[Chris is me]

I'm not really a fan of the word "irrational" to describe picking behavior. There is too much subjectivity in picking and information for all participants for this to be the term I'd prefer. Most every alliance selection pick is maximizing the value of their pick as best they can, which is by definition rational. Many examples brought up in this thread are easily defendable picks (like 973 picking the strongest scorer in their division, or a team declining for a better chance at a wildcard).

The truly terrible picks are from teams that are unprepared, don't care enough, or don't have the ability to get the data they need to make a good pick. Usually a combination of factors here, as even teams without formal data can make decent picks sometimes, and even teams who don't care enough might pick a first pick with good information to work with.

The odd, suboptimal but workable picks happen for a variety of reasons. Different seeds have completely different incentives for their alliance picks. The number 1 seed at regionals / districts generally wants to eliminate variance as much as possible, picking a consistently strong first pick and a consistently functional second pick. The 2-4 seeds want teams that may be less consistent but have a high enough ceiling that with some luck they can outscore the top alliance. The bottom half of the draft either wants to go for a strategy where 3 decent scorers beats 2 great scorers, or they want to pick super high variance teams with high point ceilings and hope for some luck and preparation to help them through the eliminations.

I'll add more thoughts later if I get the time.