My thought is trending indicates either recently acquired knowledge, or short-term memory, whereas average indicates knowledge over a longer period of time. So, if your average and your trending are fairly even, that means you're scoring about the same now as you did 4 weeks ago, so it's likely that the knowledge is more in your long-term memory. This would, then, make it more likely that you're able to recall it at test-day regardless of your prep in the last couple days prior to the exam.
Trending being high enough to pass but average being below pass would mean that your trending is being boosted by things you've just learned, which may still be just in short-term memory. So, your success on exam day may depend on your last 24 hours before the exam and how much you can remember in short-term from that time. If your short-term in MCQs is based on seeing answer explanations from 20 minutes before to answer the question in front of you, then you'll be in trouble on exam day when you've had no answer explanations for 3 hours by the time you get to SIMs. However, if your short-term is from studying the night before when you're doing MCQs, then as long as you reviewed within 24 hours of the exam, you should be fine.
Note: this is theory only, and shouldn't sound like info in short-term memory is bad or useless. I swear short-term memory is the only way I passed these exams; if I'd taken them a week later, I probably would have failed, cause short-term was the only way I could keep up with the details. The overarching themes, I can still remember, but the details? Review 24 hours before was my key to passing. So, short-term memory can still pass, but a high trending score with a low average means it's new knowledge; new knowledge is more likely to be in short-term; so exam performance is likely dictated by your ability to use your short-term memory when you have a large gap between trending and average, in my working theory.