New difficulty level?

[Seravy]

I've been thinking and I believe there should be one more difficulty level.

Currently we have :

Easy - The player receives some advantages and the AI is handicapped. No challenge, even for first time players, as long as they know the basics about how to play a strategy game.
Normal - Minimal AI advantage. No advantages for the player and the AI is not holding back in basic gameplay, but can't use some of the nastiest tricks available. No myrran wizard, and no random AI wizards, everyone is default.
Hard - Medium AI advantage. The AI is not holding back at all, except for one or two particularly evil tactics that abuse even higher AI resource advantages (stasis spam). AI wizards can be custom, and there is a Myrran wizard. Very challenging for average players and mildly challenging for veterans playing weaker strategies or random wizards.
Extreme -Large AI advantage, AI goes all out on strategies. Unforgiving difficulty level where only strong strategies can win and large mistakes result in a loss.
Impossible -As above but even harder, on top of the best strategies and perfect, flawless gameplay, you also need luck to win.


So...what's missing?
Well, first of all, there is no difficulty level that's like "Normal" but has random enemy wizards and myrran wizards and offers a fair challenge for average players. Hard has that all but already gives a lot of bonus to AI, is probably too difficult for what it is meant to be.
Second, there is no "convenient, for fun" difficulty level for veteran players. Hard is too easy - unless you go for a particularly crappy setup, play poorly and have bad luck, you'll overwhelm the game. Extreme is kinda too difficult - I've been losing more than half my games on it when playing average wizards - not weak ones, but not the "top tier" impossible builds either. There should be a difficulty where average wizards will result in an interesting game and I don't think that exists at the moment. Looking at the raw numbers, extreme gets twice as much bonus as hard which is a lot.

So...this is what I think might be ideal instead :

Novice - Same as easy
Mage - Same as Normal
Wizard - Random wizards, Myrran wizard, but AI resources in between Normal and old Hard.
Lord - Like current Extreme but resource bonus between it and Hard.
Lunatic - Current Extreme.
Master - Current Impossible.

(other suggestions for naming levels are welcome)

If we decide to do this, we'll need to look at each individual detail difficulty level affects while doing it.
See http://www.realmsbeyond.net/forums/showt...Difficulty for more (albeit probably outdated on resources) information on the effect of existing levels.

[MrBiscuits]

Well the difficulty level between hard and extreme would be great for me since I can win almost every time on hard, but struggle a lot on extreme. The jump seems massive, although looking at the bonuses they don't seem as different as I'd imagined.

Anyway for me on extreme the game seems to get bogged down into wars of attrition late on which I inevitably lose due to the AIs large bonuses and it gets frustrating. I can only do well with barbarians regularly.

As for naming how about using the unit experience level names?

Recruit, regular, veteran, elite, ultra-elite and the other one you get with crusade + warlord which I've forgotten.

[Seravy]

Unit levels...don't sound very wizardly, otherwise nice idea.
Also, shorter words are preferred, as we need to fit 6 words in the space that had 5 before. (and the longest one determines the space needed or all of them, unless it's in the last slot.)
Ultra-Elite would be 11 characters so that's not very good for this.

[MrBiscuits]

Unit levels...don't sound very wizardly, otherwise nice idea.
Also, shorter words are preferred, as we need to fit 6 words in the space that had 5 before. (and the longest one determines the space needed or all of them, unless it's in the last slot.)
Ultra-Elite would be 11 characters so that's not very good for this.

Oh right, perhaps you could use some of the hero levels instead?

Or just names of levels of a wizard such as:

Apprentice (or novice if too long)
Mage
Conjurer
Sorcerer
Warlock
Arch-Mage

[Seravy]

Here is the detailed information on what all the effects of the current difficulty levels are (first 5 columns) and what I think would be a good idea for the new system (the 6 columns behind it)
   

[MrBiscuits]

Are you not putting casting and upkeep between the two other levels? Could have casting as 55 and upkeep as 45.

Looking at that I can see why impossible is called impossible.

[Nelphine]

I'd be inclined to drop the wizards power bonus (150), production bonus (125), the lords production bonus (140) and increase the lords casting multiplier (55-60). I'd also drop the chance for specialist if mono realm on wizard. Custom for the first time doesn't need the specialist.

I'd also call extreme 'arch-mage', and I'd call impossible 'lich'

[Seravy]

Power bonus - I considered it but the AI is still and always vulnerable for excessive and wasteful mana crystal use in combat - any decent difficulty level needs a high modifier there unfortunately.

Production bonus - Maybe, but feels a bit risky, as production is "THE" snowballing stat - a small change has massive effects on all categories.  As this is the stat with the highest effect, and we already think the existing gap (25%) is too drastically high between extreme and hard, I tried to make sure the steps are smaller (18 and 12) and the new level is exactly halfway between the old hard and extreme.

Casting bonus - the lower power, production and population settings already mean the AI will have somewhat less casting skill, and weaker spells. If we also change the cost modifier, the new level will be way to close to the current "Hard" instead of being a somewhat more playable version of "Extreme". Also, magic is a qualitative difference - overland casting cost is the main determining factor for doomstack quality (and frequency of one showing up), which I'm satisfied with on the current extreme - what I'm not happy about is the crazy growth they get while I'm busy dealing with those doomstacks, and the excessive amount of medicore stacks they send which keeps all my armies busy and wins through attrition. In short, I want the difference between Wizard and Lord to be a "Quality" change, more than a "Quantity" one, while "Quantity" should be for the next two levels.

Specialist - Mono realm wizards have a lower chance to pick their "primary" retort, instead they also have a chance for Specialist. I don't see how Specialist for mono is any more valuable than the "primary" retort of either mono or dual wizards, which are 50-75% and don't depend on difficulty.

A few other rows I'd like to mention :

Diplomacy - I generally tried to change the modifiers to be a bit less harsh on high levels, especially in areas that result in all the AI allying together and/or declaring war on the human player without a good reason. I feel the AI is playing well enough now to be a threat even if not constantly at war - or more like the game being near unplayable if the player can't avoid 3-4 way wars.

Not sure about the max size for neutral cities - or if it even matters, as they are unlikely to stay neutral long enough to grow to larger sizes on higher difficulty levels.

Also not sure about raiders and monsters - hard to change the formula there due to size issues, but leaving it as is would mean the amount and frequency of them increasing on what becomes of extreme and hard. Ideally I would want these to be as is now, except "Wizard" and "Lord" being both a copy of hard, but that's hard to do - so the table shows what it becomes if I leave the formula unchanged.

[Nelphine]

*shrug* I consider casting skill to be the defining factor for AI difficulty. I already think overland casting modifiers are too low across the board. I'd like to see imposdible/master/lich be at 50%, and then base everything around there. Disjunction and globals are just too AI centric with the overland casting bonuses.

Production bonus is the primary snowball bonus, I agree. But its also the primary factor in tons of crap happening while you deal with doomstacks (which comes from power and overland casting).

Power production: AI already get thousands of mana extremely easily - they still produce far more than that, and just convert to gold to buy everything. And despite that, you can drain an impossible AI of mana even without doing it on purpose, just from their own attacks against you. Your power bonus simply doesn't matter when you're using 10000 mana in 3 turns.

Which means power just leads to a bigger snowball effect in the beginning. Which causes the most problems against newer players who aren't expecting that much snowballing that early.

[Seravy]

I don't think the AI will turn mana into gold.
If it has enough mana, it'll stop producing it and spend it on skill and research instead. Those ensure the AI stays a relevant threat in magic, even if the other, economic values are lower.

I haven't played enough Impossible to be able to tell if the 40 casting multiplier is too low or not. Having it 50% on both Impossible and the level(s) below is also an option. I believe the massive production, power, gold, etc bonus ensures the AI casts more overland spells through raising casting skill faster, so there might not be a need to add a further bonus through the casting multiplier.