Caster of Magic Release thread : latest version 6.06!

[Seravy]

The current chances for having "2 picks" as treasure are :

"20% chance for two doublepicks, 50% chance for one doublepick, 30% chance for no doublepicks"

So 70% of the maps are guaranteed to have one. Finding any additional picks should not be something the player can count on, especially as the pick found would be random anyway and others might find it first. Even if the map does contain it, there is less than 4% chance the player is the one who gets it and it rolls the specific 2 cost retort they wanted. So trying to rely on that is a bad idea. If you want a retort, you pick it when you start the game.

(maps have 1-5 single picks by the way, and you can get extra by finding a very rare spell prior to the time limit for those becoming available : there are 2-5 of those. So the total single picks available is 3-10 on each map, doubles are 0-2 but most often 1, making the overall amount 3-12.)

[Nelphine]

Incidentally, there were no more picks in my game; 5 singles and a double on arcanus, 1 single in the towers, 1 single on myrror (everything broken before very rares could be found).

While I admit this isn't enough to state there is a problem, is it possible to put slight requirements based on plane? Like at least 25% of all picks/very rares have to be on each plane? (In this game that would be at least 2 (or 3 depending on rounding?) of the picks/very rares would have been on myrror?


Also, I'd keep an eye on how weak the lairs can be that picks can be found in. I got a single and a double from the first 8 places I opened (the single had only a very very small amount of mana with it).

[Nelphine]

Another thought: can the turn limit for breaking towers be based on how well the human is doing on their plane?

I think there are four cases, including two cases we don't want: 

Case 1, the human is in the middle of a 2-3 way war on their home plane, and the AI from the other plane breaks through some towers, and declares war on the human or blocks anyone from switching planes while also gaining territory (even if it just scouts so it can city curse without reprisal) on the humans plane; the human simply can't cope.

This is what the current solution fixes.

Case 2, the human is in a 0-3 way war, and an AI from either plane breaks the towers, and people from both planes (that may include the human) gain territory on the other plane, and ai from each plane declare war on each other.

This case is ideal, as effectively both planes make up one big plane. However, it's extremely difficult to predict when this will happen, so avoiding it is a necessary sacrifice to avoid case 1.

Maybe we could add a big diplomatic penalty if you are from different planes? Not particularly important.

Case 3, the human is doing well on their home plane, and either the AI from the other plane or the human starts to open towers, and the human will be able to gain ~equal territory on the other plane as AI from the other plane will be able to gain on the humans home plane in the months/year immediately following the opening of the tower (which could include no one gaining territory).

This is the case the new rule aims to create.

Case 4, the human completely dominate their home plane, and are able to build up at their leisure to be ready to block many or all of the towers (whether they can defeat the towers or not). 

This is the second case I believe is bad. I'd like to suggest that towers can be cracked by ai if: turn equals x (whatever it is now), OR any other tower is already cracked (as they can now), OR if the human has at least 3 times the overall graph of all the living ai on the humans home plane combined (or if all ai on the humans home plane are defeated) AND the human has at least 80% the overall graph of the combined ai on the other plane.

[Seravy]

I think case 2 is extremely unlikely to happen due to the sheer difference in troop strength and resource availability between the planes. The Myrran wizard will conquer the territory of anyone else they manage to start a war with. The only possible scenario of mixed planes is if the Myrran doesn't declare war on anyone (or if towers broke so early not even the Myrran has a good army yet, but that's rare.).

So no this case is not ideal - in fact this is the case that made me want to fix the problem. The Myrran wizard breaking through and kill stealing 2/3 of the territory of the enemy I was finally strong enough to start to conquer is unfun, and unbalanced.

The way the game works, in 1410 the Myrran will have dozens of stacks of armorer's guild units (hammerhands, etc) often with adamant, while the arcanus wizards have... well, nothing that has a chance to stop that.

Case 4  - This is the description of what happens if you play a strategy too strong for your chosen difficulty level or get extremely luck. I don't think it's a problem, and blocking the 9 towers would imply you have 9 strong stacks to attack all 9 towers simultaneously. Otherwise breaking the first one allows the AI to break the other 8 (or whichever they have in reach of an army) the next turn. I'm not saying that's entirely impossible but 9 stacks good enough to break a tower in 1416 isn't that trivial to get. (Life wizards might have an easier time with buffs.)

OR if the human has at least 3 times the overall graph of all the living ai on the humans home plane combined (or if all so on the humans home plane are defeated) AND the human has at least 80% the overall graph of the combined ai on the other plane.

While I agree with these numbers, I don't think that's a very likely scenario unless the human is playing a difficulty significantly lower than they should. In which case they probably want the easy victory anyway...
Also, a human that powerful not breaking towers is extremely unlikely - players are greedy and want the tower and myrran treasures, ESPECIALLY if they are already winning. Even more so considering that the game does tell the player through the BGM if they are leading, not to mention breaking the towers is the correct move - letting the Myrran wizard develop or take treasure when you have the capacity to fight them is bad due to their resource bonus and superior race (unless the difficulty is low but then see above) - at least assuming the plane size difference doesn't stay at the current +20/-20.

[Nelphine]

Case 4 was what I just did in the game above, on Lunatic. And in 1412 there were exactly 4 lairs/nodes left that the AI had not yet conquered, so it's not like there was much treasure remaining that I could have grabbed even being greedy.

Yes, it is unlikely; but it also means that if someone accidentally achieves that (someone who decides to pick a master level strategy, but they've only just started at expert; they don't realize it's that strong), they're left with an unsatisfying game. That condition doesn't do anything to people at the correct level, but it does help with people who are using powerful strategies.

Also, i've literally never listened to the BGM in this decade. I can't imagine I'm the only one who prefers other music/watches videos/has family talking at the same time as they play.


Case 2: I've seen that, oh, a dozen times between version 3 and version 5. The most common example is a barbarian arcanus AI who breaks towers early (human is also Arcanus). Due to being barbarian (and sometimes gnolls can too; it also happens with lizardmen occassionally), they have so much strategic strength that the myrran AI can't effectively fight them. As the barbarian AI expands, they get weaker (strategic wise) due to accumulating more powerful races, so war breaks out with the myrran AI, and both the barbarian and the myrran AI steal territory from each other. It's a very enjoyable experience when it happens. I do agree it's very difficult to predict; it's based on race choice, personality choice, sometimes realm/retort choice; it requires specific diplomatic moves from the human. So I have no idea how you could code a method in to detail when to break the towers. Maybe, if the AI is on the same plane as the human, and the AI is stronger on the overall graph than the combined AI on the other plane, and the AI is not at war with the human then the AI can break the towers? (And add in a hefty diplomatic penalty for wizards who start on opposite planes so that they naturally tend towards war. However, I dislike that one, as diplomacy with the final AI should have some meaning, but I wouldn't want to restrict it to AI only.)

Anyway, not worried about case 2 actually being included. I do think case 4 is important though.

Any thoughts on my post 3 above? (#2016)

[Seravy]

4 lairs/nodes left that the AI had not yet conquered, so it's not like there was much treasure remaining that I could have grabbed even being greedy.

That's impressive, I didn't expect the AI to be that good... but anyway, if nothing else, the 9 towers are a big amount of treasure, not only are they all high budget but they also guarantee a spell.

Anyway, my problem with this is, it's difficult to implement, total graph strength is the sum of 4 values and then you'd need to calculate that for a sum of 4 wizards... not even sure if space, the potential for new bugs, or the time to implement is the problem I just have this feeling it's better to stay away from it.


2016, well, I rather not have a requirement by plane, sounds complicated. The towers, nodes and lairs are mostly evenly distributed so the chance of all stuff ending up on the same plane is extremely unlikely, and unless that plane is Myrror and the AI clears everything before the human gets there, it shouldn't matter (as arcanus treasure does get divided between 4 wizards - now that picks are less likely to get piled up in the hard places the AI can't touch, even more so).

The point cost of picks is kinda low for how good they are but that isn't new and keeps treasure hunting interesting. I'd say a pick is worth more than the entire budget of the best lair anyway so we can't make it have a cost based on that. 1 pick requires 800 points so it won't show up in the weakest lairs. If necessary we can consider moving it up to 1000 maybe 1200 but 800 is probably enough.

[Nelphine]

On both those topics I'm satisfied with that answer. You've considered them, see the value, and also see the cost with implementing. Keep an eye on picks over time - it felt too quick to me, but that's one sample; and on tower breakage.

[teelaurila]

On the design of the late game enemies:

I would like to see that the easiest way to fight the Myrran is not to conquer all of arcanus first, but to leave at least 1 AI standing and attacking as an alliance. (Maybe the rush-est builds could be an exception.) So clearing all of arcanus would generally be a fall-back for the player for when he must.

To this end, perhaps AI diplomacy could consider the power graphs a bit more, so that the snowballing Myrran AI would more easily incite animosity on arcanus? And also the clear leader could cause the losers to like each other more, so as to help ganging up. So opening up towers to introduce the Myrran "joint enemy" into the equation could even help the player preserve his alliance(s) when he starts to win on arcanus. And when the Myrran AI would be losing ground, the arcanus alliance would start to crack.

I know it's all complex maneuvering with many parts in the equation, so easier said than done. But I wanted to chip this in, as it seems like the present idea is that the player generally would/should clear arcanus. Which I don't think is ideal for the most interesting game, as the diplomacy part gets shrunk to be only relevant so as to avoid 3-on-1 war on the player early on.

[Seravy]

I would say having one Arcanus ally should already work as well as, or better, than conquering them all. It gives both players a lot of time to build up armies and resources, as the timespan of the last war essentially becomes "free time". Considering this war would destroy buildings and eat up resources otherwise, and there would be no time to turn the conquest fruitful afterwards, skipping the war should, in theory, not hinder spell research nor casting skill.

Unfortunately, this only works well if the ally has troops with relevant strategic strength, otherwise all they'll do is feed the Myrran more cities by losing them. Even then it's a great help - the Myrran will be spending a lot of mana on these battles (especially if they happen on Arcanus), so you'll see them without being able to cast combat spells more often.

[Nelphine]

I disagree Seravy. Wars on your home plans are early enough that your skill is low, and close enough that the cost multiplier for combat mana is low, and early enough that mana from treasure generally covers a very large portion of your combat costs.

So, all of your power can be dumped into research and skill, even if you're at war.

Further, the biggest gains the human can make is to conquer amplifying towers (more important than nodes). And the last opponent on your home plane is the richest source of these, so conquering them is the single most important thing you can do while preparing to fight with ai on the other plane.

And having those amp towers on your ally simply isn't as effective as having them yourself - less instant cast overland spells, and less combat spells, which are the two most important things you can cast.

And at least in some cases, your ally may declare war on you during the war with the other plane, especially if you're chaos or death or sorcery and want to use your most powerful very rare Global's; this would result in your rear cities being at significant threat.

Finally, your ally may not declare war on the AI on the other plane (if the other is say charismatic with all the same books as your ally), which you can't completely predict.


These 4 things combined mean it's almost always the best move to conquer your home plane before risking a tower, unless the tower is broken by the AI.