After many failed approaches, I came with conclusion that the best way to practice algorithms (and math) is to trying to solve some problem which you can not solve and getting stuck, after you tired and failed to solve a problem, you can read answer how to solve a problem.
This will help your brain to remember particular method/algorithm/concept because your brain will think that this is very important problem, a sort of missing puzzle for your brain.
If you just passively read book about algorithms, you will have false feeling that you learn quickly but you will forget everything very soon and not being able to reconstruct any of algorithms you learned so far.
If you participate in algorithm contests, you will struggle a lot and after contest you will read editorials and may be redirected to CLRS or Knuth books.
Your learning curve will grow very slowly but believe me you will get very solid knowledge which won't be forgotten for a long time.
P.S. I practised a lot of passive learning by reading algorithms books. I forgot almost everything. Now, I'm switched to learning though competitions. It's hard way but very solid.
This will help your brain to remember particular method/algorithm/concept because your brain will think that this is very important problem, a sort of missing puzzle for your brain.
If you just passively read book about algorithms, you will have false feeling that you learn quickly but you will forget everything very soon and not being able to reconstruct any of algorithms you learned so far.
If you participate in algorithm contests, you will struggle a lot and after contest you will read editorials and may be redirected to CLRS or Knuth books.
Your learning curve will grow very slowly but believe me you will get very solid knowledge which won't be forgotten for a long time.
P.S. I practised a lot of passive learning by reading algorithms books. I forgot almost everything. Now, I'm switched to learning though competitions. It's hard way but very solid.