Back to the topic: If one is determined to stick to Myers's groupings (NF, NT, SF, ST) and reject Keirsey's (NF, NT, SJ, SP), then I suppose ISTJs and INFPs will be "opposite-ish," to borrow your term. But according to Keirsey, they have "cooperative use of tools" in common; both types tend to pay attention to rules and conventions rather than thumbing their nose at them.
Oh, and in Keirsey's last (?) book, Personology, he seems to have moved away from S/N being the main distinction; he elevates the directing/informing communication styles instead. In this slightly revised system, ISTJ is a "compliant enterpriser" (cooperative and directing), while INFP is a "compliant inquirer" (cooperative and informing). Still, they're both compliant, not adaptive (utilitarian/pragmatic).
On Myers's NF/NT/SF/ST vs. Keirsey's NF/NT/SJ/SP, my position is that it's probably a mistake to view
any two-dimension-based foursome as truly
fundamental in the way that Keirsey did, and I'm more inclined to think (as Myers did) that there are noteworthy things to be said about
all the two-preference combinations — and I think Keirsey had a lot of interesting things to say about the four combinations that he thought were the most significant.
On the more limited issue of whether some combinations might indeed,
as a matter of degree, be more consequential than others, you might be interested in
this post — which, besides talking about Keirsey's foursome vs. Myers' foursome, also includes a leetle correlational study I performed using a large official MBTI career sample, and a bonus spoiler about why I tend to think of the INs as my peeps.
As a last note, I haven't read
Personology, but I've always been skeptical about the extent to which Keirsey really believed in the "tool usage" duality that he introduced in
Please Understand Me II. That duality has always struck me as the same kind of messy, artificial category set as Berens' interactive styles (which also have their roots in
PUM II), and I can't help but cynically wonder if it was largely motivated by market positioning or academic ego or something. In any case, it's certainly worth noting that this supposed
entirely new basis (since the original
PUM) for his types made almost
no significant difference in Keirsey's type portraits.
I think the differences between the Keirsey of
PUM/
PUM II and the MBTI tend to be exaggerated in multiple respects, and just in case you're interested, you can read more of my take on that issue in
this post.