[Lil' Lavery]

Sure high seeds typically win, but not always. The #8 seed in VCU got all the way to the finals this year (which means they obv beat the #1 seed in the QF) , before loosing the #3 seed. Last year in VCU the #7 seeded alliance (yay for us!) upset the #2 seed. The #3 would eventually win VCU last year as well. I have seen several #2 and #3 seeded alliances win competitions, almost as much as the #1 seeds. In Purdue this year, the only "favorite" not to get upset during the QF was the #1 seed, proving lower seeds do stand a chance. Even though the #1 would eventually win Purdue, it took them 4 final matches (yes, there was a tie).
Yes, there are several dominant #1 seeded alliances, such as in Colorado and Annapolis this year, but the #1 seed doesnt always get what they want. I have seen a number of teams decline picks and start their own #2, #3, or #4 seeded alliance as Alliance Captain. Last year in VCU, team 33 did just that and went on to win VCU.
The competitions which tend to have dominant alliances are usually (but not always) the smaller and/or "weaker" ones that have very few really good teams. When 2 of these teams band together, there is usually only one, or perhaps no other alliance, that can stand a chance. But at the larger and more competitive regionals, or championship, this is not the case. There are enough good robots to be able to upset a #1 seed. Look at UTC for example. the #1 lost there, and it was comprised of what many consider 2 of the best teams around.