Tutorial:Unit Balance Tips

From Total War Modding
Revision as of 16:14, 7 February 2021 by Vandy (talk | contribs)

Originally by Cataph

PREMISE

This guide is aimed at unit or character modders in WH2, and while I assume we’re talking about introducing or editing stuff for vanilla balance, most of the tips should remain useful even if you’re doing your own battle overhaul or whatever it is you’re doing. What I intend to do is suggest ways towards tuning your stuff so that it fits in the whole shebang and avoiding the common mistakes.

I’m sure that many of you may be thinking “but balance is subjective!”, which is possibly true, but it’s especially true only because balance is fickle and does depend on context a lot. On the other hand, there are scenarios in which what you’re doing is objectively too weak or too strong.

There is also the spiky and infamous matter of SP or MP balance.

I agree that MP poses issues that SP doesn’t, but I personally disagree with those that see the two aspects of the game as incompatible. You can balance a unit for both, give or take. Of course, MP’s tendency towards min-maxing and cheesing will expose issues faster than a Karen in a retail shop.

Then there’s the famous “all units become OP in campaign” which is kinda true but sort of an excuse, and out of our hands insofar as we do our job balancing the unit’s baseline stats.

“But I want unbalanced OP units!” Sure, then read no further and we won’t talk about it.

Keep in mind this guide will avoid technicalities of how to create a unit, that’s for the unit creation guides to do.

CONTEXT: THE ROSTER

Alrighty, you’re making a new unit. One of the main features of Warhammer is that its balance is asymmetrical, in that there are many diverse factions and most of them have their own strengths and weaknesses, usually much more marked than they’d be in a historical setting of human-only warfare.

We want to preserve those factional strengths and weaknesses; the usual examples being how Warriors of Chaos are a meatheaded faction poised for a decisive melee push with expensive and elite melee units, with little in the way of missile potential, and predictably fielding heavy armour stuff. Dwarfs are a tanky and shooty faction famously suited for corner camping because of how lacky they are in terms of mobility. The Empire is a faction built for width rather than height, with only a couple units that can shine in their department. Wood Elves are supposed to be the kings of squishy skirmishers. And so on.

So while it may be your active choice to, say, add a unit comparable to Chaos Chosen to Wood Elves, and assuming you’re still reading because you still somehow want to balance them, know that you’re removing a faction’s weakness, specifically the lack of a uber heavy infantry unit.

Relatedly, you could be tempted to exactly copy the stats of a unit to add one to another faction. Hey, they can’t be OP if they’re the same as vanilla unit X, right? Wrong, some units may be deliberately stronger for a faction exactly because that’s where that faction is supposed to shine and to compensate other weaknesses. In other words, pooling good units from multiple factions into one still creates an OP chimera.

This also works from another perspective. If you’re designing, say, a heavy cavalry unit for whatever non-elven non-chaos faction, you should likely avoid making it more powerful than what the supposed kings of heavy cav can bring, aka Bretonnia. If your Halfling Rooster Knights can keep up with Grail Knights there’s definitely something wrong.

Which brings us to…

CONTEXT: THE METERSTICK

Which is like a yardstick, but we don’t do that sort of measure in these parts.

A healthy habit I found is identifying possibly two or even up to four vanilla units that you can use to delimit what your unit should be. No, vanilla balance is not perfect, and gods know I have my grievances, but that’s what we’re working with.

Let’s continue the example and say your unit, hereby The Fonzies, is to be a unit of heavy infantry with greatswords. Daring, aren’t we? And what faction are they for? Dark Elves, I decide that the Fonzies are Dark Elves.

Ok, what would your metersticks for the Fonzies be? Obviously, more heavy infantry with great weapons. Let’s look around, and we’ll see a lot of examples; we can do that either by browsing on the Custom Battles rosters and/or on https://twwstats.com/

First, we look at units from the same race: Executioners, White Lions, Swordmasters, Wildwood Rangers. Then, from other races, we have Greatswords, Foot Squires, GW Longbeards. Ok, that’s a wild range, innit?

Let’s narrow it down a bit and decide that the unit will be more or less for a tier 3 slot of sorts, so we definitely don’t want it to be remotely as good as the Executioners or the Swordmasters. Lions and WW Rangers are a good meterstick, since they are tier 3 elves with a big choppa, but both have their own peculiarities: one with pelt shenanigans plus martial prowess (BIG +12 MD while healthy), one with immpsy, a custom passive and beefier stats. A Delf unit would come with its own murderous prowess, and might be more aggressive in overall style.

When we proceed to juggling stats, we want to keep an eye on the human units, to make sure we don’t go as low as that, that we don’t increase unit size, since elves cap at 100@ultra for non-chaff, that we don’t make the Fonzies as tough as the Rangers or the Longbeards. Again, we don’t want to make Executioners look bad, since they have to coexist. So, for example, we can play on less armour, more speed, squishy stats. The point here is making sure the Fonzies don’t make other units obsolete. And since Dark Elves already have killy anti-infantry units, we should let Corsairs and Witch Elves be more efficient against lightly-armoured foes, focusing more on the Fonzies as a contemporary source of armour-piercing (perhaps with no Bonus vs Infantry), with a bit of armour.

CONTEXT: THE TABLETOP

Woooh, what a can of worms. However, if you know how to read tabletop stats, the 8E TT was better balanced than people usually give it credit for. That’s not to say perfect by any far-fetched means, but good enough for the purpose: whether your Fonzies had tabletop stats or you’d have to imagine them, TT stats can provide a further meterstick for you, or provide information and inspiration for abilities and stuff.

In other words, similarly to the conventional meterstick approach, you can also apply the same logic to the metersticks’ statlines and make sure you’re not accidentally down/overstepping. This is particularly important when you’re designing units for a race that has very definite limits, like humans or even dwarfs. A non-character human can’t be as good as a Swordmaster of Hoeth, ever, a Dwarf is slow, and elves will be frailer than most other things.

THE CUSTOM BATTLE TEST

This is the obvious place to test the Fonzies in terms of actual performance. A staple is testing for 1v1 duels. Lords will usually interfere so you want to get a “all units are generals” mod or do it yourself so you can get rid of them. Get a nice flat map and duke it out with the metersticks AND with other units. Keep in mind that the post-battle result will heal not-dead entities, note down the results like kills, losses and hitpoints at battle-end, then possibly tweak. Possibly run at least 3 tests per duel because there is always a RNG component that will cause deviations. This is particularly true for characters, where a single blow can change the course of the fight and there aren’t tens of entities flattening the statistical variation of those hits.

Keep in mind that the vacuum of a 1v1 unit duel is not how battles are going to go, you’re just testing sheer statistics devoid of the context in which the unit may have to be used. For example, you usually want a glasscannon unit to flank a battle line (which you can still simulate of course), and will probably underperform in a frontal assault. What you want to know is whether the glasscannon Fonzies are STILL rolling over the competition even in an unsuitable context, which would be bad.

Also, if your unit is for example very defensive and perhaps even with Charge Defence vs All, don’t charge with it because you’re going to waste one of its main benefits. Or if your unit is anti-large, it’s ok for it to suffer a bit against good infantry. And so on.

THE AI

If possible, check what the AI does with the Fonzies if they happen to be a weird unit. It may not know what to do with it, which may not be something that you can fix, but keep in mind that if a unit is particularly strong in the hands of the player, it should also be not completely useless in the hands of the AI. This typically happens on very polarized units or just very powerful ones. You may know that you need to focus fire the Big Nasty Emperor Dragon Of Doom with its 25k hitpoints, but the AI doesn’t and will allow it to wreck the entire army.

Excursus: I’ve seen this a lot. “Unit is very strong but if you focus fire it with five armour-piercing missile units it goes down” is not an excuse. Everything dies if you shoot it with enough firepower. Nerf it or accept that it’s a cheat unit.

THE PERNICIOUS STATS

Like I said, I’m not going to talk about how the single tables work, but I’ll point out some individual stats that are often underestimated.

  • Hitpoints: CA heavily uses these to make a unit jump from a tier to another. This is a major performance tool.
  • Unit size: on top of not going above the usual unit size for that race, keep in mind that you can manipulate it to better balance your Fonzies. For example you might have an infantry unit that overperforms no matter what you sensibly do while at 120 size, so you might want to downgrade them to 100 and see what happens. Higher numbers expose the unit to more missiles and more damage in general, adds damage soaking potential on the same hitpoints, adds damage dishing potential because more entities may be dealing that oomph, adds entities that will deal that charge bonus. Smaller numbers usually demand better stats but are nullified if you’re boosting both hp and damage. Unit size is particularly important for missile units, as each entity will fire and become an active multiplier of its missile damage.
  • Unit spacing: this is a subtle one. A unit of high elves and one of norscan marauders will behave differently in terms of how many entities get into contact, how many overlap beyond the flanks, how many will get hit and how easily when shot at.
  • Damage: even shifting one point of base damage towards ap can change the outcome of a 1v1 test.
  • Bonus vs X and Charge bonus: keep in mind that both will add that number to both Melee Attack and Damage (in the same base/ap ratio). It’s a huge thing. Charge bonus only lasts for a few seconds, but its impact can be huge especially when there is a lot of entities on the first ranks. You could increase these stats and reduce MA and damage to make a unit shine more in a very specific time or target frame.
  • Speed: on top of mattering for collision impacts and mass comparisons, the simple truth of speed is that you probably want to avoid having a unit that is both relatively fast and has three tons of armour on it.
  • Attributes: they can be cool, but nobody is cool anymore when everybody has all the attributes, so use them with moderation and understand what they do. I’ve seen many that don’t know what’s the actual effect of Causes Fear (-8 Leadership aura, unless target also causes it or is immpsy), or that make a unit both unbreakable and immpsy, which is useless. Actually, normally just pretend Unbreakable and Perfect Vigour don’t even exist.
  • Cost: costs are tricky, because there’s basically no chance that two different units priced at about 700 will be exactly worth the same in all prospects. You might be satisfied with pricing the Fonzies at 800ish due to their performance, and then a 700-cost anti-infantry unit comes and completely wrecks them. It happens. Just try your best. More importantly, keep in mind that there is no cost high enough to justify a unit from single-handedly nuking half the battlefield.
  • Animations: this is actually quite a painful note. Many unit attacks are quite animation-driven in WH and will need special attention. Typicall it’s chariots, each dealing damage according to its parameters in collision attacks, charge impacts, draught mounts and crew members, which for the moment I’ll spare you. Sometimes it’s slightly straightforward and an animation will only support so much reload speed or so much splash attack size. Sometimes you have Nakai hitting like thrice during the same charge animation. In practice, you don’t have a foolproof way of telling how something like a big monster or chariot unit will behave before testing.

THE CAMPAIGN

Like we premised and like it’s well-known, campaign-side tends to eventually bring enough buffs to make even a gobbo a solid unit. More or less. However, if you’ve fine-tuned the Fonzies well enough to be decent in custom battle and in respect of the points above, chances are you did good and the unit will grow similarly to others once junctioned to vanilla unit set buffs or made your own to be similar.

Of course, if you are making your own campaign effects to be +100 armour for Fonzie units or some such, it might be the case of putting that back into perspective.

Where you place the Fonzies for recruitment also matters a lot. Quite obviously, there is a huge difference between putting a good unit on the main settlement chain, on the barracks chain, or on an isolated building that does nothing else. Especially when you’re adding very powerful or otherwise game-changing stuff, it might be a good idea to make sure the owner has to go out of their way to get it.

Relatedly, there’s tier, which is very important with CA balance, and sometimes you can see a more advanced unit completely replacing an earlier one because it’s simply better. Some people are fine with that, I personally despise it, but it’s a thing. My own recommendation would be to try and maintain, if at all possible, a tradeoff if the Fonzies are a virtual upgrade/downgrade to something else. Basically, avoid the Longbeard syndrome.