Experimental version

[Seravy]

Experimental 4 is up!

Experimental 4
-Fixed bug : AI Plane Shifting and Astral Gate use checks for ocean wrong - might shift units into water or fail to shift.
-Fixed EXP3 bug - the number of starting commons was wrong for books 7-10.
-Sprites have been reverted to the non-experimental version
-AEther Binding now produces SP equal to the current turn instead of the casting skill, and the AI difficulty bonus for power income applies to it.

. If you can't teach the AI that for garrison defense, then I think you need to teach them to research better units faster.

Summons are at the top of the list for each tier in the new research procedure. That's as fast as it gets.

Other things the AI can rely on self-defense :

-Hiring heroes -we removed this from the early game but they'll still be there after turn 30+ so aside from some ultra early attack on the first wizard, this is working.
-Wall of Fire - Quite effective at countering early attacks.
-Heavenly Light - +1/+1 is a major boost on common units with stats in the 3-5 range.
-Fortress Lightning - Equal forces won't work, the player needs significantly superior forces to actually win.

I believe this part works fairly well for most wizards. Although, and this might be a problem, Nature uncommons are poor garrison units. Cockatrices and Spiders have a lower garrison rating than Sprites, and the new AI will do them in spiders->cockatrices->basilisks order, so it might take some time - leaving sprites the main garrison force for a while. Fortunately they are powerful enough to work, unless the player has Fire Storm.

the AI can afford to try to get better, faster.

The AI is not a human player. Aside from immediate needs, the AI distributes power based on personality traits and retorts. Don't ask for the AI to ignore its own personality and objective, that goes against the definition of what an AI player is. It's not a matter of being able to afford it or not.

[Nelphine]

Yes it does, that doesn't mean you can't generally up research across the board - generally all AI tend to be slow on research; and/or switch the order so that nature gets great lizards earlier. We're talking about change to avoid sprite abuse - that is going to have a lot of repercussions throughout the entire nature tree, which will affect other realms as well. I think it will be for the better.

[Seravy]

Yes it does, that doesn't mean you can't generally up research across the board - generally all AI tend to be slow on research; and/or switch the order so that nature gets great lizards earlier.  We're talking about change to avoid sprite abuse - that is going to have a lot of repercussions throughout the entire nature tree, which will affect other realms as well.  I think it will be for the better.

Early game resources are too scarce to research more specifically then, and summoning has a better return of investment through conquering neutrals, nodes, and protecting territory from the human player. If AI research is too low, the best way to improve that is to raise the research AI multiplier, but that has no effect in the early game because no research buildings are built yet. The idea of spending more on research is outright impossible to do on Normal and Hard difficulty - there is not enough to spend, it's barely enough to fuel the starting casting skill.

btw If Lizards come before Cockatrices, that gives Cockatrices a +200 priority penalty - it triggers "equal or higher tier summon already know, research later". Whether that's good or bad, idk, cockatrices are very hit or miss situational units...but they fly and have very good strategic combat value, better than lizards. So it's probably bad.

I believe we should postpone this discussion.
Until Catwalk presents the treasure data on Sprites and it proves a change is needed, we are just wasting time trying to solve a very hard problem. I rather focus time and resources on other issues - Sprites have already taken away weeks with no actual progress on them.

We should start thinking about retorts instead, specifically Famous (too weak) and Inquisitor (unbalances early game AND too weak later).

Edit : opened a thread for these two retorts here : http://www.realmsbeyond.net/forums/showt...p?tid=8740

[Catwalk]

I think it's time I quietly back out of this project. I think that you, Seravy, lack the game competence to make good balance decisions. You're great at coding and ideas, but you're not a good decision maker. You are also not good at absorbing input, and your priorities also seem to be worlds apart from my own. I honestly don't understand where you're coming from on Sprites, and I probably won't. By your own admission, you haven't even tried out the Sprite strategy. It's trivially easy to solve, you're seeing problems where none exist. You put pet interests above game balance, and that's a big problem. In order to fix the problem with an incredibly broken Sprite strategy (I can demonstrate 10 wins in 10 Impossible games where I win with Sprites), you prefer fixing everything but Sprites. But the problem is Sprites. The above comment about refusing to improve the AI because it has a certain personality trait is also mind boggling. Personality should add flavour and variety to the AI, not handicap it.

With that said, I'll refrain from further comments on the mod, and wish you good luck in achieving your goals with it. I'll probably keep posting game reports once in a while, as I do find the mod entertaining in spite of its shortcomings. And I'll admit that I'm still having fun with Sprites, I'm still getting better at the timing of the individual steps in it.

[Seravy]

Personality should add flavour and variety to the AI, not handicap it.

Adding flavor and variety IS a handicap by definition. It means you aren't doing the best strategy all the time - doing so has neither flavor nor variety, it's repetitive, boring, and effective.

Does this mean you are not going to provide the data on sprite treasure hunting? Not that it matters if I don't know how to solve the problem right now but knowing if there is or isn't one should have been the first thing to decide. Because, unlike you, I'm not 100% sure there is a problem with sprites. As far as I remember every test we did used some overpowered retort combination to accelerate it.
I made this new version because I thought if it works better there is no need to spend time thinking about "overpowered or not" anymore, but that plan failed.
If you aren't providing the data, I'll have to get it on my own.

By your own admission, you haven't even tried out the Sprite strategy.

This, I don't understand. I spent like an entire week testing nothing but that as far as I remember. And all I found out was how a whole bunch of retorts break the game. I admit I haven't tried Sprites without retorts in the past half year...neither did you...didn't we agree that you calculate the data instead as that's more effective than manually playing dozens of games?
If you mean I haven't tried the NEW sprites vs lairs, yes, true. You said the change didn't help and I accepted that. The new unit was so ridiculously powerful against wizards that lairs didn't even matter.

[Nelphine]

The problem is you've already stated you are OK with where most spellcasting retorts are (possibly excepting conjuror). If retorts are OK, and retorts + sprites wins, then sprites are the problem.

But realistically, spellweaver is too strong (far more than twice as good as any other spell retort including channeler) and specialist is too good comoared to 1 more book, no matter what book number it is and spellweaver + specialist + sprites is all you need to make sprites insane. Conjuror makes it much worse, but is overkill.

[Seravy]

The problem is you've already stated you are OK with where most spellcasting retorts are (possibly excepting conjuror). If retorts are OK, and retorts + sprites wins, then sprites are the problem.

I'm fine with the individual effects of retorts. The interaction when stacking 4 of them is a completely different problem. We've already dealt with Archmage+Spellweaver.

I expect specialist to be less of a threat - you need 9+ books to get any cost reduction from books to stack with it. If it's still too good, we can change it to a 12% cost reduction, I believe this has been discussed a few months ago but back then we pretty much left it at "doesn't really matter, it's not overpowered".

Conjurer is left to test and possibly change. My preferred solution, if necessary is to make it mutually exclusive with the archmage/spellweaver, but I'm not entirely sure there is a need yet.

Is there any other retort left that boosts summoning that I missed?

And back to Sprites, they haven't been a problem 3 months ago. So the problem might as well be the Earth Lore change and/or the lair monster distribution change. If nothing else works and both sprites and retorts are fine on their own, we even have crude solutions available such as "the cost of this spell cannot be modified".

I still don't understand your problem with Spellweaver. (wasn't it you who asked for 33% mana instead of 25%? Or someone else? I guess someone else...)

Archmage/ Spellweaver
+28% Overland skill (1/3 picks) / +50% overland skill (2/3 picks)
+28% combat Skill (1/3 picks) / None
+10 starting skill (1/3 picks) / None
None/ +33% power income (unknown, I'd say about 1 1/3 picks sounds about right.)

[Nelphine]

You may recall, what I said was that mana focusing was too weak at 25% because mana generation isn't terribly important; I also said cult master was worse than mana focusing despite being a power multiplier instead of mana. As a note, with eaely rush tactics like sprites, I've been shown very good reasons why even mana generation can be very powerful.

However spellweaver increases power generation, not mana. This is much stronger. (Power is about 25% better than skill which is about 20% better than research and mana).

The old astrologer was limited in that its bonus only applied between turns. This made it so that ideally you always wanted to be casting between turns, making spellblast a real problem. Additionally, the other things you got from astrologer were very niche benefits. So effectively you got 50% more overland skill, with some minor benefits and a minor but noticeable drawback for 2 picks.

Now you get true 50% overland skill, and 33% power generation, all for 2 picks. Considering I consider overland skill to be far more important than combat skill (past 60-100 skill anyway), and I already considered archmage better than channeler, and the result is spellweaver is incredible

[zitro1987]

I'm not concerned with spellweaver's power. It tends to match my personal idea of fun (more overland skill in late game) without appearing to be stronger than stacking two 1-pick retorts. I think I was the one that requested 33% power.

I'm getting repetitive, but this retort vs book argument can still be mostly resolved by adding more generic benefits (initial or percentage power) after the 7th or 8th book (regardless of mono-multi realm)
*you suffer if you have few books (few spells, weak intro power/skill, fewer initial spells, risk of not getting spells)
*you suffer if you have few 2 or less retorts (too obvious to explain)
*therefore, by pushing stronger benefits by choose more books over retorts, players have more choices, making it more fun.

[Nelphine]

Anything in mind zitro? While I realize you've been saying it, I think my problem is I'm not sure what you really mean, so it just slips away from my thoughts when I'm trying to work something out.