1) Remove some other creatures and find room for at least 3 Timber Wolves. They are quite tough for their cost, they work really well with Wounded Prey, they have no egregiously bad matchups (zombies and Iron Golems aren't great matchups, but they're a lot better than Dire Wolves or Emerald Tegu in those matchups), and they're about as expensive as you can play regularly out of a Lair without disrupting your ability to take other actions if you don't have time to significantly increase your channeling.

2) Pull Cheetah Speed for Mongoose Agility. You're already fast and run some fast creatures. Additional ways to get fast aren't as valuable as the ability to get away from your opponent's creatures in a single action. Being able to ignore hindering is huge. I always want to run at least 2 Mongoose Agility with the Johktari Beastmaster.

3) Take out the Hail of Stones. You already have a big investment you want to make with your full actions (bow), and when you're not doing that, you generally want to be repositioning. I much prefer taking range 0-2 attack spells on the Johktari Beastmaster. With Academy you can opt for Kajarah, in which case it makes sense to use more range 0-1 attack spells, but when I'm playing a Johktari Beastmaster centered around a bow, I like to maintain range 2 as much as possible.

Counterstrike never flips an action marker. If the creature making the counterstrike has a guard marker, the guard marker is removed, but the action marker isn't flipped.

That's the way that I play the Vine Tree.

I like the ideas of both books, but I feel like both of them could be improved with some refinement and switching mages. The Priest is able to daze with his attack, but the Priestess can cast healing spells much more efficiently, and is better at stacking armor in order to make healing spells even more efficient. Your opponent can't strip your armor off with corrodes very effectively when you can just remove them all with one action. The higher base channeling would also help with being able to summon slightly larger creatures.

Healing spells should not be an alternative to armor, they should complement armor. The stronger you are defensively, the more efficient a healing spell on you is, since it takes more work to deal that much damage to you.

The Druid could have been at least 2 rounds faster, and hit harder as a Beastmaster. Yes, the Druid has more defensive wiggle room, but you want to go on the offensive if your opponent is playing a stronger economy than you. Spamming enchantments can let you do a reasonable amount of damage, but eventually an opponent that has 5 times as many creatures as you will overwhelm you.

You spent too much on playing every conceivably useful enchantment when you could have had another creature up with similar enhancements and another backup creature in case one of those got killed instead. You put too much emphasis on that one threat at a time, which made it much easier to deal with. Your opponent was able to keep pace with you by just casting a healing spell every round, which means you weren't getting to the point of dealing enough damage, although a Poison Blood would help.

Why play Harmonize on Fellella? That just makes her a more tempting target, and it's not that more efficient than other economy options.

I like this as an opening better:

Turn 1 (19): Fellelle, Pixe Familiar -> Enchanter's Ring (5)
Turn 2 (14): Pet Timber Wolf -> FD Enchantment -> FD Enchantment (0)

The Timber Wolf gets stronger faster, you can use Joined Strength, and the option to use Vampirism + Packleader's Cowl makes it much harder for your opponent to win a direct engagement.

The way you were playing, it would have been better to put Teleport on a Mage Wand and Teleport the Timber Wolf out immediately after attacking each round. That way it wouldn't have gotten overwhelmed, and you could have tried to force your opponent to move out, where your vine markers could have come into play more meaningfully.

Tele-pit is the strategy that comes to mind when you mention waiting. You play big powerful, immobile creatures like Darkfenne Hydras, Iron Golems, and Devouring Jellies while setting up a method of locking down your opponent. Then you Teleport your opponent in and kill them before they can escape. Your opponent doesn't really want to directly engage you because your creatures are more mana efficient in a direct engagement. You still need to defend yourself from attack spells, but this strategy is best implemented by a Wizard anyway, and Voltaric Shield + armor and a Gargoyle Sentry or 2 helps a lot with that kind of thing.

There are other times when waiting is advantageous. If I have a big economic advantage, I would prefer to have a little time for that economy to generate resources before engaging my opponent if they'll give me the time. I'll use the Warlord as an example. If I play a Warlord with heavy economy (Barracks, 2 other outposts, Battle Forge, and Meditation Amulet), and my opponent is producing two thirds as many resources as I am, they are the one that will want to start fighting right away. If they'll let me sit back and build up 4-5 creatures before I'm ready to move out and engage them, I'm happy to keep building my incremental advantage.

Sending in creatures in isolation is a good way to get them killed for minimal return. I've seen this happen with a wide variety of creatures. So while this isn't an argument for waiting, it is an argument to be cautious of rushing. Sometimes the deliberate approach is better.

Altar of Carnage will never be a good distraction because if they kill your Barracks, it loses most of it's utility. If they decide to attack your economy at that point, they won't try to destroy the Altar directly, they'll either kill your creatures (if you use vulnerable ones like Goblin Grunt) or your Barracks. Either is a much more efficient way to make playing the Altar relatively useless than attacking the Altar directly.

It can pay back it's cost in 2 rounds pretty easily if you have 4 creatures, which is faster than any other economy card in the game besides Meditation Amulet. In that sense, it's actually a better mid game economy card than the other possibilities.

I don't use it anymore because it's an additional investment in the Barracks, and the Barracks already forces you to commit heavily to it before it operates efficiently. Once the Barracks is gone, you're reduced to spending that mana on the Battle Forge or Gurmash. I haven't found Gurmash to be particularly exciting, and flooding him with mana won't change that in my opinion. Battle Forge benefits somewhat from the mana, but by the time you've played a Barracks, 2 other outposts, 4 creatures, Battle Forge, and Altar of Carnage, you should have enough equipment played that most of the time the extra mana will be of somewhat limited use.

Besides, if my opponent leaves me and my conjurations alone to build when I'm generating 17 mana/round, I'll win anyway, no altar needed.

I haven't played Pixel Tactics 3 that much, but yes, it's fun. I played a lot more of Pixel Tactics 1, where the balance is heavily skewed towards orders. 3 is more evenly balanced between different kinds of strategies and uses of cards, but I don't remember the specific cards in that set at the moment.