This post will be something relatively close to liveblogging my reactions to reading through the D&D Next (I still think of it as 5th edition) material. Note that I haven't played or run any of it yet, and won't until Monday. Also, I can point to things here that contradict very recently posted material. I will not be spoiling any Caves of Chaos material here, since I asked my players not to read it.
Specifically, they've stated that they're changing the Hit Dice for wizards to d6s and rogues to d8s, but that isn't reflected in these documents. Also, they've listed the fighter's hit die here as d12, but that post specifically mentioned that they would have d10s. As the article suggests, they've also made the Hit Die itself carry a little more systemic weight - in this case, it is also a close cognate of 4e's healing surge mechanics. Your Hit Dice plus (Con modifier x level) is effectively your limit in nonmagical healing per day.
The long rest mechanics are explicitly written to require either 1) a 4+ person party or 2) some span of the rest that has no one on watch. Hey, look, it's the first rule I absolutely must remove: if you take any strenuous action during the long rest, you must start the rest over. So basically, if the party is attacked during the long rest, you gain no benefit. I get that this is meant to encourage the party to return to town, but it's a huge impediment to wilderness adventuring. On the other hand, the long rest is a 100% restore (if you had at least 1 hp). If you didn't have at least 1 hp and don't have a magical source of healing to get you to 1 hp, you're stuck in the Stabilized condition for 2d6 hours. Guys, there has got to be some better way to do this - I'd like to see a Long Rest grant less than a full heal, while also restoring all spells. Natural healing does need to be more robust than in 3.x, but less robust than 4e. I want my low-level parties to need time to be up and ready to fight again.
Humans have no racial ability listed. This makes me cranky, especially in light of the dwarven immunity to poison (oh, how I hate it when players have immunities!) and the much-less-surprising elven immunity to sleep and charm effects. Races don't really have visible drawbacks, though it's hard to tell based on what might be numerical effects that are already factored in. One of my fellow commenters points out that they may be better off during ability score generation, but it's hard to tell - there are no character creation rules here.
Not sure what's up with the halfling rogue using one die size larger for dagger and sling attacks. But that pales in comparison to my objection to +1d6 sneak attack damage every level. Sneak attack scaling was too steep in 3.x already! In combination with the halfling's Naturally Stealthy racial feature and the Lurker theme's Ambusher feature, it looks like the halfling should be able to gain advantage kind of all the time, as long as there's someone larger in the party. At least hiding takes up your turn.
Interestingly, they've made the melee cleric the "tank" of the party by
giving it the Guardian theme (grant disadvantage to enemies attacking
targets within 5 feet of you). I had been under the mistaken impression that characters had three major choice points in creation (race, class, and theme). I expected theme to be relatively minor in terms of how much power it gave your character, so I was surprised when they revealed that "being a defender" was something you got from your theme. I felt that this was at odds with it being a minor aspect of the character, and further felt that it would invalidate a huge range of theme options (since so many parties will want the fighter to be a defender). I would have gotten all of this beforehand if I had read this article, but it somehow evaded me. It turns out that characters have four major choice points: race, class, background, and theme. Background now does what I had expected theme to do, and theme carries the more significant mechanical weight. Backgrounds are still pretty nice for granting skill bonuses, but you don't have to choose between, say, blacksmith and guardian the way I was fearing.
I'll have to improvise a stunt system into existence to give fighter-types more rules support. I liked the powers that 4e gave fighters to help them characterize their fighting style in evocative and rules-meaningful ways. The options here are like someone took 2e and bolted on just the tiniest bit of 4e (specifically, the Slayer automatically deals a small amount of damage even on a miss).
Giving the wizard the "magic-user" theme just seems silly to me, but its effect is perfectly reasonable and useful. But thinking about the statement that my wizard is specialized in magic kind of hurts my brain with its redundancy - this would be less of an issue if more of the visible themes seemed feasible for the wizard in the first place. I'm looking forward to writing a wide variety of themes and backgrounds to cover things that fit into my conception of a setting.
The Guardian's defense mechanic is super stripped down, as it simply inflicts disadvantage. The lack of "punishment" as part of the mechanic is odd to me as a long-time 4e player, but I'll reserve judgment until I've seen it in play. The Guardian (which is, again, a cleric - a pretty tolerable paladin, if you're willing to squint) gets a stickiness mechanic at 3rd level. The only problem I have with this is that I got used to thinking of it as a core role function, but it's probably fine.
The melee cleric has the death ward orison, which is completely unlike 3.x's super-powerful death ward effect. This is something the cleric can use as his action every turn, but probably won't - still, giving the cleric a constant ability to counter necrotic energies is interesting. "Resistance" now means "half damage from that source," end of story - this disappoints me because it sort of fails at granularity, for my tastes.
The armor chart is incredibly weird to me. Armor is divided into Light, Medium, and Heavy categories. There's no Armor Check Penalty, but there's a note that you gain disadvantage when making stealth checks in medium or heavy armor. Light armors allow you to add your full Dex modifier to AC, medium armors allow you to add half your Dex modifier to AC, and heavy armors allow you to add nothing to your AC. For basically any arrangement of Dex scores, medium armor is the worst of all options. The more magic-focused cleric wears scale mail and has 15 Dex - which gives him the same AC as if he had spent 25 gp fewer and bought studded leather. I get that the designers didn't build in many slider bars for armor, but medium armor will need fixing - the only time you should be using it is if you have a very poor Dex score and you can't wear heavy armor. I am glad to see that your character class doesn't exactly dictate what kind of armor you'll wear, though.
Since these rules aren't written on an assumption of a gridded map with minis, a much wider variety of spell areas are back in action. Hokay. I've been running non-gridded combat in Mage and OtE lately, but I'm likely to adopt miniatures rules for my D&D games unless there is a really compelling reason not to.
So, um. This whole advantage/disadvantage thing. It's not 4e's combat advantage, which I kind of expected; it's "roll twice and take best" (advantage) or "roll twice and take worst" (disadvantage). That's cool and all, but it means that shield of faith makes the target immune to sneak attacks for its duration, since it causes attackers to have disadvantage, and advantage and disadvantage cancel each other out. This is kind of a big deal. It is awfully fast to play, but I have major concerns about the consequences over the course of play.
The different language in the cleric and wizard character sheets for spell preparation is somewhat perplexing. It looks like clerics cast spontaneously from a fixed list (based on choices presumably made at creation and level-up), while wizards are pretty much completely using Vancian magic. Both classes do these things as add-ons to their at-will effects. This wizard is showing a lot more variety in his at-wills than the clerics are, but that comes from his theme.
Now for a summary of the work as a whole. While I've made a number of negative comments above, I like the general simplicity of what I see here, and I feel comfortable making plans to hack it - it looks much more kitbash-friendly than 4e, though the early-release materials of 4e also looked much more kitbash-friendly than the final publication. The character sheets are much simpler than in 3.x or 4e, so maybe my players won't feel as much of a need for DDI character-management tools. I'm looking forward to playtesting this material, and I can make guesses on how to backfill enough character creation material that I could try offering more options to players once we've exhausted the playtest material.
My favorite change in this edition (at the moment, anyway) is introducing both Background and Theme. Classes can carry some skills, but the majority of the character's skills and a minor feat come from background. Theme carries something significant and mechanical that could usefully apply to any number of different classes. Background feels a lot like a class for non-combat situations - something I'd been thinking of adding anyway.
My least favorite change, all in all, is returning the flavor of encounters to something more like 2e. I felt that 4e's incorporation of terrain and obstacles into encounter design was an inspired move that made huge strides in just making a set-piece encounter interesting. These rules present a more retro feel, quite self-consciously. I'll need to focus on encouraging players to do cool, creative things, because even the best of us can fall into a rut when our combat turn comes around. Since interrupt/reaction actions are almost completely gone (this is both good and bad) and there's much less mechanical focus on setting yourself or your allies up for the big, devastating attacks, it's a lot easier for a player to tune out (and start up a side conversation) whenever it isn't his turn. These kinds of things drove a particular playstyle in 4e - it wasn't for everyone, but it went over very well with my PCs.
What I wanted this to be - what future rules modules may turn it into - was a more rules-flexible 4e. What I got was not that, but something that is still playable and potentially very fun. I dearly wish they had gone for a greater degree of rules transparency, and there are definitely areas, like initiative, where I wish they had taken a few more risks in innovation. Still, I'm comfortable making that hack myself after playtesting this weekend (ahem, in a way that doesn't violate my agreement with Wizards - so we're talking post-launch here), and with highly modular rules, it's all but guaranteed that they'll eventually publish it themselves.
ETA: I thought about it just a little more, and decided to add: the actual best thing about this text is that the D&D it presents suits my aesthetics of fantasy quite a lot better than the later portions of 4e's lifecycle.