I wasn't looking at it from a consistency point of view. To illustrate the issue I have with the hand limit I just use the following example: Player B has won the coinflip, it is round 1, player A has a decent lead in points and has 6 cards in hand; it is the turn of player B. In old Gwent usually you'd have to check if you can take the lead within two turns to not go down two cards. Because of the hand limit player B can go down two cards if they have at least 4 cards in hand at the end of the round, then drypass and go into round 3 with even cards. So in the scenario described above they don't even have to check wether it is possible take the lead with two cards (without overcommiting), because it doesn't matter. So let's say player B plays another card. In old gwent player A could decide to pass if A felt like B needs to play another 2 cards to take the lead. This isn't the case anymore, because B can go down two cards and still go into round 3 with even cards + last say by just drypassing next round, so player A wouldn't get any benefits from passing in such a situation..... I'm missing the feeling of having to evaluate every turn if I should pass and the possibility of punishing my opponent for not passing or being punished by my opponent for doing the same.
It never is a one-issue problem. The hand limit does both of those things to limit you. Moreover, there were cases where despite the importance of card advantage, you could still go 2 cards down in the old Gwent (sometimes you had to), Notably, against the old Reveal deck, the golden rule was to never pass. I don't like that this evaluation is gone either.
As for consistency, better blame other questionable changes like allowing only two bronze copies.
Like above, it's never one factor causing the whole problem. This could be mitigated if you had 3 sets of similar engines and certain thinning capabilities. Same way old engine decks (that worked at least) had 2 sets of 3.
ST had elven merchants or whatever the unit was called, which could tutor spells.
Aside from the fact that this was their only one, it functioned just like Vicovaro Novice. I am talking about a unit thinner, because engines are based on units. Consistent access to the units that power the engine is important.
I agree, but I had a big problem with all the tutoring in the game after midwinter. It wasn't really about the consistency, but having 5-6 cards in every deck, which just play other cards is kind of incredibly boring. That being said I think HC doesn't have enough thinning. You just have to be careful than designing thinning cards. The worst design approach (and what we've seen a lot after the midwinter update) are unconditonal tutors like post-widwinter pirate captain. A better design approach are conditional tutors. Examples are old slyzard, which required a unit in the graveyard or the current Wild Hunt Rider, which requires you to have the biggest unit one board. These conditions are rather simple to meet so there could be certainly more challenging requirements. Another option would be to make it event based. Like one very simple example would be something like "Every two times you apply bleeding to a unit, play a copy from your deck".
I like that idea yeah, an engine that is also a thinner. But it wouldn't fit in either version of Gwent due to the rampant removal. The conditional approach seems to work better, a possible vampire tutor could be what you described, but a one time deploy ability. A Dimun Pirate Captain that fit in an Axemen deck could be "if there is a hazard on the opposite row, play a Dimun unit from your deck" or maybe even "when this card is discarded from your hand, play a dimun unit from your deck" for the discard synergy. Something along those lines.
In general I think Open Beta gwent started losing interesting card effects gradually but that reached its peak during midwinter. Where Dol Blathanna Bomber (then Trapper) initially was: "Ambush. Spying. When a unit next appears on the row, flip and damage all units on this row by 5" (later became adjacent, and 4), and Vrihedd Sappers were "Spying. Whenever you play a special card, damage the entire row by 4" (again, over or under-powered is not the point, as values could be tweaked, but how interesting the effect was), now we had "spawn an incinerating trap on the opposite row" and "flip after 2 turns". Not to mention the old Drummond Shieldmaiden which was one of the best designed card effects of the entire game imo
Notable exceptions of course existed (like Vandergrift, an excellent card effect, even though it lacked distinct synergies), the effects became much simpler and what most people affectionately called "dumbed down"
While in the game now, more complex abilities do exist but with virtually inconceivable synergies, effects themselves and conditions.
This affected all factions but the worst contenders were the top tier decks (Skellige with corsairs and pirate captains, Alchemy with Viper Witchers, dwarves and most elven decks).
I'd have liked if we got something like "Spying. If Spying at the start of your next turn, damage an enemy on this unit's row by the number of alchemy cards in your starting deck, then move to the opposite side" for the Viper Witcher. Without affecting any of its basic properties (power, base strength, ability to be resurrected by Ointment), you single handedly nerf the deck's rampant control capabilities and give the opponent a 1 turn window to counter play (lock, or in the case of spy decks, infiltrators). For Fangs of the Empire, I had a similar idea to the old Vrihedd Sappers, except with Alchemy cards and moving to the opposite side when the opponent passes. Both units representing the need to sneak past enemy lines, sabotage/assassinate, then return to their camp.
Though I am thinking Alzur's Thunder should have only targeted enemies. After all, the whole point of spies is that you don't know they are there to target them
As for over-consistency, I had that problem as well, but well, like I said, choosing between that and inconsistency, I'd definitely choose that.