Here's a quick example of actual play in Blades, where I can discuss exactly what I expected from the player side and how I could clearly track what the GM was doing when making decisions.
The score involved ending a war with another gang, by ending the gang. We had determined that they were holed up in a well fortified location, with the entrances booby-trapped and covered by prepared firing positions. Any direct assault was out the window (we, in play, repeated botched gather info checks, so things kept stacking against our interests). However, one solid success sequence during info gathering identified a weakness -- a weekly supply shipment. However, it also showed they had supplies delivered and then brought them in themselves. This set up the score -- my character would join the shipment crew (we had this pull) and try to finagle getting into the establishment to plant a bomb, which would disable the lower floor defenses and guards and allow a rush by the other, more martially oriented characters. This led to the engagement roll, which we got a 5 on, so the initial scene was going to be Risky in position.
With this, the GM framed the scene at the door to the establishment (a bar, currently closed), with the target gang ordering us to drop off the supplies outside. They were wary and watching, so anything done here could clearly spiral (hence Risky). I made a move to settle things by saying I had brought some good Skovland whiskey to these fine Skovland patriots (the gang is ex-Skovlander rebels, mostly, for those familiar with the setting, it's Ulf Ironborn's crew). I got a success. Here's where I'm going to talk about mechanics. For this role, I was leaning on my playbook as a Slide because it means I have excellent disguise abilities, with fine tools for this -- this ups my tier level for disguise attempts. Further, I am Skovlander, so I could easily affect the accent/manner/custom here. This meant the GM was obliged to consider this, and he did. The position was still Risky, but my leveraging of my playbook and my background meant that I received a bump to effect. Had the GM ignored these things, and left Effect at Standard, then he would have been clearly ignoring the system mechanics for tier and for potency and a number of the principles of the game. This is very clear in play. The GM here is good, so great Effect was on the table, and that meant my success should move towards my goal of getting inside the establishment to plant a bomb was achieved, or greatly advanced. As the GM had not established that this task was more difficult (no clock was put forth in framing to overcome), this was a direct success. Had the GM tried to waffle here, and start a clock that this action didn't fill, then this would have been obvious in play and a violation of the established processes of play (and principles). So, we moved the action to inside the bar. The situation wasn't resolved -- I still had actions to perform to accomplish the intent (plant the bomb, get out, set off the bomb), but the initial obstacle was completed. So, now the GM frames a new obstacle inside the bar -- I'm in and without suspicion (well, without more suspicion that would be normal). He frames a new wrinkle that there's a potential innocent -- a bar maid -- still here. Whether or not she's in with Ulf's crew or a hostage is unclear, and this is an obstacle that asks my character a direct question -- do I care? It's a nifty way to be a fan of the character by asking this kind of question, and presenting that there's a wrinkle we didn't expect and that could be a problem, because she's working where I wanted to plant the bomb.
Play continued from here, but this should be sufficient to illustrate that a system with clear procedures and principles serves to deliver play that is constrained by the system in a coherent way, but still allows for the play of the game to be unpredictable -- that GM and player input can result in wildly different outcomes while still being entirely within and guided by the system of play.