Cleveland was exponentially less likely to win 21 in a row than the Dodgers lose 11 in a row. Simple probability theory tells us it's not even close.
It depends on how you model it, really. If you assume a "hot" team is playing .700 ball, and a cold one is play .500, it comes out roughly the same. About 1 in 2000, or once every 10-20 years worth of games.
If you model them based on their winning percentages going in, you have it backwards. 1 in 500,000 chance of the Dodgers losing 11 in a row. 1 in 250,000 for the Indians.
I think the former is more realistic, and it makes sense for it to happen every ~20 years with a few teams per year capable of it.
Interesting corollary: a .700 team is more likely to sweep a 3 game series than lose game 1.