Not risks versus rewards, because the two are not polar opposites locked in a tug-of-war. They exist semi-independently, hopelessly entangled but able to be measured independently.
Reading through this post on risks gave me mixed feelings. I was happy that other people are doing similar things to what I laid out in The Law... I was once again upset that numerous people showered the author with adoration and praise for revolutkonary thinking, when I've been stomping around in this territory for the better part of a decade.
Different actions are flavoured by varying outcomes and varying risks. This is an inherent part of "The Law". The actual rolling of dice remains consistent, but it's the potential positives and potential negatives that lock the die roll into the unfolding narrative.
A good action result might give a benefit without a penalty, a step in the right direction where the risk has not manifested. A lesser result could go one of two ways... it might allow success at a price, or if might indicate failure it that price is not met. An even worse result might require the price to be paid as the risk manifests itself, but the action still sees no benefit to the characters.
The Apocalypse World engine kind of does this, in a crude way, ut keeps everything on a linear track.
10+ = Success/no-Fallout ➡️ 7-9 = Success/Fallout or no-Success/no-Fallout ➡️ 6 or less = no-success/Fallout
It feels like Vincent Baker took his "Otherkind Dice" system and decided it was too innovative for the masses, so he'd dumb it down a bit. The only similar analog I can think of are the dark years of D&D 4th edition, when the dual axis system of alignment was reduced to a linear scale. Instead of offering law versus chaos searate to good versus evil, it simply followed a track of "Lawful-Good➡️Good➡️Neutral➡️Evil➡️Chaotic-Evil"... it left out the fun diversity of Lawful-Evil characters and those who were Chaotic-Good...as well as those who were straight Lawful or Chaotic. A chunk of the potential richness gone.
If we tie those concepts back to action results, perhaps saying that the "Good/Evil" axis is analogous to success/failure, while the "Law/Chaos" axis is analogous to "Fallout/No-Fallout", we can start to see one of the issues I have with the AW engine.
10+: you're Lawful-Good ➡️ 7-9: you're either Lawful-Evil or Chaotic-Good (you decide) ➡️ 6 or less: you're Chaotic-Evil.
It still leaves out all those interesting Neutral cases, but if I were going to try and add them back in, I can only think of needlessly complicated ways to do it. The system just isn't built for that level of nuance. Yes, I can hear the voices of a thousand Apocalypse-Afficionados and Bakerites all screaming "But, mah system! He's picking on it... PbtA can do anything". No, no it can't. I appreciate black-and-white photography, but don't expect it to capture the vibrant colours of a flower. Different game engines do different things well.
In the post I referred to above, Rob still takes this mixed binaries approach. He divides the potential risks into a few areas, and makes them mix and match to adapt to a variety of situations (which I think is good), but he still links a pair of binary yes/no results into a linear progression.
As a tool for flavouring actions, it's great. I also understand that his examples are for injecting into an existing game, in this case FAE, rather than generating a new game from scratch; so the application of these risks has to be applied in a way that won't disrupt the existing systems too much.
The whole thing maps fairly similarly to the way I run my sessions of "The Law", but by breaking the success result and the risk/sacrifice result, I was able to be more nuanced by showing varying degrees of success balancing off against varying degrees of sacrifice... rather than just binaries on a simple linear path. I went with the idea that a player chooses the obvious risk associated with an action, and then that risk might manifest into a sacrifice with a moderately bad die roll on ond of the dice, or might have additional fallout chosen by the GM if the roll is abysmal. All of these elements of fallout are derived from the story and feed back into the story, separate from the successes. I'd considered this like driving a car, where the rate of success is like the car's speed, while the sacrifices and risks determine it's direction. This works well for a non-railroaded game because it allows a full range of movement throughout the game environment. Everyone wants to get somewhere, and regardless of how fast you go, you might not get there because something might divert you the wrong way. (Meanwhile someone sacrificing speed for precision might achieve the end goal sooner, just because they were careful).