Hey Ryan - thanks for the question, and for taking the time to read today's blog! I figure since I wrote it I'm perhaps the right guy to respond here
Part of this reply is going to depend on how we classify "rarity," so let me begin by trying to clarify what I mean when I describe something as unusual. It's a sliding scale, of course--however much I may wish it there's no
objective delineation between "rare" and "common"--but in my mind anything that appears only a handful of times per year in LR, or perhaps once every few years in LG, can be suitably thought of as infrequent and surprising.
So to your first question there are two considerations: what are the test makers doing that should qualify as rare, and why are they doing it? For the early years of the modern LSAT there weren't enough candidates for anything to count as common or uncommon (rarity means breaking from tradition, and traditions take time), so I don't count those tests much in this discussion. Instead, let's consider the past several years and the unique moments they've contained.
If you follow the blog you've perhaps caught wind of a few outlier game instances, like the Pattern games in both June 2014 and December 2015, as well as a Circular game (reportedly) in February 2014. Those two types are extremely uncommon, and reappeared without warning, much to the annoyance of test takers.
So that's the kind of outlier event I'm referring to. Ditto were we to see a Mapping game, or Profile Charting, or even another Pattern or Circular (although the surprise factor should be lessened by their recent reemergence).
Why do the test makers use those types on occasion? I think you've actually answered that: to keep people on their toes! LSAC knows that test takers these days tend to be extremely well-informed about the exam, so the LSAT becoming too consistent, or too predictable, is a constant worry. They shrug off that predictability in the only allowable way, which is to very occasionally include something that most people wouldn't expect, but that's still permissible under the constraints of test fairness.
And that idea of "fairness" is important. There's only so much fringe content they can include before the test itself becomes unrepresentative, which is why I think it's safe to say you won't see an LG section with two Patterns and a Mapping. One example of rarity? Fine. Fair. Two or three? That test is suspect. So I'm pleased to report that there is in fact a ceiling for strange.
Which leads me to your second question, and it's a good one as well. Obviously an unexpected game causes quite a fuss, in part because that section tends to be the most challenging for people, and in part because one game in four stands out so much more than one LR question in 25 (or 50). So oddball games get a lot of attention, and deservedly so.
But that's not to suggest that LR doesn't permit surprising variability as well, both in terms of what people should rightly anticipate, and in terms of some misconceptions most people hold. Take a question type like Evaluate the Argument (and if you don't know what that is it doesn't matter here; you'll learn of it soon enough if you're in a class with us): Evaluate questions are the single least common type on the test historically, often with years passing between appearances. And yet in the past few years nearly every LSAT has had at least one, and sometimes even two. So that recent emphasis caught a lot of people off guard, myself included if I'm being honest. I know how to solve Evaluate questions so it was of no personal consequence, but considering how little attention the average person out there pays to that q-type I know multiple Evaluate problems was a legitimate issue.
The same could be said with a heightened emphasis on Cannot be True, or Formal Logic, or the re-inclusion of "double question stimuli" (as you mention; interestingly they used to present that question construct all the time, then phased it out entirely, but lately we've been hearing steady reports that experimental sections are featuring it again, signaling its probable appearance on an exam in the near future. See Dave Killoran's discussion of this for more info:
The Rise and Fall of Double-question LSAT Logical Reasoning Stimuli ). Likewise, a section of LR that has an unusual composition in terms of question frequency, with, say, 6 Justify and only 1 Flaw, or 7 Weaken but no Must be True. Those are extreme cases (and hypothetical; neither has ever happened I don't think), but they'd fit the definition of rarity.
Secondly, and as I alluded to above with LR, there's the issue of misconception. For instance, did you know that Strengthen questions appear, on average, twice as often as Weaken? Most people don't, and thus devote as much, if not more, time to studying Weaken than Strengthen, despite Strengthen being literally twice as important points-wise. So even with an entirely typical LSAT people often are caught off guard due to their own misunderstandings about what to expect.
Finally, you'll note that I didn't mention RC in my blog. The reason why is that that sections seems to be by far the most consistent of all: you'll always have four passages, only one will be comparative reading, science/law/humanities are going to feature, etc. To change any of that the test makers would have to announce it well ahead of time, as they did with the introduction of comparative reading in June 2007. Failing that announcement you can rest easy on RC.
My point in all of this is that a great test taker is going to be ready for anything, while still setting expectations according to what's most likely. Know all the conceptual types, but also know how to prioritize your time and energy so that the most common elements get the most attention.
I hope that helps!
Jon
