Over the past week, I've been reading through the December playtest package. Overall, I like some of the concepts/changes. Unfortunately, I haven't had a chance to play or DM yet, but perhaps I get a chance this weekend.
I did however get a chance to convert the 4 PCs from my group (A Wizard, Fighter, Cleric and Rogue), and the 1 Rogue PC I play in ShadeRaven's group. It wasn't too difficult to retool, and the process sparked my thoughts.
At first, I thought it would be confusing to have Skill Dice, but after converting the PCs, and thinking about them more, I actually believe that having Skill Dice may make it easier for the flow of the game. Instead of tracking skills and looking at the character sheet to determine which skills have +3, +4 or +5 etc. (from additional training at even levels), players can now just remember which skill dice they use with any trained skill. I think that will make action/action resolution slightly quicker and more immersive for the players.
Picking Rogue Talent and skill tricks was fun. Of course, some seem more useful than others, but at least there are options to make the type of rogue a player wants to play. Interestingly, Sneak Attack (although it can grant 2x combat expertise dice in damage) seems less useful than other talents. If I'm in a party with 1 or 2 fighters, chances are, they'll be dealing so much damage that trying to gain advantage and then sneak attack losing advantage may be too time consuming. A rogue might be better served just standing beside the fighters and attacking normally (especially with Artful Dodger - which seems like the king of Rogue Talents to me). There are some interesting feat, trick, talent combinations though, so rogues will be very different depending on play style and player choices.
I was also wary about the way casters have to prepare 1 plus level spells, but after doing it, it seems as if that method does give a lot of freedom and choice/responsibility to make meaningful choices each game day. Utility spells and rituals need work. It seems that read magic and detect magic should be part of the Wizard class more directly than though choice of spells. Also, many of the utility spells that players would generally avoid preparing unless they really knew they would need them, should probably have a ritual version at the appropriate spell level. Also, it seemed so much better to create a Scholarly Wizard than either of the other traditions.
I like the idea that the PCs initial "to hit" scores are lower. They don't seem like super killers now, and with the boost to the monster "to hit" scores the game should be more challenging. Actually, some of the weaker monsters may have too much of a boost (giant centipede, rats, dire rats, goblins, kobolds). This is especially true because these types of monsters usually attack in larger groups (some gaining the Mob Tactics ability), which will make them much more challenging than in past packages. Additionally, after reading some of the forum posts, I am starting to believe (as do some other posters) that bounded accuracy needs to expand a bit. I'd rather see a wider distribution of "to hit" scores for monsters, and a wider distribution of AC scores as well. Some of the more fearsome monsters like dragons, devils, demons, more powerful elementals, etc. should have higher AC even if it is granted through magical or other worldly powers (especially the Arch devils and demons, and most powerful dragons). If a 10th level fighter has +7 or +8 to hit, I could definitely see some of the more powerful monsters having AC 22 or 23. There should be some monsters that the party would fear attacking (not just because the monster does a lot of damage or has nasty attack abilities, but sometimes because they are just really hard to damage).
I'm looking forward to testing this package. We'll see how all these changes change the experience.