To start off, it might be good to talk a bit about the book as a whole package. The cover is amazing, and I have to give Wayne Reynolds credit for both exceeding the original Core Rulebook art while paying homage to it with another epic battle against a dragon. This book is long for a playtest document: 428 pages! One might plausibly argue that it was too long for most players to digest during the very short window in which playtest reports were being evaluated; though, on the other hand, it's best for the designers to get as much feedback as possible about every aspect of the game. Similarly, this is not some sort of bound print-out--this is a professionally laid-out book with obvious devotion paid to attractiveness and readability. The folks at Paizo are probably the best in the business at this aspect of publishing, and the amount of time they spent on a book that was destined to become a forgotten curiosity just a few months after release is impressive (and maybe a little sad). The artwork inside is a mix of color art re-used from previous Paizo products alongside new B&W sketches of the Iconics by Wayne Reynolds.
The book is written in very precise, technical, and frankly awkward language. I know this structured approach is designed to avoid ambiguity about interpretation of rules, which is important, but it makes the book read like a computer programming textbook. I'm also not a fan of the glyphs used throughout the book--I find them confusing and unhelpful. There's also a lot less setting flavour than I would have expected considering that one of the major decisions Paizo made was to integrate their rulebook line with their Golarion campaign setting. Things are still kept pretty generic (with a few exceptions), which is a missed opportunity. As a sort of overall first impression, I wasn't excited or inspired but I what I read. There wasn't a lot of "I can't wait to try that out!" I'll also go ahead and note that my view of the playtest rules as a whole are probably coloured by my experience with the Doomsday Dawn adventure (reviewed elsewhere), which didn't go so well with my group.
CHAPTER 1: OVERVIEW
[omitted due to space restrictions]
CHAPTER 2: ANCESTRIES AND BACKGROUNDS
In the Playtest Rulebook, "ancestry" has been substituted for the more loaded term "race." The gist of the idea stays the same, as characters are selected from the usual array of human, dwarf, elf, gnome, and halfling. There's one new addition to the list, goblin, that touched off a massive debate in the Paizo forums because in PF1, goblins were almost uniformly pyromaniacal evil creatures, and their inclusion here seems more of a marketing gimmick than anything. In addition, the old races of half-elf and half-orc no longer receive their own write-up in the playtest but are instead a subset of the human ancestry.
Ancestries also work mechanically different than races did in PF1. Although there are a few fixed features of each ancestry, the vast majority of mechanics are chosen by the player from a list of "ancestry feats" unique to each ancestry. This starts the common characterisation of the playtest as "everything's a feat." The problem from my point of view is that the ancestry feats system doesn't differentiate from biological features (for example, giving an elf bonuses to Perception because they have keen hearing) versus cultural features (for example, giving an elf bonuses with elven weapons). This means, depending on player choices, some elves have keen hearing and some don't, which sort of undermines the point of having advantages and disadvantages premised on race/ancestry to begin with. It also creates clunky storytelling because characters get one ancestry feat at first level, one at 5th, one at 9th, one at 13th, and one at 17th--which means (for example) my gnome may have no idea how to use gnome weapons until suddenly that part of his ancestry inexplicably kicks in at level 13. There is a good infusion of Golarion-specific flavour in the description of the ancestries. Humans are all treated alike mechanically, but there is a one-paragraph description of different ethnicities in the setting.
Another big change in the playtest rules is the inclusion of backgrounds. Every character gets one background, and the background grants two ability boosts, a bonus skill feat, and increased proficiency in one knowledge ("Lore") skill. The backgrounds are very broad and generic, like "Street Urchin" or "Warrior." Backgrounds are replacements to the traits system of PF1, and, while I acknowledge there were way too many traits, backgrounds are very boring and forgettable. Traits could allow you to do really interesting little things to assist in a variety of character builds, and were also often grounded in the regions, faiths, or cultures of Golarion. The background system here is something fine for an introductory game for newcomers to RPGs, but lacks the flavour and depth of the old traits system.
[Languages: Omitted due to space restrictions]
CHAPTER 3: CLASSES
The classes in the playtest are exactly the same as the core classes from PF1, except for the addition of the alchemist. I thought this was unfortunate, as it would have been a perfect opportunity to show some real creativity and originality to freshen up the game. There's nothing here that made me say "wow, I never thought about being a xyz before!" As with ancestries, the classes offer some fixed elements (hit points, bumps to saves and skills) but are mostly sorting devices for a selection of abilities that players can choose. Every other level, characters get a "class feat", and these are unique to that class--so there's a big list of alchemist class feats, rogue class feats, etc. It does allow for a lot of options in build. A lot of the classes get special powers which require "spell points" to use--I'm not a fan of the terminology, though I understand having a consistent standard across the board is better than PF1's penchant for giving a point-based resource pool with a different name to seemingly every class.
I should mention that, although officially alignment is still in the game, its importance has been reduced dramatically. Instead, classes with some sort of moral restriction have a new mechanic called "anathema" which is designed to better describe the restrictions and penalties for breaching them. Given the enormous controversies in past decades over alignment, it probably is time that a better system was developed to give players more guidance. There's little if any mention of Golarion in this chapter--it's almost like the ancestries chapter was written by someone who liked the setting and the classes chapter was written by someone who didn't.
There's no way I can give a full summary of each class, but here's a quick and dirty rundown of some things:
[omitted due to space restrictions]
CHAPTER 4: SKILLS
Skills have been condensed to just seventeen--far fewer than PF1, but the consolidation looks pretty reasonable. A more significant change is that there are no longer skills points (ranks); instead, there are four levels of proficiency (untrained, trained, master, and legendary) and each level gives a very small difference in the PC's bonus when using that skill. The difference between being "untrained" and "legendary" isn't actually that big numerically, so the randomness of the dice could create some surprising results. However, some uses of some skills are "gated" behind different levels of proficiency. I prefer the gradual increases of PF1 as more intuitive and straightforward (while acknowledging there were too many types of bonuses that could elevate things to ridiculous levels).
The descriptions of the skills and their respective uses are very clear and specific, though the DCs are very discretionary--a problem that could lead to the "treadmill" effect (where GMs increase the DC to do the same thing at different levels of the game, so players never really feel their characters are getting better at anything). Other changes include the addition of critical failures (something I've long house-ruled into PF1), a much-cleaner and easier to use system for crafting, and the surprising integration of the Athletics skill into the combat portion of the game.
CHAPTER 5: FEATS
Characters gain a small number of "general feats" and a large number of "skill feats" as they level up. A lot of the feats here speed up the number of actions it takes to do something, but sometimes in ways that complicate the otherwise-simple three actions/round premise. There wasn't a lot in this chapter that was particular exciting, though I did like the level 15 "Scare to Death" feat--on a critical success with Intimidation (and a failed Fortitude save), the victim just drops dead! I wish there were more fun, flavourful things like that.
CHAPTER 6: EQUIPMENT
One of my favourite things about the playtest rules is the rarity system used for mundane equipment, magical equipment, and even spells. This gives the GM a much better system for determining how "exotic" they want their game to be, and in a way less likely to require extensive house-ruling and ad hoc judgments.
Some other changes to the way equipment is handled is the adoption of the Bulk concept from Starfinder, because apparently adding more than whole numbers is too hard for some gamers. The currency standard shifts from gold to silver, which is one of those "changes for the sake of change" that really annoy me--what does it matter? Equipment can now take a certain number of "dents" before breaking (kind of weird for shields), and there's a ton of new traits for weapons that add both depth and complexity to the system. Switching grips takes an action, and is going to lead to some really pedantic arguments in encounters. In addition, critical hits do different things depending on the weapon's type. I could get behind the changes, though I do wonder if Paizo has strayed from its goal of speeding up and simplifying the game to draw in a broader audience.
CHAPTER 7: SPELLS
Spells still work fundamentally the same way they did in PF1: there's still prepared vs. spontaneous casters, slots, spell lists for different classes, schools of magic, and spell components. Magic can still do anything and everything, though the playtest goes some way to curbing the martial-caster disparity by reducing the sheer number of spells that casters get. One of the innovations I really like is that each spell, even the non-damaging ones, has a sliding scale of success, which will lead to some more interesting and less-predictable outcomes during encounters.
[rituals & teleport: omitted due to space restrictions]
CHAPTER 8: ADVANCEMENT AND OPTIONS
This chapter includes the rules for levelling up a character, which is
very straightforward. The playtest version of archetypes are introduced here, and they're very different than PF1. There are regular archetypes (cavalier and pirate), "prestige" archetypes (Grey Maiden), and, in the biggest change, "multiclass" archetypes (Cleric, Fighter, Rogue, and Wizard). The inclusion of the multiclass archetypes means that the PF1 method of multiclassing by taking a level in a different class doesn't exist. All of the archetypes function by requiring class feats to instead be spent on archetype feats. I'm not really sure what I think, but I do like the new archetypes better than the new Background system.
There's a short but important entry in this chapter on animal companions. The "pets" system is much cleaner and better balanced in PF1. In this new system, it takes actions by the PC each round to command animals, which helps solve the action-economy advantage that pets had in PF1 (not to mention that a lot of animal companions used to be tougher than some martials). There's also a new "work together" mechanic which allows the animal companion to contribute to combat success without having to run through a full suite of actions for them. I'm a big fan of the approach here.
CHAPTER 9: PLAYING THE GAME
This is an important chapter, and there's a lot here. Some of it I've already averted to: every d20 roll you make in the game (whether skill checks, attack rolls, saving throws, etc.) has a chance for a critical success or a critical failure, and all of these rolls follow the same "levels of proficiency" system I talked about back in the Skills chapter. It's an attempt to bring added consistency to the system, and although it looks a bit weird at first, I think it's a pretty elegant way to streamline the game without losing complexity.
The most well-known change from PF1 is the switch to a "everyone gets three actions a round" concept. The PF1 distinction between different types of actions (standard, move, swift, immediate, free, miscellaneous, etc.) is gone. The advantage of the new system is clear, even if it's not quite as straightforward as it sounds (you can't just cast three spells in a round, for example, because now each component of a spell requires its own action, nor are you going to want to attack three times in a round because there's incremental penalties for trying).
The death and dying rules have been changed significantly from PF1 in a way that's so complicated I don't even want to try to summarize it. The PF1 method with negative hit points is much better in my opinion. I know, I know, some people don't like negative numbers . . . but we shouldn't have to dumb down a game to cater to the lowest common denominator.
[omitted due to space restrictions: Hero points]
A major change that I'm not a fan of is the idea that there are three fixed "modes" of play (Encounter, Exploration, and Downtime). I don't actually mind downtime being a separate thing (it's good to have a system in place for when adventurers aren't actively adventuring), but the way Encounter vs Exploration mode is articulated is very clumsy and leads to weird, forced results. This is another example of how I never heard of groups having problems with the distinction in PF1, so the "it ain't broke, don't try to fix it" precept should have been followed.
CHAPTER 10: GAMEMASTERING
[omitted due to space restrictions]
CHAPTER 11: TREASURE
I've already talked about my love of the rarity system. The playtest has another system that proved very controversial: resonance points. The idea here is that each PC has a certain number of these, and each magical item that they use (are "attuned" to) takes up a resonance point. The goal was to reduce the "walking Christmas tree" effect of characters wearing 20 magic items, and it also allows for the elimination of the "item slots" idea. I think it was more trouble than its worth, and the way Starfinder just sets a fixed limit (two, I think) on wearable magic items is much easier.
Snares get a surprising amount of attention here, which I guess is important since rangers can get a lot of special abilities in setting them. I don't mind the concept, but I'm not convinced they're very practical in most gameplay.
Instead of magical weapons and armor getting fixed bonuses (a +1 sword, for example), the playtest introduces the concept of runes. These are swappable and give the item they're attached to a special power. It's very video-gamey, and I guess reminiscent of how Starfinder handles weapon fusions.
This is a big book (428 pages), and I've unfortunately had to skip over a lot. To sum up though, I'd say as a physical product the book is really impressive for a playtest document--even a playtest document that people had to pay for (unless they wanted the free PDF). There's a real mix of things I love (the rarity system, the three-action economy) and things I loathe (Backgrounds, the Dying rules, etc.). I think the biggest surprise for me back when I started going through the materials is that PF2 was not going to be a mere update or streamlining of PF1 (like Starfinder was, though in a different genre). Instead, PF2 was going to be a whole new game, and this playtest made that very clear. Whether that's a good or bad thing, of course, is up for you to decide.