That's entirely backwards. The second option is orders of magnitude easier to crack, since it is just 7 words and a number. As the sentence even makes semantic sense, some crackers should have an even easier time guessing that password.
Meanwhile the first option consists of 15 random characters, so wordlist-based attacks don't work and one would have to default back to the much slower / 'legacy' byte-by-byte bruteforcing approach.
Maybe I didn't explain it well. And what I did was just an example. Furthermore, the world does not only speak in English.
The idea I mean is that it doesn't necessarily have to be random in human eyes, but rather random in machine eyes.
This is a little bit what I want to say:
That's true but I always wonder what kind of prediction can someone find, for example, in this password: 'railWayZDanieAccCausticCornUebung'. I'll explain: Railway is railway, ZD is Russian word, short version of здарова (Hello), Zdanie is also Russian word здание and means building, Acc is a short version of Account, Caustic is caustic, for example caustic soda, Corn is a corn and Uebung is a German word Übung that means practice.
I agree that humans are very bad at randomness but I'm curious why these combination of words doesn't sound or look random.
Maybe my English is not the best and I can't explain it in the best way.
I apologize.
My mate, I honestly suggest you go and change all of your passwords, now..

I also wouldn't trust passwords created by a password system, because if a system can create them, it can also crack them.
Either way, you can rest assured that I have my own password model, different in many aspects from those discussed here, which, as you may understand, I will not share here.
