im going to try to hit this one square-on, so please take everything i say here in the helpful spirit in which it is intended!
first, to get it out of the way - unless you find it very helpful, don't try to think about initiating the fastbreak, or how fastbreak would really work, or anything like that. i have found that fastbreak really does not line up with what we would intuitively expect it to be. it a very good set, and i like it a lot, but at best is it very loosely related to any real-world conception of fastbreak you are likely to have.
fastbreak is a lot like all the other offenses. think of it like them as your baseline, except with a few little tweaks:
1) you play at a higher tempo, so players on both teams, but especially yours, pay higher fatigue. there are more possessions, which fatigues both teams, but you also burn more fatigue per possession running FB than any other set. im pretty sure it works that way, i wouldn't swear to it, but overall with fastbreak, i am pretty well informed relative to my understanding of HD in general. i really like fastbreak personally. it and motion are my favorites. a useful device is to think of FB as a mini-press, like, half a press, when it comes to the increase in possessions and fatigue
2) you can't run slowdown. this sucks, making fastbreak much more desirable for teams who hope to make deep NT runs as the favorite (as opposed to the many folks who are hoping to make a deep NT run but by pulling a couple upsets).
3) fastbreak gives up more turnovers - about 1 more per game (at a game at an average amount of tempo) than any other set.
4) fastbreak scoring formulas are materially different than the rest of the game. ath/spd weigh more, and lp/per less - especially on 2 point scoring. fb is a great set for 2 and 3 point scoring both, but i think it is hands-down the game's best set for 2 point scoring (but still, by a small margin - 3s are still more efficient in fb, just like all other sets, in 95%+ of circumstances). you can be really efficient, both in terms of 2pt% and fouls drawn/FTA, and you can do it with less of the scoring stats (lp/per) which are so highly valued by others. the nice consequence of this is you can basically just get a couple elite 3 point scorers, as classically defined (per being the most important stat), who will take on heavy scoring workload, and you can pretty much fill in the rest with whatever - if you are prioritizing stuff other than lp/per, you are probably getting strong defenders and rebounders and guard skills guards, which will usually have good ath/spd. those high ath/spd players without great lp/per will be decently efficient for you, solid mid-level scorers, and certainly better than they'd be in other sets. FT is more important for 2 point scorers in FB relative to other sets, as a logical extension of all of this.
otherwise, you basically want to build your team much like you would any other team - the fundamentals of it are the same. however, many folks make a really big mistake and trying to build with this really rating-focused understanding of what makes players and teams tick, when you really want an ability-oriented outlook. i am afraid you are making this mistake, with the way you ask the question. it should be about abilities, not ratings. when you think of what makes a great team, you should think things like, a pg who is strong defensively, with really good guard skills to limit turnovers and to increase the efficiency of team mates' scoring. of course, on an individual basis, you need to be able to know things like, what ratings actually contribute to this ability, like for example how pass and iq are the only ratings (and stamina, if you want to think of it like that) that impact the 'increase the efficiency of team mates' scoring' ability.
its hard to answer your question because you don't talk about what abilities you are wanting to prioritize. and really, the SF is the swing position. like any other scheme, in fb, your SF is your opportunity to make your team whole. what role does your SF need to fill? are you strong def and guard skills (not including scoring) wise at the 1-2, but weak offensively? then you need all the per scoring you can get at the 3, no matter what the cost. have a dearth of rebounding, but solid everywhere else? you need a strong rebounder - and so on. so what im kind of getting at is, the way you ask what attributes to prioritize, with the sole context of a fb sf, i think is shows you are thinking of these things in too trivial of a manner. so, please allow me a long digression on the subject!
nobody should have a static idea of what they are looking for at a position - what you want is a dynamic idea - where you contemplate multiple roles you could have players fill at each position, and try to understand what ratings allow the player to achieve those roles. but really, outside of standing how ratings interact to create abilities like per scoring, per defense, post scoring, rebounding, guard skills (which you could call a product mainly of limiting turnovers and of increase the efficiency of the scoring of the rest of the team) - you really want to be thinking abilities. you want to think - i need at least two strong per scorers on my team - not, i want two guys with 90 per.
one last note on this subject. with 12 players and about 12 ratings each and with countless interactions between those ratings (counting off iq, ft, etc), its a really large number of factors, trying to understand what makes a team good - which is the ultimate goal, building-your-understanding wise. understanding what makes a player good, and really, what makes a player good at a specific function, those are just building blocks you assemble to try to build teams that work well together.
anyway, if a player has a few high level abilities (like off, def, reb, guard skills - those are the big four - you can subdivide, but those 4 get you pretty far - with caveat that you should recognize per scoring is more efficient than other scoring in the vast majority of cases, and by a significant amount, so you need a lot of good 3 point scoring, generic offense isn't enough) - if they just have a few abilities, its way easier to think of how 5 players might combine to form a full set of players. i want 2 who are really good rebounders, 2 who are really good scorers, 1-2 who are supporting scorers, 2 with good guard skills, and 4-5 with really good defense - something along those lines. that is concise, that can be understood, tested, and refined. it is way easier to build your analysis thinking of players and teams in ability terms, than with all the complexity and numbers involved in talking about all the ratings. abilities just cut through to what matters. i don't care WHICH of the 50 combinations of ratings make my player an elite 3 point scorer, all i care is that he IS an elite 3 point scorer.
team building is the single most important part of the game - and it is good you are working towards it, which is clear from your question. but i encourage you to try to think about abilities, not ratings, because it puts you on a path to a deeper understanding, by bringing you closer to what matters, and by abstracting away from all the numbers and gobledegook that doesn't.
... so, to your actual question. it totally depends on what you need for your team. if you need a strong per scorer, then definitely the first. if your guards are on the whole really good, you may get almost no marginal value from having 1 more guy who can shoot the J, so B would be far better.
and really, nothing about that question is all that unique to FB. by far, the most important factors are the team building principles, what makes a player good at something, what roles do you need to fill on your team, etc. if you were to ask, how is the question different for FB than any other offense? i suppose i would say, stamina obviously matters a little more across the board, so there's that. rebounding is no more or less important in fb, although fb is more forgiving to some styles of play that are variations of 'small ball' and include less rebounding - but its not like its more important in general. there are a tad more turnovers, but the SF is a small part of that and i think the impact on bh/pass's value for a SF, compared to any other scheme, is really quite small (but not 0). the bigs ones are, if this SF is going to score 2s, ath/spd are more important than other schemes, lp/per less, but for 3 point scoring you really want that per like anything else. the relative value of spd to bh for fb 3 point scoring is really not much different than anything else - and both are quite important.