One factor you might not be thinking about is that if you use DDG first then you're getting a biased sample. You only try Google when DDG fails. So you see some times when Google succeeded where DDG failed, but you don't notice the times when DDG succeeded but Google would have failed.

As a counterpoint, I also use DDG with a !g fallback. I'd place the success of fall-backing to Google at 20% at most. (i.e. most of the time, if DDG fails, Google will fail too).

I sort of prefer Google Results too. Specially when searching in Spanish.

What I dont like is the amount of captchas Google makes me solve (perhaps because I dont use Chrome or keep my Google account open)