a move will be applicable in any situation just because it was the strongest in a given one. As of right now, I'm focusing a bit more on the opening, and isn't it then good to analyse and see what move was objectively the strongest reply - in that case, the situation is in fact going to be the same every time, right? 

The term "depth" here is a technical one related to chess engines. A human might look so many moves deep when thinking about a position, but a chess engine must look much deeper for its analysis to be valid. At lower depths, engines evaluate positions incorrectly and make huge errors. It's true that people should only trust an engine's evaluation once it has looked deep enough, somewhere around 20-40 ply.