Currently, on the tactics trainer, you get giving a rating depending on how well you do the puzzles given to you. However, you are currently being given random puzzles, which means that the rating you receive may not be as consistent as one day you may get lots of puzzles on subjects that you are good at recognising and the next day you get lots of puzzles on subjects that you are not very good at recognising. Instead, I propose that the following system could take place:

Tactics are split into groups eg Deflection, Back rank mate etc. When you begin Tactics Trainer, you are given a certain number of puzzles about each group. Once this is done, you can then have a rating for each of the separate groups. From then on, the puzzles which you do are mainly on your weaker areas, but with some of your stronger areas in as well.

Simplified Example: Person A has just started on chess.com. He goes onto Tactics Trainer for the first time, and scores the following in the 5 areas tested:

Deflection: 2/2, Back Rank Mate: 2/2, Pins: 1/2, Forks: 1/2, Skewers: 0/2

Next time he goes on tactics trainer (out of the 10 puzzles he uses), he gets 4 puzzles on Skewers, 2 on Pins, 2 on Forks, 1 on Deflection and 1 on Bank Rank Mates due to his scores from the first 10 puzzles. After doing 20 puzzles in total, Person A now has the following scores:

Deflection: 2/3, Back Rank Mate: 3/3, Pins: 3/4, Forks: 2/4, Skewers: 3/6

Person A gets re-evaluated and the next day he will have a different number of each type of puzzle due to his new weak points. This ensures that Person A will get more puzzles on his weak areas and less on his stronger areas. For example, on his third day of training, Person A will receive 3 puzzles on Forks, 3 on skewers, 2 on Deflection, 1 on Pins and 1 on Back Rank Mates.

This process can be continued until Person A becomes good at recognising all of these problems. The puzzles can then be made harder to ensure that Person A gets better at recognising more advanced tactics.

I would like to hear other people's opinions on this subject and other methods of improving the Tactics Training rating levels.

As long as I am not aware which pattern is being used beforehand, I am okay with the idea

Perhaps the way that the tactics are extraced by the Chess.com computer means that they cannot be easily classified.

In the actual chess games that determine your chess rating, you cannot expect to meet equal ratio of different tactic themes. It is up to you to adjust to what your opponent is playing against you.

Based on that I would say that the current rating system of the tactic trainer is equally fair as the overall chess elo rating system - reflecting your performance so far based on the opposition you have been facing.

I personally like the complete randomness and unpredictability of the current tactics trainer.