It is a sound argument that a game reviewer should have played the game he is reviewing first, and it is rare for a reviewer to have played extensively, but after one year with Pathfinder 2nd Edition I can safely say that any review I offer will be based on approximately 350 to 400 hours of actual play time, in a campaign which ran from levels 1-19 (with 20 imminent) over a 50+ story arc. In the course of play I think the following are fair observations on how balance and the unique crit/fumble range in PF2E works.
I feel like there's a subtle, hidden learning curve with Pathfinder 2nd that will make for a refined experience with the next major campaign arc. As you can probably guess, if I am planning a second campaign arc then I must like it. So here goes....my observations of Pathfinder 2nd Edition after a year of weekly play between a campaign that ran from 1-20 (with 20 about to hit), and three shorter side campaigns on Wednesday running from levels 1-5, 5-9 and a fresh level 1 campaign that is still underway (just as soon as our Wednesday group exausts our Call of Cthulhu side trek).
PF2E is an extremely balanced game. More so than any modern iteration of D20 over the last 20 years of design, PF2E has achieved something no other system has really pulled off. High level play in D&D 5E felt better than in its predecessors but could still be a slog (chiefly due to hit point bloat). PF2E manages to make high level play feel very similar (and at a similar pace) to lower level play. I simply have not worried about our level 19 game getting stuck in a multi-hour long combat like I used to in the pre PF2E days; that can only happen when you throw a signficant and deadly high level threat at the group, and even then if you design within the encounter parameters you still shouldn't see long combats that often.
The upside is this means the GM has a high level of control and understanding of how an encounter will play out. This is a strength for sure, but also a weakness; savvy players will get comfortable with encounter range expectations, and if you break those expectations it can be jarring to them. Likewise, the ranges are so tightly defined that as GM if you are throwing anything in to an encounter that is more than CR -4 out of range then you might as well just roleplay the encounter: "Okay group, seventy challenge 1 orcs attack you level 10 group, tell me how you wipe the floor with them," is a perfectly valid way to resolve that encounter in PF2E*. Conversely, if you as GM throw something more then CR +4 at them then you should do so with full warning that they might as well flee for the hills.
Some of the reason for this tight range of balance is due to the next design element....
The Crit Range Mechanic Changes Everything
In PF2E, you have a +10/-10 range of success that profoundly impacts play. If you run a couple random sessions of Pathfinder 2E without really understanding the mechanic it can feel extremely swingy at first, but in fact the opposite is actually happening: the math is extremely precise and predictable in PF2E, and it creates a unique set of what I are assumed intended consequences in the encounter design and combat experience.
When you get 10 better than your target to hit then you land a critical hit. When you roll 10 under you fumble. Saving throws often default to what is called a "Basic Save" which in most cases means spell and hazard/trap effects also define what happens when you fail or succeed by 10 or more with greater effect. In many ways its simply a codification of a mechanic which had style elements back to classic 3rd edition D20 mechanics, but now extrapolated to a near universal rule in PF2E.
One you get used to the notion, the result means that you can figure with a high degree of accuracy the odds of success and failure leading to dramatic success or dramatic failure against threats in the game. When a group that is fighting a gang of foes that are challenge rating -4, for example, they are not just four levels better in terms of skill, but four levels higher in their degree of achieving a critical success. Likewise, if you are placed against a foe which is challenge +4 (four levels better than you), then you are actually equally more likely to achieve critical failure by four degrees. This is why I feel that the balanced encounter range of -4 to +4 degrees is a bit deceptive; it's accurate, but much of the risk (or ease) of the combat is attributed to the margin by which you achieve a critical success or failure.
For example: as a player, you may realize that if you have a +10 to hit (a reasonable chance for a level 3 fighter with 16 Strength and expert training in his weapon), and your foe is AC 23, then you will on average need to roll a 13 or better to hit the enemy. You can only crit against that foe on a natural 20 since you can't roll 10 higher than the target number (AC). But if your foe is AC 15, that means you will only need to roll a 5 or better to hit, and that means you will also crit anytime you roll 15 or higher. So against an AC 23 foe you crit on a 20 only (5% chance), but on an AC 15 foe you crit on a 15-20 (30% range).
In PF2E an example of an AC 15 monster is a giant rat (challenge -1, meaning it's a weak foe against level 1 PCs, or worth level-2 for XP). By contrast a challenge 4 monster (werebear) could have AC 23. So....if you do the math, our hypothetical level 3 fighter would find the werebear to be CR+1 vs. a group of 4 level 3 characters, and the giant rat would be CR-4 against the same four adventurers. That werebear would be a dire threat one-on-one, and it would take a small army of giant rats to make an impact against the group (though they could still nickel-and-dime a solitary PC to death).
To further emphasize how critical this threat range is, the werebear has a +13 to attack while the giant rat has a +7 to it's attack. The well-armored 3rd level fighter might have full plate, giving him an AC of 21 (base 10, with +6 for the armor, and +5 for training). So our rat will hit this guy on a 13 or better, and only crit on a natural 20. Our werebear will hit him on a 7 or better, and will crit on an 18-20 (15% chance).
Those +/-10 ranges make a huge difference as the challenge level gets further from the level baseline. Furthermore, in PF2E if you roll a natural 1 you fumble anyway, and crit on a natural 20, unless you happen to have been unable to otherwise hit without rolling a crit....in which case the crit converts to a normal attack! This is logical, but it means that, to use another example, a group of typical level 3 characters like in the above example, when facing a challenge 7 opponent (so level+4, the max difficulty advised in the game) may be in dire straights. A typical challenge 7 foe is the Ogre Boss, who has an AC 25 and +19 to hit. Against our puny level 3 fighter he will hit on a 2 or better, and will dish out a critical on a 12 or better on the die (45% chance)! He normally hits for 1D10+11 piercing but against a level 3 foe almost half the time will deal 2D10+22.
At level 7 that same fighter will find the ogre boss a tough but fair fight.....but at level 3 that ogre boss will clean his clock.
Anyway, the crit/fumble range has a profound impact on how you must consider challenges in PF2E. It really does mean that you must take the rules seriously. In consideration, though, the rules scale experience with difficulty. That Ogre Boss is worth 40 XP to a group of level 7 foes, but is worth 160 XP to the group of level 3 adventurers. Under the encounter design guidelines, that means that 4 ogre bosses are an extreme encounter for a group of level 7 PCs, and 2 werebears are a severe encounter for a group of level 3 PCs. One werebear should be a sufficient "low" threat for 4 level 3 PCs however, as would one ogre boss for the level 7 group. In theory up to 4 giant rats would be a trivial threat, and by XP budget alone 16 rats could pose a extreme threat but in reality they really don't. Likewise, the giant rats would be worth exatly zero XP to the level seven group and are best described as a descriptive one off, such as "You killed a bunch of giant rats," rather than waste time killing them. The rats pose zero threat....they can't even crit against the level 7 group, and will barely have a chance for a normal hit.
One consequence of all this is that leveling up impacts the game's pace more. If you want, for example, kobolds, goblins and giant rats to be a meaningful threat for a while then don't let your PCs level quickly....they will outstrip those foes by level 4-5 and consequently a 20-session arc in the giant rat warrens will quickly become a merciless slaughter as the PCs earn the exterminator achievement!
*Using swarm rules to simulate an orc horde as a threat for high level PCs would be a way to do this, as well....