This is one of those ideas that is simple in any context other than gaming: In most situations, when you describe something happening (“Diana Thunderstone punched the bruiser in the jaw…”), there are clear consequences that you can also describe (“…and he went down like a sack of bricks.”)
In gaming, we introduce uncertainty into this equation, so that if Diana’s player rolls higher than a 10, the guy goes down like a sack of bricks, but if she rolls less than 10, he doesn’t. We introduce different layers of complication to this (Maybe she only makes a little progress towards knocking him out, or maybe she takes some damage or whatever).
But if you peel back all that machinery, it comes back to the simple idea of an expected outcome. It’s the way things are supposed to go.
For all that this is a very simple idea, it has roots spread throughout gaming, many of which come back to the question of how we decide what is a reasonable expected outcome (and by extension, who gets to decide it). This is complicated further by the fact that we deal in realms of the imagination, which means finding any kind of shared agreement on what is reasonable can be tricky. How tall is a giant? How much weight can a pegasus carry? Can painting a vampire protect them from the sun? The potential array of questions is endless, and while it is possible to answer some of them, we end up having to turn to other tricks to deal with them.
The most common trick is, of course, to fold these things into a system, so the system provides the answer. Often the system (usually including some amount of randomization) fills the gap of all the things we don’t know to allow us to come to an agreed upon answer. Analyzing every single detail of skill and position makes it very cumbersome to determine if Khadgar hits the orc, so we take that uncertainty and throw it in a box labeled “To hit roll”, and now we have an answer. Or, rather, we have reduced it to a narrow range of expected outcomes, and given ourselves the tools to select between them.
But even an example that simple still depends on a broader idea of an expected outcome. There are the “obvious” assumptions – Khadgar needs to be within reach of the Troll, he has to have his sword, he needs to not be incapacitated and so on. System addresses many of these, but they are sufficiently obvious that there is rarely much need to look anything up or otherwise step out of the action. More tenuous are things situational events – what if the orc is distracted? What if it’s knocked on its ass?
System can address these, and a well constructed system (like 5e) offers tools (like advantage & disadvantage) to quickly systematize edge cases. However, this is one of those areas where GM skill and experience plays a strong role in making the right ruling and continuing to move on. The most obvious example in almost any system is the question of when to roll the dice. I mention this casually, but the decision of when to roll and when not to roll is incredibly powerful and usually falls explicitly into the GM’s hands.
Now, this idea of an expected outcome tends to hover over any situation in play. If there are greatly divergent expectations at the table, that creates a problem. If there’s great divergence between the expectations of the players/GM and the system, that can be a deal breaker (in fact, many game designs begin from this disconnect, looking to deliver an outcome closer to their expectations).
It is super interesting to look at PBTA in this regard – it has a weird relationship with expected outcomes (as it does with difficulty), and there’s a case that you might be better off going in without expected outcomes. At the other end of the spectrum, diceless games lean very heavily on expected outcomes, usually largely dependent on the GM’s sensibilities.
To my mind, this reveals something kind of critical about GM power. When you talk about giving GM power, I think people imagine very different things, often involving dropping elephants from the sky. It gets confused all the further when you have SUPER STRICT limitation on the GM except they’re still the owner of the entire fictional world – not sure what problem that solves.
But for me, when I say I want the GM to have a lot of power – more than the designer – I am speaking directly to expected outcomes. I think that system is a poor substitute for good judgement in deciding what outcomes(or range of outcomes) would most satisfy that table at that moment, and it seems silly to not acknowledge that authority. It’s not about purple beams from the sky, it’s about responsiveness. It’s about the moment of play.
Or at least that’s the expected outcome.