They took away a ton of Warrior's best card draw/cycle tools (and all of their bad cycle tools too). Control Warrior's consistency with getting cards like Brawl was a big part of the problem with the card. Its overall powerlevel has remained the same, but we can probably expect Control Warrior to have it in their hand less often, making it a more reasonable inclusion.
It's definitely not too random to see play. If Peon read "Battlecry: Add a random spell from your class to your hand" it would be very playable in value-oriented decks. The fact that he has to survive damage to get the spell is the real challenge here.
Shurima is really strong, no doubt. It has the tools to compete. I think however that the new Shurima cards got intentionally released stronger to make shurima -which has only half the cards of any other region- playable. If they nerf the overtuned 1 drops later on and push back when the number of shurima cards gets more I am fine with that approach because unlike targon, shurima is not a toxic control region.
This is probably exactly right, and exactly why they were dumb to set up their new regions/expansions like this. Without a full set of cards up front, they're forced to overcompensate. (I also don't really get the whole "some regions get expansions first" thing, but oh well.)
Spider Tank's mech tag may synergize with certain archetypes in a way that makes it relevant despite being (mostly) outclassed by updated class cards like Warhorse Trainer and Void Terror.
There seem to be two major questions here (and a brief, but seemingly irrelevant discussion of talent trees): should weaknesses/resistances to spell schools be a feature of Hearthstone, and should those weaknesses/resistances be implicit based on typing.
As far as the second question is concerned, I think this answer is plainly "no." Hearthstone is a relatively user-friendly game, and the lack of discoverability associated with those connections would be very hard to deal with. First, you'd need tons of helptext on every spell to explain what is or is not weak to it, then you'd need some tooltips to show you which enemy minions are of a given type, and all of that would have to be overlaid onto the game interface. It would be a nightmare.
The first question is a bit more interesting, but I think it's also a bad idea outside of anything but a very small number of Legendary minions. If it isn't very narrowly scoped in its application, you could easily get serious balance issues with the game. Imagine, for instance, a 2 mana 2/3 with "Immune to Holy spells." That minion is immune to all most Priest and Paladin removal spells, so putting it into play against those classes can create a sort of "virtual card advantage" - basically turning off one or more cards in your opponent's hand. Having a card which offers tempo and card advantage exclusively against a subset of classes could warp the metagame arbitrarily.
That's not to say that these effects shouldn't exist at all, but that example serves to highlight the way in which this sort of mechanic is more powerful than it looks on the surface, lopsided in its effects, and therefore inherently hard to balance. It starts to look a lot like Inspire as a mechanic - something which is overly rewarding a player for doing the kind of thing they'd be doing anyway (in this case, developing their board). It should be noted that Inspire was a largely failed mechanic because of how significantly underpowered the minions had to be in order to make it fair, and I think you'd get exactly the same thing here - either the minions in question would be underpowered and useless, or they'd be good enough to break the metagame.
I think FortyDust has it right by saying that the big thing here is that there's really no competition for the 7-drop slot in Paladin (particularly after rotation), so this card is actually probably very good in context. Where this is probably particularly relevant is in Pure Paladin, but with only High Exarch Yrel left to support that archetype after rotation, the archetype may be irrelevant.
It seems like Peon is probably outclassed by Wandmaker. In the early game you have decent odds of triggering this frenzy, but it will be hard to do later, and in the early game you'd almost always prefer getting a random 1-drop spell over a random spell because you can utilize it sooner. But, with so many cards rotating, maybe the pool of random spells will be really good for one or more classes, and maybe the pool of great 1-drop spells will decrease, helping Peon to edge out Wandmaker.
Obviously this looks a lot like Lightsteed - maybe triggering off of holy spells (rather than healing) will make this better? One of the challenges with Lightsteed is that the buff is pretty minor on top of the mana you're already spending to target your own minions with healing (often via hero power), and the fact that it included non-friendly characters made it worthless with Circle of Healing. So, perhaps the randomness here is offset by getting to spend your mana doing valuable things instead of just board-wide heals, etc.
Razormane Raider is surprisingly playable with Animated Broomstick - if you line your trades up right, he could clear an enemy and deal 5 to face. That's probably not enough to get this card to see play, but it's something.
Feels like a meme card, but maybe there will be one or two combo decks that want him. Also, depending on how many strong battlecries and frenzy effects are in the game, he might be a decent (albeit super greedy) value play. For example. you could bounce Ysera the Dreamer, potentailly swapping the 4/12 for an Emerald Drake and getting that battlecry again. If that's how he's most useful, then Shadow Hunter Vol'jin will probably be a bit too greedy to be played much.
[Hearthstone Card (Imp Swarm) Not Found] is solid reload for a Zoo deck, but these imps die easily to most classes' AOE - even after rotation (e.g. Immolation Aura, Consecration, Holy Nova, Lightning Storm). That doesn't mean it's a bad card, but might be a little underwhelming against any control deck. It's really going to depend on what synergies are out there.
Bru'kan is clearly super powerful, and between him and Imprisoned Phoenix you might be able to set up some crazy OTK turns for Shaman if they get the right Nature spells. That said, the only Nature face damage in the Core set is Lightning Bolt, so even with 2x Phoenix in play, you're only looking at 20 damage max burst with that combo (though there are some ways you could probably eke out another 6 damage with Rune Dagger and 2x Novice Zapper). Ultimately Bru'kan is probably better as a way to burst down/finish your opponent after solid midrange play.
Desperate Prayer is probably bad, but a zero cost spells have utility with Gadgetzan Auctioneer, and a free heal effect could have value with any Holy spell synergy or a "whenever/after a character is healed" type cards
Sunscale Raptor seems awesome if you can get him down early, but given the high number of 3 damage removal tools in the core set (Eye Beam, Quick Shot, Holy Smite, Lightning Bolt, Drain Soul), if you don't trigger its frenzy pretty early, I think it's going to be very easy to deal with.
The other challenge is Hunter's historically weak draw tools, but maybe the new Selective Breeder helps you to really snowball this effect.
I haven't seen that - can you cite the remarks about card draw please?
They took away a ton of Warrior's best card draw/cycle tools (and all of their bad cycle tools too). Control Warrior's consistency with getting cards like Brawl was a big part of the problem with the card. Its overall powerlevel has remained the same, but we can probably expect Control Warrior to have it in their hand less often, making it a more reasonable inclusion.
It's definitely not too random to see play. If Peon read "Battlecry: Add a random spell from your class to your hand" it would be very playable in value-oriented decks. The fact that he has to survive damage to get the spell is the real challenge here.
This is probably exactly right, and exactly why they were dumb to set up their new regions/expansions like this. Without a full set of cards up front, they're forced to overcompensate. (I also don't really get the whole "some regions get expansions first" thing, but oh well.)
Spider Tank's mech tag may synergize with certain archetypes in a way that makes it relevant despite being (mostly) outclassed by updated class cards like Warhorse Trainer and Void Terror.
There seem to be two major questions here (and a brief, but seemingly irrelevant discussion of talent trees): should weaknesses/resistances to spell schools be a feature of Hearthstone, and should those weaknesses/resistances be implicit based on typing.
As far as the second question is concerned, I think this answer is plainly "no." Hearthstone is a relatively user-friendly game, and the lack of discoverability associated with those connections would be very hard to deal with. First, you'd need tons of helptext on every spell to explain what is or is not weak to it, then you'd need some tooltips to show you which enemy minions are of a given type, and all of that would have to be overlaid onto the game interface. It would be a nightmare.
The first question is a bit more interesting, but I think it's also a bad idea outside of anything but a very small number of Legendary minions. If it isn't very narrowly scoped in its application, you could easily get serious balance issues with the game. Imagine, for instance, a 2 mana 2/3 with "Immune to Holy spells." That minion is immune to all most Priest and Paladin removal spells, so putting it into play against those classes can create a sort of "virtual card advantage" - basically turning off one or more cards in your opponent's hand. Having a card which offers tempo and card advantage exclusively against a subset of classes could warp the metagame arbitrarily.
That's not to say that these effects shouldn't exist at all, but that example serves to highlight the way in which this sort of mechanic is more powerful than it looks on the surface, lopsided in its effects, and therefore inherently hard to balance. It starts to look a lot like Inspire as a mechanic - something which is overly rewarding a player for doing the kind of thing they'd be doing anyway (in this case, developing their board). It should be noted that Inspire was a largely failed mechanic because of how significantly underpowered the minions had to be in order to make it fair, and I think you'd get exactly the same thing here - either the minions in question would be underpowered and useless, or they'd be good enough to break the metagame.
Why would this change back to 2 damage? It still can't target face, which was exactly the trade-off that Blizzard made when they upped the damage.
I think FortyDust has it right by saying that the big thing here is that there's really no competition for the 7-drop slot in Paladin (particularly after rotation), so this card is actually probably very good in context. Where this is probably particularly relevant is in Pure Paladin, but with only High Exarch Yrel left to support that archetype after rotation, the archetype may be irrelevant.
It seems like Peon is probably outclassed by Wandmaker. In the early game you have decent odds of triggering this frenzy, but it will be hard to do later, and in the early game you'd almost always prefer getting a random 1-drop spell over a random spell because you can utilize it sooner. But, with so many cards rotating, maybe the pool of random spells will be really good for one or more classes, and maybe the pool of great 1-drop spells will decrease, helping Peon to edge out Wandmaker.
Obviously this looks a lot like Lightsteed - maybe triggering off of holy spells (rather than healing) will make this better? One of the challenges with Lightsteed is that the buff is pretty minor on top of the mana you're already spending to target your own minions with healing (often via hero power), and the fact that it included non-friendly characters made it worthless with Circle of Healing. So, perhaps the randomness here is offset by getting to spend your mana doing valuable things instead of just board-wide heals, etc.
Razormane Raider is surprisingly playable with Animated Broomstick - if you line your trades up right, he could clear an enemy and deal 5 to face. That's probably not enough to get this card to see play, but it's something.
Lots of stats for the cost, and a little face damage - Mankrik is a great card that could slot into just about any deck.
It doesn't take that much work to make Blademaster Samuro a very effective AOE. Seems like a great card for a lot of board-centric decks.
Feels like a meme card, but maybe there will be one or two combo decks that want him. Also, depending on how many strong battlecries and frenzy effects are in the game, he might be a decent (albeit super greedy) value play. For example. you could bounce Ysera the Dreamer, potentailly swapping the 4/12 for an Emerald Drake and getting that battlecry again. If that's how he's most useful, then Shadow Hunter Vol'jin will probably be a bit too greedy to be played much.
[Hearthstone Card (Imp Swarm) Not Found] is solid reload for a Zoo deck, but these imps die easily to most classes' AOE - even after rotation (e.g. Immolation Aura, Consecration, Holy Nova, Lightning Storm). That doesn't mean it's a bad card, but might be a little underwhelming against any control deck. It's really going to depend on what synergies are out there.
Solid tempo card that synergizes with Bru'kan - seems great.
Bru'kan is clearly super powerful, and between him and Imprisoned Phoenix you might be able to set up some crazy OTK turns for Shaman if they get the right Nature spells. That said, the only Nature face damage in the Core set is Lightning Bolt, so even with 2x Phoenix in play, you're only looking at 20 damage max burst with that combo (though there are some ways you could probably eke out another 6 damage with Rune Dagger and 2x Novice Zapper). Ultimately Bru'kan is probably better as a way to burst down/finish your opponent after solid midrange play.
Desperate Prayer is probably bad, but a zero cost spells have utility with Gadgetzan Auctioneer, and a free heal effect could have value with any Holy spell synergy or a "whenever/after a character is healed" type cards
Sunscale Raptor seems awesome if you can get him down early, but given the high number of 3 damage removal tools in the core set (Eye Beam, Quick Shot, Holy Smite, Lightning Bolt, Drain Soul), if you don't trigger its frenzy pretty early, I think it's going to be very easy to deal with.
The other challenge is Hunter's historically weak draw tools, but maybe the new Selective Breeder helps you to really snowball this effect.
Druid of the Plains is pretty great on its own, and should be a staple of any Midrange/Ramp Beast Druid.