Sprites are underpowered?

[Seravy]

Considering they caused me to lose one game and contributed absolutely nothing in the next, I'm worried about Sprites now. They might no longer be worth using, and/or spending a starting pick on.

I decided to run the sprites treasure calculating tool for new maps.

I measured 4300 and 3400 average points on two sets of 8 maps, which is a fairly sharp drop from the old maps where we measured 5900.
Considering the consensus was that they are not worth it on Myrran maps where this number was ~2900, there is a chance Sprites end up underpowered and need buffing. - I'm pretty sure we all agree they are a horrible unit to use against other wizards due to their huge cost and nonexistent durability - 80 mana for a unit that can be killed for 10-20 (fire bolt x1 or 2, or other common spells) is not playable even if it is fairly good if it survives. Unfortunately it doesn't have enough ammo to capitalize on the advantage of being a ranged flier very well and every time a new battle begins, it gets hit by combat spells again, rapidly decreasing the size of the army.

Now, the treasure change was needed anyway - no doubt about it, the old unit values were poorly balanced, easy units were worth a lot of treasure. But that doesn't mean the game is fine when a unit that is supposed to be somewhat above average (Sprites are still a nature creature) is actually below instead.

The main problem I see is, this is ultimately a luck based strategy. If the average is "barely worth it" it means a significant percentage of games will be ones where they are not, and the cost is high enough to cause those games to be lost (you both need to give up a common spell pick or two with earth lore, and pay the massive cost of sprites - especially bad if the first target you find requires 5+ of them yet you find no other targets later.)
Even the games where it will be worth it, won't produce as much advantage as other, non-luck based strong strategies do, ultimately making this strategy bad. (If this wasn't bad enough, sprites were one of the most fun part of the game - they provided a unique strategy no other realm or spell did.)

So... the task is the same, test sprites, except, this time, please prove that they are WORTH using, and can provide enough advantage to be on par with a war bears or giant spiders or ghouls based strategy.

Considering they were already at best equal to other strong strategies in my 3 tests, the 30-45% drop in available treasure is worrying.

[Nelphine]

I regularly protect nodes with one sprite. They die a lot, but I almost never lose the node. And at least 30% of the time I don't lose the sprite. It also allows me to escort magic spirits with minimum of effort (even if it isn't perfectly effecient)

I'd never try that with war bears.

By itself, this shows me that sprites are worth it.

[Seravy]

Wait, so 70% of the time you lose the sprite?
Isn't that the same as losing the node? There won't be units left on it so the enemy can put whatever they have nearby on it or send spirits.
(also, 80 mp for something that dies in 70% of the battles is a horrible deal.)

[jhsidi]

Tried out a short run, did pretty well. My setup: 3 nature, 6 sorcery, omniscient, sage master, conjurer. I was thinking I'd go the focus magic route with my sprites, but didn't bother as it happened.

Nearby-ish: neutral high men; ruins with phantom beasts; skeletons; and a node with phantom beasts / warriors just by a chaos wizard with lizardmen.

I took out the targets in the order listed above, which netted a small amount of mana from the first 2 victories: about 260, plus I was building marketplace first in my outposts for alchemy. By 1203 I had the 4 targets down and had researched giant spiders, thanks to sage / omni w. sorc / halflings. I then summoned a stack of 6 spiders, joined it with 3 sprites, and went after the chaos wizard, who had a capital full of hell hounds / halberdiers. I squeaked through (fire wall is really nasty!) and by 1205 I finished completely destroying that wizard, adding 6 cities. Wouldn't have been possible without the initial acceleration from just 2 summoned units of sprites (enough to take down the phantom beasts).

Of course, some games now have no targets. The way I approached it was, I cast earth lore before investing a lot in sprites. Once I saw I'd get payback, I went ahead and poured all resources into my sprites. Won't get my victory by 1408 or anything crazy like that, but OTOH sprites gave quite a good start. In case all targets had been poor for sprites, I could have done war bears instead and still gotten by decently.

One add'l note: I did have mostly nature nodes near me (only mentioned the good targets) but these now look impossible for sprites: every one contains stone giants, lots of spiders, or lots of sprites. If the goal was just making the strategy more difficult rather than closing it entirely, I think nature nodes may have overshot. Nodes w/ just cockatrices as a threat seemed reasonable to me (your sprites are on the same playing field as a bowman would be versus a pikeman), but I haven't seen a node like that since the rebalance.

[Seravy]

Cockatrice wasn't directly changed so it's due to the shifted probabilities - there are a lot of different types of Nature monsters so the chance to pick cockatrice is most likely lower, especially to do it at the same time as picking another easy monster for the other slot. In particular the chance to pick two "possible" monsters, or have only one type is overall far lower - in the old system if the game rolled a high number (like 4-6, 50% chance) then due to how costs are, cockatrices were a likely pick for the primary monster - they are pretty far from other monsters in both directions (spiders are ~100 cheaper and earth elementals ~100 more expensive) so overall over 25% of all nature nodes had it as the primary monster - and while the secondary could have been something like a spider, even if it was, it often only meant one or two of those monsters.
In the new system however the game just picks two monsters at random from the 12 possible Nature creatures, and then a random amount for each - an 1500 budget node now has the same chance for each expensive monster as cockatrices while in the old system it required a roll of 1 out of 6 to get any of the 3 very rares in, and no more than 3 to get the rares.

On the whole, including strong monsters (even if it means a lower amount) is probably more frequent.

[jhsidi]

Yeah, it still feels like I see more strong combos than before. But overall the new system feels like an improvement. I'd rather have the challenge of fighting 1 stone giant than the equivalent strength of war bears, for instance -- it's just more interesting gameplay (if occasionally frustrating).

[Nelphine]

My single sprite virtually always wipes out all the enemy. Then I send another sprite out to guard it again. Occasionally I lose the node, bit its definitely rare. This is one reason why a few games back, if o went to war with the peaceful halfling who's empire intertwined with mine, I would have lost so many nodes.

80 mana for something that maintains a node (dozens of mana per turn depending on power setting) and wipes out hundreds of hammers + mana worth of attackers (not counting the mana the AI and I use during the actual combat) is worth it.

Obviously it doesn't work against every wizard type, but it works in the majority of games I play. And flying really is key to making it work, because most times, the AI has huge swarms of things like halberdiers running around, which means its entirely their combat spells vs my combat spells + sprite.

[Seravy]

I don't really see how 4 ammo is going to wipe out an entire enemy army.
...and you can't use combat spells because the node counters it which is why sprites work instead of getting killed the first time the enemy casts a spell. (By the time your combat spells are so expensive that they have a good chance of success, you won't be using sprites)

I've used sprites to protect nodes before and it generally worked like this for me :
-1 sprites is not enough. Yes, if I get lucky and the node always counters everything it works. If not they die in the first battle. So I need at least two because gambling with a node is not a good idea, especially when it is far, and it usually is because if it's in my own territory I don't really need to defend it - whatever units I have nearby can take care of enemies.
-Sprites were able to stall for a few (about 2-3) battles before the AI gets enough combat spells past the node effect to kill them
-The sprites don't do enough damage to kill larger groups of enemies - if such a stack attacks they are worthless. Yes, I might kill a unit and they'll be forced to retreat, but they will be back next turn and the turn after and the wizard will succeed casting something before running out of units
-Of course if the enemy army has flying, thrown or breath units, such as ghouls or hell hounds (or more sprites than myself) then they're dead. Same if they have a javelineer, or just more than 1 bowmen.

So yes, they are fairly good if you have no way to have real forces on the node due to distance, but they aren't going to work against any real attempts to take the node and every one you leave there weakens your capacity to take out other targets as they only work in larger numbers against...well, the best targets that don't even exist anymore.

[Nelphine]

I dunno, I rarely have a problem with the node countering my spells, unless it also counters the enemy spells. Maybe my games see a lot more melee on arcanus than you do. I 'normally' play life on myrran, so my experience is not super high.

I think the best (and albeit super niche, so not worth relying on it, and not what I'm basing my conclusions on) example I've got is my current analysis 2 game, where 1 sprite on each of the four nodes in his territory is holding off the entire AI.. Who is 9 life 3 sorcery, no focus magic that I've seen, no angels, and the only cities he owns are gnolls, even if he has 11 cities on the same continent.

[Seravy]

Oh, I have one more thing to add - All tests that were conducted to prove sprites are overpowered used Earth Lore for obvious reasons.

However, for tests to prove sprites are not underpowered, we should not use it - a good spell is viable on its own, combo potential is only an extra. I don't think there is a common spell in the game that's useless without knowing a specific other spell. So the suggestion to start out by earth lore already shows there is a problem - every other summoning spell works without having to explore first - you'll find something to attack, if nothing else, another wizard, or you can even use them for self-defense. Sprites on the other hand, well, they are bad for both. 2 mana maintenance in the early game is too steep for a garrison unit and their low durability and difficulty to replace makes them a poor garrison overall (yes, they are amazing if you have 4-5 in the same city but no one can afford that)