So I tried a second simulation of the combat - this time with carbon copies of the actual characters the players used. I handedly defeated the encounter in a little over 4 turns. Half the party (the sorcerer and rogue) were unscathed in the battle. The champion did drop at one point, but was quickly healed and returned to the fight. All-in-all, the cleric spent two 1st level spells and his focus spell, and the sorcerer spent one 1st level spell and his focus spell. No one needed to use a Hero Point.
I did this using the same basic tactics that the used. The only differences were that the rogue never stayed next to an enemy (after making his strike, he always backed away) and the cleric decided to step forward and engage the enemies earlier in the combat, which helped take some attention away from the champion.
Other than that, the enemies didn't win initiative over the entire party in this simulation, and I don't recall them landing a single critical hit. So I don't know if the dice were that "on fire" for me when I GMed last week - but is that enough to make me a Killer GM every 3 or so sessions?
And this wasn't a case of the party "barely" won - it was a resounding victory. It doesn't seem like in a system like Pathfinder 2e (or even D&D 5e) that this would be within the margins of error.