Guardian Spirit

[Seravy]

Does anyone use these? Because I don't (unless my army needs +1 resistance really badly - but not for the melding ability).

-I feel it's inadequate protection for my nodes - yes, there is a 80% chance it will prevent melding attempts but at the rate AI players send spirits that won't save the node. Being able to hold it a year more is of course nice but I don't plan for losing nodes - I rather put garrison on it so I have a 100% chance of keeping the node.
-It's also inadequate protection for the AI's nodes - to begin with, they garrison all nodes anyway (albeit some might have week garrisons), and once the garrison is defeated, taking the node away is just a matter of turns - summoning some additional spirits until one works.
-Despite being unless, it manages to be extremely annoying, when your 12th spirit fails to meld a node it's hard not to think "this is the most retarded game mechanic ever".
-Also, it's a random chance based spell in Life of all things, not Chaos.

There are plenty of things that could be done with this unit instead...

A. The node cannot be melded, period. However Disenchant Area can remove this effect just like Warp Node. (this would probably be bad for game balance and AI)
B. Other Guardian Spirits have a 100% chance to meld a node with a Guardian Spirit
C. Drop the melding ability altogether and come up with something more interesting, or just leave it as is for the resistance bonus.
D. Add "This creature can be summoned in combat when defending your cities only" as a new, more interesting ability.
E. At the end of every turn when there is a node with a Guardian Spirit with an enemy unit standing on the tile, materialize a Guardian Spirit unit and engage the army as the defender, allowing the wizard to cast spells to protect the node. (don't think I can do this one but might be worth a try)
F. something else?

[zitro1987]

My take:

Guardian Spirit - 60/12 cost, 1 upkeep
Combat stats - same
*Some kind of defensive bonus (like 'guardian retort') for your troops when defending the node - in place of the harder to meld mechanic
*Can be cast in combat while defending cities, costing only 12 skill.

[zhantul]

I tend to use these a lot to pick of small groups of enemies when facing death wizards. They are strong enough to handle simple enemies and are safe from possession. Even a single spearman is costly to defeat otherwise. (bless is also an option, I know. Multiple options are better.)

Also, They are a cheap, common summons with reasonable strenght. Life has no other common summons. I like the idea of the 100% meld chance. That restores a bit of balance for the human player.

A great but not overpowered improvement would be to raise their hit points by one, to 11. That way, they are strong enough to survive a doom bolt. That should help with the warlocks that any dark elf ai sends out surprisingly early in the game (on extreme or higher anyway).

A final option would be to have heavenly light active in battles taking place on a node with the guardian spirit melded. The bonus is small, yet noticable and otherwise unable to be cast due to being a city spell.

[Nelphine]

I think guardian spirits do what they are intended to do. Especially on nodes a long distance from the enemies city, they can make it feel impossible to get that node. (Consider, with a magic spirit, I can be sure that I only need to cast 1 magic spirit to take it from the AI. With a guardian spirit defending the node I will typically send a stack of 6 magic spirits, which forces me to spend much more mana. But it doesn't slow me down. On the other side, the AI has no idea there's a guardian spirit there. I've seen an unprotected node go 30 or 40 turns without being taken over by the enemy, because they only send one spirit at a time. If it takes that spirit 10 turns to get there, its a very long time before the next one gets there.)

However, I have never ever spent the mana to summon a guardian spirit. Intercepting incoming spirits is almost always more cost effective, because by the time I lose my garrisons I typically have mobile offensive stacks, OR I've already planned to lose the node. If its much earlier, the guardian spirits extra cost isn't worth losing a cast of endurance or heavenly light or focus magic. Human has too little spare casting skill.

The AI on the other hand can't do roving mobile stacks. They also have loads of overland casting skill. I think guardian spirit is basically as important as the guardian retort.

It allows for certain styles of play from humans, and it makes up for things the AI lacks the ability to do.

[Seravy]

I think it's marginal for the AI.

There aren't enough picks, so the AI will not have Guardian Spirit at the start, and it's relatively far on the research list so they won't have it in the early game. The AI also cannot "upgrade" the spirit from magic to guardian, so they'll only have it on nodes obtained midgame or later, which..implies the human can afford the 5 spirits already and not care.

If we want the AI to hold on to nodes better, we can raise the minimal garrison requirement from 3 to 5 or even 7.

I like the "heavenly light in battle" idea, and it would be even better if the node provided an extra 3 power the same way as heavenly light.

[Nelphine]

I'd never waste time on heavenly light combat bonus. Literally its a waste to garrison nodes. Instead you keep a roving stack near several nodes. The AI wastes troops garrisoning the nodes for you, and you can still intercept and kill all the spirits they send towards the node.

The benefit of the current guardian spirit, is to force the human player to spend more mana, and more importantly, casting skill. This is (for it's cost) just as important as to the AI as the guardian retort. I agree its marginal. And if you add in the extra power generation, then I'll use it in a heartbeat. But I'm not sure why you'd improve this, and not improve the guardian retort. To me they fulfill the same role in the game.

[Seravy]

I'd never waste time on heavenly light combat bonus. Literally its a waste to garrison nodes. Instead you keep a roving stack near several nodes. The AI wastes troops garrisoning the nodes for you, and you can still intercept and kill all the spirits they send towards the node.

The benefit of the current guardian spirit, is to force the human player to spend more mana, and more importantly, casting skill. This is (for it's cost) just as important as to the AI as the guardian retort. I agree its marginal. And if you add in the extra power generation, then I'll use it in a heartbeat. But I'm not sure why you'd improve this, and not improve the guardian retort. To me they fulfill the same role in the game.

Oh, you used "as important as" in the sense of "as worthless as". I see. I thought it meant "as essential as" and you were saying guardian spirits are necessary for the AI.
I see two main differences :
-I want to pick the guardian retort but not the spirit. (I admit I didn't have enough picks for it lately, but the same can be said about 75% of the retorts in the game. I also can't afford Myrran, Famous, Chaneller, Charismatic or Astrologer, Artificier etc)
-The guardian retort is not marginal, it provides a massive combat boost in a very large portion of my battles, and in the more important half. (Yes, attacking is more important but I can build strong enough offense to do it. Building strong enough defense in an empire of 20-30 cities on the other hand is very hard when the enemy has rare+ combat spells.). The spirit is marginal, it affects a small portion of the game (nodes) and only delays the inevitable loss, while the Guardian retort can provide enough edge to never lose your cities.

btw there is one case where Guardian Spirit matters, and that is Suppress Magic. It makes the already unaffordable 1000 cost of melding a node into 5000...which is hardly relevant I guess. I wouldn't spend 1000 on melding a node either.

[Nelphine]

Right, but 2 picks is a much larger opportunity cost than 50 mana. For its cost, they do the sane thing. They only help in very particular situations, both of which the human player can largely avoid anyway. Mobile rovibg stacks are better for defense of both nodes and cities. And static garrisons should only be used when the AI has an overwhelming numerical advantage. If I have one stack that can move 7 squares per turn, then I can hit 2-6 stacks with it before any if them get to the city/nods. Only when the AI can saturate even that speed does the retort or the spirit matter.
On the other hand, the AI cannot use roving stacks. Therefore, it is important for them to have static defenses, and anything that boosts those is helpful.

The guardian retort is not worthless. It has a very important place in the game. But its not a place the human player should be aiming for, because that is trying to beat the AI at its own game. Same thing with the guardian spirit.

[Seravy]

very particular situations, both of which the human player can largely avoid anyway.

Battles in my cities is something I can never avoid on high difficulty. Moving stacks certainly are better...until there are too many enemy units to catch them all, or the enemy comes with ranged units and giving them the first turn is suicide...or if they have strong spells to cast on turn 1.
Or until they produce pegasai or wyvern riders and hit from a distance of 5. Or get wind walking stacks.

Also, moving stacks won't do a thing when you break an alliance to start the last war and there are 25 enemy stacks on your continent, like, in my current game right now (I'm starting the war next turn) :

   

This is a game where the enemy is significantly weaker than me. If they weren't, the map would look like this everywhere on both planes, and those would be archangels.

Being defensive works well for the human. The AI will repeatedly send in stacks that can't beat it because it's dumb - having Guardian makes this work better - a larger percentage of your defensive battles will be won. For the AI it's the opposite - the human player never sends stacks that cannot win, so they need guardian to raise the amount the player needs for a successful attack - it's not helping them actually win the battles (unless the human player fails to notice the retort or is inexperienced). The overall result is the same - the Guardian gets to keep more cities longer - but in case of the human player they also get to kill more enemy troops with it, which the AI won't.

btw, if you want to take out enemy stacks before they reach your cities, assuming a visibility range of 4 and speed 2 enemy units while your units are speed 4, you have 2 turns = 8 movement to hit that stack. So your stack can cover an area of 8 in all directions, which is like 9 cities, sounds good? Well in practice you need like 4-5 of those stacks because multiple enemy stacks appear at a time, and not from the same direction. You also need at least 4 units in every city so a surprise attack by 1-2 fast fliers (or something invisible) won't take it out. At that point you have as much units as if you just had 9 in every city...and the latter gives you more unrest reduction.

I feel we already had this discussion before...

PS : I don't mind reducing the cost of Guardian to 1 but it will be overpowered as hell and I would use it in every single game. Probably not the best idea.

[Nelphine]

Absolutely. But my garrisons don't have to all be sky drakes. They can be spearmen, while my roving stack is sky drakes. And typically I find that I need 2-4 roving stacks for ~10 cities (and thats against a sorcery ai with windwalkers). Which is still far easier to acquire than 10 cities with proper defenders.

But yes. My point is that I think guardian spirit is as effective as that retort. Which means for the AI it replaces a decision making ability the AI doesn't have, while for the humans it can be used by those who like that tactic, and not used by others.

Just as I think that because you find the guardian retort effective, doesn't make it effective for me - but the reverse is true. I wouldn't presume to tell you that its ineffective for you because I find it ineffective.

Similarly, I think the guardian spirit is just ad effective. Which means, for me, its not effective. I think it is effective for the AI. You don't think its effective. But I don't think you should take that to mean its not effective.