Numerical value assigned to each move, accurate?

As I new player, I found that I seem to evaluate the strength of each step is just .... very different from how the computer calculates it. 

A mediocre move seem to get rated as somewhat good. And whenever I make a move that I consider good, it just gives me like 15 - 20 pts or so. Whenever I captures something, it doesn't give it too much value either. It's strange because I get 100+ points for just strategic moves that may or may not work in the long run. I also don't understand why capturing a queen/rook only gives me like 20 pts or so. 

I think what bugs me that most if whenever I make a step consider as good the computer will think it is mediocre. The way I see how good a move is just completely different. I'm getting crazy because of this. Can someone explain to me that I'm just a bad player and the computer is right? lol 

Thanks.

You kind of answered your own question - you are a newer player. Computer evaluations take many positional and tactical aspects into account, most of which you are likely not familiar with. And, computers look at games from the highest ELO ratings, so again it is understandable why your move evaluations differ greatly from the computer.

Just stick with it, read a few books for beginners, do some of the lessons on Chess.com As you become more familiar with opening theory and positional play, perhaps the computer evaluations will start to make more sense. It is a process. Good luck with your game.

@zoechau Have you ever seen those horror movies where people in the theater shout "don't open the door!" and yet the character opens it and gets ...? People in the theater do that because they are aware of the line in the story. The computer may give a low value to your moves because you may be "opening the door" without you being aware of the consequences.

I have absolutely no idea what "points" the opening poster is talking about. This makes no sense. Maybe I am missing something crucial?

He's probably talking about the evaluation numbers from an engine analysis, and believing they're supposed to represent the value of each move.

An evaluation can be wrong or not but is not an opinion.  On the other hand, not having a perfect chess algorithm does not exclude correct evaluations, like the K+Q v Q result.

Yup, I just noticed from another post.

Is there any engine evaluation in the world that goes up by 15-20 points every move, and even by 100 points occasionally? I don't think so.

So you want to be picky when a beginner describes 1.00 and 1.20 as one hundred and one hundred twenty.

You may be right. Still, what I am saying is the same. Is there any engine evaluation in the world that goes up every move? I don't think so.

Pay attention to the big swings i.e. where you miss tactics that win a good amount of material. The rest of the time you can ignore it. The first reason is because you're somewhat new, so the "proper" move may not even be good for you (because it requires a lot of technique and understanding).

The second reason is because I assume you're using chess.com's auto analysis thing that runs in the browser and doesn't think very long. This kind of analysis is more of a marketing gimmick than anything useful. It can catch big tactical mistakes, but for other moves don't put much faith in it.