One of the major challenges that most GM’s will run into is the fact that tabletop roleplaying gaming has both mechanical and narrative elements, and while this is intended to be a resource for storytelling the two can often trample on top of each other and lead to a gaming experience which is frustrating and annoying for players and the GM alike, requiring careful intervention to save the narrative.
If I had a dollar for every time a Shadowrun player rolled a critical glitch on their defense I’d be pretty poor, but when it happens it’s utterly disastrous, especially if they failed to dodge something that wasn’t supposed to hit them. Obviously, this is a case of mechanics conflicting with the narrative, and it is easily solved by simply ignoring the result of the roll or re-rolling it, preferably using Shadowrun’s built in Edge, which is a hedge against this sort of thing, but in some situations the GM must choose between character death or severe injury (especially bad at the beginning of a run), or ignoring the rules for a minute.
This is one of the reasons why some systems use a hybrid static/roll or entirely static method of resolving some simple conflicts; if the chance of a character actually losing is rather slim, there’s no reason to bother with the roll, especially if it doesn’t make narrative sense for them to do so. On the other hand, the simulationist attitude teaches that it’s necessary to run through the system if not every time than certainly most of the time. This is presented for three main reasons: fairness, “realism”, and “feel”, but I’m not satisfied with any of these arguments.
When I was taking an education class in college (a special needs related class), my professor told us “Fair isn’t equal; fair is what everyone needs”. Putting everything within the framework of the exact rules just to be “fair” doesn’t make much sense if one player has consistently bad dice roll outcomes and constantly dies or fails. Realism, on the other hand, purports that it’s necessary for everyone to play by the rules constantly because that way you have guaranteed accurate results. On the other hand, however, does it make sense for an 18th century sailor to have difficulty tying modern tennis shoes? Other than using an unorthodox knot, no. The “feel” of the game is the last thing that I hear people complain about. As far as that goes, I’ve done enough game design to know that some of the best made mechanics can still fail to deliver exactly the feel they were intended to, and sometimes it’s best to have GM per-case judgment on outcomes and effects.
So how can you save your narrative?
Fudging results works well, preferably when rolls are done behind a hidden screen so as not to appear to be playing favorites. This is harder in some systems than others. For instance, in the case above I tend to roll for NPC’s while my players roll their own results, meaning that I can’t lie and say “You got four hits, so the tank shell flies past harmlessly and detonates on a tree.”. Of course, this has to be used sparingly; only truly disastrous rolls should be changed or you’ll take away the appearance of danger. It helps if you think ahead a little-you could try to just negate any lethal hit preemptively, but if you think about your potential damage sources in a session you could go and negate the largest hit early on, leading to more drama in the conclusion since the player is back in the hands of probability.
Alternatively, just ignoring the results is the other simple option. This is easy and possible without too much effort, but it’s something that requires both consent of the players and a logical explanation to really work. As such, it’s more than a little difficult to pull off something that will be satisfying and work well for all the players. You can also change the results; saying “You just got killed in one shot, but I don’t want your character to die so he’s just unconscious instead.” or something along those lines. I’ve found that it works best to explicitly tell players when you do this (and make sure they’re fine with it), since not only does it seem pretty authoritarian if they don’t buy into the changes, but it also leads to some system confusion, for instance when the character wonders why the security guards with pistols did the same amount of damage as a shot from a tank.
The third possible change is that you can simply create house-rules to customize difficulty. The main difficulty with this is that you have to create them ahead of time and be consistent, so not only do you have to know what you’re going to do rather early on but you need to implement them universally. Neverwinter Nights is a video game based off of Dungeons and Dragons, and includes difficulty settings that basically allow players to ignore enemy critical hits or other such rapid death sources. Ruling that players cannot glitch on defense in Shadowrun would be a similar rule, or limiting enemy combat successes, would achieve a similar effect, though obviously this has to be considered in other games. The rules that work the best to prevent sudden unexpected results are the ones that limit the extremes of the probability curve; consider the player’s worst moments and think about how to make them better-removing critical hits on players in Neverwinter Nights means that they’ll never take an attack with a damage multiplier from most foes, rather than potentially suffering double or triple damage from a foe that wouldn’t normally be a threat but now does significant damage out of the blue.