Hello, thank you for all your work on this mod. It's really great!
Quote:Spell of Mastery
Costs 10000 to cast. Researching it is 4 times slower than any other spell's.
Not a fan of this spell, as it allows you to outright win the game as long as you can defend your fortress. With any decent casting skill, you can cast the old version in 20-40 turns, way too fast for this effect. Even faster if you reduce the cost through the Runemaster retort. At this cost, you have to survive for a much longer duration without being able to use overland spells, making it an actual challenge if you go for it. Needless to say, with the improved power sources, and faster city growth, the old research cost was way too low and the spell was rather easy to obtain. Unfortunately the maximum possible cost was already used, so the only way to achieve this was to cut the research points generated while researching it instead. Researching other spells will also have an equally lower effect on reducing the research cost.
I think you may have gone too far in trying to balance Spell of Mastery. I have only played one game where I tried to go for it, so maybe I need to try again to get a better sense. But by the time I had no other spells to research, it still was going to take me over 200 turns to research, dumping all my raw mana into it. I then built a few more cities, putting Nature's Eye and Heavenly Light in every city and outpost, bought a bunch of Libraries and Wizards' Guilds, cast Tranquility and Life Force, and still it was going to take me over 100 turns.
I then did a bunch of military aggression and claimed several nodes.
Now in my game (after about 20 turns of researching but not investing a lot of raw mana into research) it's still 77 turns away from complete when I set my research to maximum. I have 14 cities, 165 total population, all Dark Elves (I took Inquisitor at the start), the Sage Master retort, Libraries and Wizards' Guilds and Universities in every city, and the massive global enchantments AND city enchantments that are supposed to boost your research power to its maximum. I've also claimed 100% of the nodes on Myrror and at least 30% of the nodes on Arcanus. I'm not going to wait another 50+ turns when I could just be ending it in about 30 turns with my overwhelming army of Nightmares.
Is Dark Elves a mistake? Do I need to go Halflings instead?
If the only way to get Spell Mastery is to wait around for 150+ turns when you could be ending the game the other way, what is the point of it even being an option?
My suggestion, based on this one game (so maybe I don't know what I'm talking about yet), would be to change it to one third research speed or even as fast as half research speed. I think it should still be a viable option for people who want to go for it and are set up to do so from the start (Sage Mastery + Dark Elves, or Sage Mastery + Halflings).
Sorry, that's a bunch of text complaining and I don't want to complain overall, I think the work you have done is amazing and I appreciate it! This is really just one small part that maybe could be re-rebalanced.
(August 13th, 2016, 18:26)Settemio Wrote: I think you may have gone too far in trying to balance Spell of Mastery. I have only played one game where I tried to go for it, so maybe I need to try again to get a better sense. But by the time I had no other spells to research, it still was going to take me over 200 turns to research, dumping all my raw mana into it.
...
Think about it the other way. Once the AI reached their very rare spells, you still have 200 turns to get yours, fight it out, and win the game. Which is usually the most fun and epic part of the game.
In quite a few extreme difficulty games, the AI was over 80% done researching it by the time I won.
I haven't yet been in a situation as a player when I needed it (which doesn't mean I didn't use it for fun or to save time - yes, hitting next turn 200 times is still faster than attacking 40 enemy cities one at a time), but I can easily imagine the AI having an unbeatable capital, if I get unlucky with my spells and end up without a very rare summoned unit for example.
It's mostly meant for large and huge maps where there are more resources and it takes a lot of time to win the normal way.
If looking at the numbers :
with 40 cities fully built, assuming each provides 20 research and 20 power (can be higher with elves, halfling etc or on rich maps with lots of crystals), you get 1600 RP from cities. You most likely control enough nodes for another few hundred and get some extra from spells or heroes, so I think 2000 is a reasonable estimate for an endgame research capacity on large games. That allows 500 RP per turn into research, SoM costs about 50-60000 to get, so 100 to 120 turns.
My personal experience is I usually have power in the range of 1200-2000 in big games near the end (before fighting the last opponent or during).
While 100 turns are a lot, if a direct war has a high risk of losing (for whatever reason), waiting it out can be better as long as you can maintain peace. Also, there are situations where you can hold your position but can't advance (enemy has defensive enchantments on their cities, a meteor storm kills your units on the way, the enemy has naval supremancy and you can't cross the ocean, etc) in which case SoM will be what you want to invest into.
Also SoM victory is a good option if you don't have a big empire and big armies but have a huge power income anyway (lots of nodes with astrologer, armageddon, dark ritual, life force, dwarves on a mineral rich land with everything transmuted into a crystal etc) and can at least defend what you have.
2.64
-When the AI decides on an intercontinental attack targeting land, it will not consider own land units in the stack even if the target is on land because the land units are not participating (the land attack system already takes care of that possibility earlier so it shouldn't attempt to here. Note that windwalking makes the entire stack count as nonland as it should)
-When the AI decides on an intercontinental attack targeting land, it requires at least 2x as powerful army as the target, otherwise no attack is done. This is to ensure 1-2s of incoming units do not attack a city with only a few guards then get killed by ranged attacks or spells. Instead they'll go to the gathering point to build a larger stack and attack through the land attack system (which is much more accurate with unit power)
-The AI now considers units up to 14 resistance a valid target for Syphon Life
-Runemaster and Aether Binding now increase the priority of using Dispel Magic or Dispelling Wave in combat by 50% instead of a flat +18 (which caused using the spell on a single enchantment way too often)
-The AI has a +10 priority for using Dispelling Wave and Dispel Magic in ocean battles (if the spell removed is one that enables water movement, the unit drowns so it's worth going for it more than usual)
-Fixed bug : Wraith Form unit enchantment is not recognized as a source of water movement when removing items from heroes.
-Fixed bug : Turning off water walking/flight causes the unit to drown even if it still has some sources of water movement. Also fixed Wraith Form never checking for drowning this way. Note that it won't drown units if wind walking is turned off. Fixing that would be far too difficult. Pointless, too, the units drown at end of turn as they should.
-AI engineers will now properly stop before entering enemy cities while building, instead of not building at all towards them. Note that they'll still move into enemy units on the way as well as into lairs and nodes, it's only for the destination tile.
I just realized the AI isn't attacking adjacent enemy units through the city walls despite that being legal and a pretty good move. I can't believe I didn't realize this is a problem earlier, I was so used to it. On the positive side, I don't think I would have been able to fix this without the other changes in 2.6x so it's perfect timing :D
Stuff like this is really hard to to, the wall prevents movement and no attacking can be done if movement is not valid...so the AI has to simulate a "hole" in the wall where the enemy units are to be able to attack...plus it has to be able to determine if the unit can be reached or not when selecting targets, otherwise it might try to target a unit that's too far from the wall to hit...which was done about 2 or 3 weeks ago, lucky. And with both of these, I still had to remove a part where it explicitly says if it's not a unit at the gate then do not move to attack things outside, and at the gate do not move at all, even if the enemy is already inside. Not sure about this latter part though, it's questionable. Should the unit at the gate move and attack if an enemy is coming from the side or behind the walls? if there are any idle units that cannot attack, they'll most likely replace it but if not, the gate stays empty...although that most likely means every unit inside was able to attack something, which might be worth it?
Attached is a save game which seems to crash the game 10% or so of the time during the AIs movements, it crashed during my normal play, then ive tried reloading it multiple times and just seems random, even if I try steps that I thought crashed it before. Just normal mom weirdness?
Have recorded a video of it crashing as well, so I can upload that.
(August 15th, 2016, 16:02)grazzy Wrote: Attached is a save game which seems to crash the game 10% or so of the time during the AIs movements, it crashed during my normal play, then ive tried reloading it multiple times and just seems random, even if I try steps that I thought crashed it before. Just normal mom weirdness?
Have recorded a video of it crashing as well, so I can upload that.
Before I check, are there roads in the battle and all units on your side flying? I've just found a battle a few hours ago that crashes 100% of the time but only if there are roads and I don't summon a walking unit. If I disable the roads it works fine. Also, the units are dwarves so the 1 movement might be related. (also, does the crash happen while the units move, or when the AI is (trying to) cast a spell?)
I assume there is a bug in the shortest path algorithm which I can't find related to tiles that cost half to move (maybe it generates an infinite path somehow?), so I'm trying to rewrite the whole thing and see if the bug disappears that way. Might be the same problem for you so I want to try fixing it on my side first.
Edit : the new shortest path finding works like a charm, my 100% crash combat is no longer crashing. Now let's see yours.
Its not in battle, there are roads and none of my units are flying. The video has been uploaded. Looks like the last AI unit moving is an engineer, so maybe that has something to do with it.
(August 15th, 2016, 17:10)grazzy Wrote: Its not in battle, there are roads and none of my units are flying. The video has been uploaded. Looks like the last AI unit moving is an engineer, so maybe that has something to do with it.
Oh, so it's an overland crash. Those are random most often, unless you manage to reproduce it consistently by reloading. (and by consistently I mean at least, like 30-40% of the time in the same 1 turn)
From the looks of it, green just finished their turn, so the crash is happening in blue's spellcasting phase. Except it doesn't, I tried hitting next turn about 7 times, got no crash.
It's either random or the spell causing it is chosen rarely by the AI. If the latter, eventually you'll get a file where it happens 100% the time (as the AI already started casting the spell the last turn but finishes on the next)
Edit : I tried moving all your units to the same position you had at the video and no crash. Since the AI's unit movement has no random elements in it (except settlers), I strongly doubt that is the source of the crash, it should be either spellcasting or random. (or maybe other stuff the AI does in the turn like continent ratings or taxes but those would crash a lot more often if having bugs I think)
(sorry Mal)
