RIFTS uses two options for most of its rolls. One of them is to roll 1d20, add all appropriate modifiers, sometimes compare against another roll or a static number, and then declare success or failure- always roll high. The other is to roll d%, do as with the 1d20 roll, but always roll low.
Let's work with this. You usually do the d20 High roll for combat/saves, and the d% Low for skills. That indicates a trend; when there is a known or bounded range of probability, you use the d% because Roll Low works for when there's a known boundary of probability and when there isn't you go with the Role High option. (Yes, you could convert the game to go entirely one way or the other, but if that's what you wanted you'd be using a d20 System or Basic Roleplay product and not Palladium.) Look, I'm not going to get into the math of that--experts in statistics are welcome to throw in their pair of pennies below in the Comments--but I think you get where I'm going with this.
We have Saving Throws and Horror Factor checks on the d20 Role High schedule. While the Target Numbers are static, what makes them unknown is what the character subject to them can add to the roll, and by putting them on the d20 schedule it also means that automatic success and failure no matter the odds is on the table at all times. We like these features when we want either that possibility of success or failure no matter what, or we have one side being partially or wholly unknown in their probability.
We have skill checks on the d% Roll Low schedule. Here we have a character operating within a specific realm of expertise, with a known degree of proficiency, and the checks are really checks to see if that character can properly apply their expertise to the situation before them. As is so common, this general principle is not applied uniformly, but it is still very much the case. We want to go this route whenever similar conditions arise.
So, here's the suggested House Rule: if your man can't act knowing all the variables at hand, then roll the d20 and go high and let the GM figure out if he passes or fails; if your man does, then have the GM declare the odds as a percentage, roll d% and go low.