Armor system overhaul

[tmewett]

Currently armor is mostly a linear progression, with the only caveat to upgrading armor being increased stealth radius. Also, it is usually recommended to enchant weapons over armor when doing a melee build.

Two ways to make this more interesting that we discussed:

• [ ] Split armor class system into hit probability and damage absorption
  • Allows more melee hybrid builds, because in some cases you will be able to deal small amounts of damage with an underpowered weapon, where now you would struggle to hit anything
  • Different monsters have different weights - big and slow are easy to hit but absorb more, small and quick are hard to hit but take a lot of damage
• [ ] Rework armor types to give unique benefits. Some ideas for this:
  • The player would start with no armor -- would require some enemy re-balancing in early depths.
  • Types of armor:
    • Leather armor (stealth): base strength 10, lower stealth range by 1 for each enchant over zero
    • Mirrored plate (reflection): base strength 12, increases reflection chance for each enchant over zero (similar to reflection runic)
    • Scale mail (standard): base strength 14, increase armor rating by 1.5x the enchantment value instead of 1x for other types of armor
    • Thorned splint armor (reprisal): base strength 15, causes reprisal damage like a runic of reprisal
    • Padded ring mail (absorption): base strength 16, reduce damage from attacks by a random amount between 0 and the enchantment level like absorption runic
    • Base armor values would scale with strength reqs along the same curve as current armor, with scale mail starting with a couple of bonus points.
    • Then we’d remove runics of reflection, reprisal and absorption, and rings of stealth.

[SanatMishra]

Some thoughts I've had since the meeting:

• I'm not sure if absorption would make lighter unenchanted weapons more useful. Any appreciable amount of absorption for an enchanted broadsword would make a flail or war pike much worse in comparison.
• If we give every enemy some amount of absorption, it would have a pretty similar effect to giving them extra HP but with added complexity. Cutting a monster's HP to give it high absorption would make it weak to magic, and vice versa, but going too far on this spectrum would make certain monsters like revenants for weaker weapons.
• It would be cool if the we balanced the armor types to be pretty similar in overall value when enchanted, so that you have choices when picking an armor to enchant. Maybe chain mail gets a few extra base armor points in return for not having an effect?

[tmewett]

(3) - that is one of the ideas above: scale mail gains more armor points per enchant

(1) - the idea is that the expected damage for each weapon vs each creature will be largely the same as it is now. Whereas now weaker weapons will struggle to hit a well-armored creature, but hit fully when they do, with this change they will hit more often but less per-hit. The result is you're less likely to say miss 3 in a row and do 0 dmg but the weapon DPS is similar long-term

[SanatMishra]

I see your point with scale mail.

What I'm saying in the first bullet is that stronger weapons lose less of their damage than weaker weapons on the same absorption. If a monster has an average of 5 absorption, flail average damage goes down from 11.5 to 6.5 (43%), but broadsword average damage goes down from 18 to 13 (28%), or even less if it's enchanted.

[tmewett]

Oh, I was thinking about absorption being proportional to incoming damage. % damage resistance, essentially

[HomebrewHomunculus]

Apologies for the long post that follows - just trying to get a feel of what problems and solutions are on the table here, seeing as damage reduction sounds like quite a sweeping change in mechanics. Please feel free to rein me in if this goes too far off topic, or if this isn't the correct forum for this level of discussion!

---

Currently armor is mostly a linear progression, with the only caveat to upgrading armor being increased stealth radius.

• Personally, I'm not sure whether linearity or simplicity is necessarily a flaw in itself. Likely you weren't implying tanything of the sort, but I think it's still worth singing the praises of simplicity, seeing as there are so many maximalist games out there, and Brogue being one of the few minimalist ones. So, before discussing armor mechanics, I'd like to preface by saying: different games model the effects of armor to different degrees of detail, strategic depth, and realism-simulationism, depending on the game's aims. Brogue benefits from using a simple model - it's part of its identity, in my opinion.

• Just as historical trivia, Dungeons & Dragons itself has toyed around with damage reduction (DR) across the editions, but I think it has always been found lacking and is not used much presently. The current iteration is still a simple "both armor and evasion make you harder to hit, bigger weapons deal more damage" model.

• Of course, D&D must use flat integer amounts for damage reduction, due to using humans for calculation. Brogue, thanks to the computer, has the added option of using percentages ("damage resistance"), as you've already mentioned. That makes it perhaps easier to boost the power of low-damage weapons: while a flat DR threatens to overcome the weapon's damage entirely, a porportional DR means that low-damage weapons suffer less from it, in terms of absolute damage, than bigger ones.

• Nevertheless, any equation that combines damage reduction and damage avoidance will be far more difficult to balance than one which involves only the latter, I expect.

• Player expectations. One thing to watch out for, when adding layers of detail to an armor system: it might send the signal that this is an attempt to model some realistic aspects of combat. It may lead to only further debates over which aspects of medieval combat the system fails to model well - such discussions are common on tabletop RPG forums. Having a more intentionally "cartoony" system sends a clear signal that there is no attempt to do such a thing, and invites less comparisons to reality.

---

Also, it is usually recommended to enchant weapons over armor when doing a melee build.

Allows more melee hybrid builds, because in some cases you will be able to deal small amounts of damage with an underpowered weapon, where now you would struggle to hit anything

I like this this line of thinking - attacking the specific pain points that come up during play could be a fruitful direction to take the discussion in. So, to clarify things, one important question:

What is the design problem that the present discussion attempts to solve?

Is it that:

1. Armor is uninteresting and lacks strategical variety
2. Unenchanted weapons are too weak, and therefore enchanting weapons before armor is a strategic no-brainer. (Or, equivalently: the benefits of enchanting weapons are too strong relative to other options.)
3. Unenchanted weapons will tend to miss almost all the time, which slows down gameplay (the expected number of attacks-to-kill grows out of control).

(Obviously, I'm not aware of what was discussed in the meeting mentioned above, but from this thread it sounds like all of the above are included in the mix. Hopefully I'm not misinterpreting too much, as it seems I may be going off incomplete information here.)

I think 2 & 3 are related problems, but 1 is largely a separate one. Any solution to the armor problem should be considered independently of the weapon problem, at least initially. That is my intuition because the weapon problems are likely to find a shared solution having to do with the combat algorithm or enchantment system. In contrast, armor variety/strategy seems like something that can be solved purely within equipment design, without necessarily any changes to combat algorithms, since monsters don't use equipment but do use the same combat algorithm as the player.

There are many possible options, so I'm just going to throw out a bunch. Since I'm new here, please forgive me if these have been discussed before - I'm mostly going off from reading the wiki and a small part of the source code.

Problem 3, where monster survivability grows effectively without bounds when the attacker has low accuracy, is something that D&D tries to handle with "automatic hits" - the "natural 20 always hits" rule, which was included far before doubling of damage from critical hits. It acts as a backstop for low accuracies, guaranteeing a 5% rate to hit when monster defense is very high. Of course, Brogue doesn't use such a linear accuracy, but rather a curve which guarantees that the chance to hit never goes to 0%, in practice.

The curved to-hit progression of Brogue means that enchanting a weapon gives compounding returns. If I'm reading the formula right (from the wiki), your chance to hit a troll (defense 70) is:

((0.987^70)*(1.065^(net enchantment))

where "net enchantment" is (weapon's enchantment level + any strength bonus/penalty). Thus, increasing your weapon's enchantment level (ignoring strength bonuses for the moment) from +0 -> +5 -> +10 takes your chance to hit from 40% -> 55% -> 75%. The odds to hit improve first by 15% for the first five levels and then by 20% more for the next five - the opposite of diminshing returns. The only diminishing returns you will see is when your weapon is so highly enchanted that it already has at 100% chance to hit your intended target. (For example, ogres have defense 60, so at 12.50 net enchanment you have a 100% hit rate, so there's no gain from further enchantments. For trolls, this cap is at 14.75 net enchantment.)

This scaling is further compounded by the fact that enchanting equipment decreases its Strength requirement. You get an accuracy boost not just from the +1 on the weapon, but from the increased Strength bonus (or reduced Strength penalty). So, it's actually +1.25 net enchantment gained per scroll of enchanting (assuming you meet the Strength requirement).

• One brute option is to simply reduce enemy defenses across the board. This will make less enchanted weapons hit more often, and highly enchanted ones reach the 100% hit rate sooner. Overall enemy survivability will be lower, which can be counteracted by increasing HPs across the board by some number that results in a similar turns-to-kill number as currently. However, instead of disincentivizing pumping weapons full of enchantments, taking this option might actually incentivize players to enchant their weapon quickly, up to a 100% hit rate, before beginning to enchant other items. It also doesn't remove the exponential scaling of accuracy with net enchantment.

• Another option is to reduce the base value 1.065 which determines chance to hit. This, likewise, will not remove the exponential scaling, but does make the growth slightly slower.

• Yet another option is to make the weapon's "nominal enchantment level" (the "+X") map to the "net enchantment" (the exponent of 1.065) in some way that is not linear, but instead has diminished returns. Just to give an example, assuming a scaling exponent of 0.9, your net enchantment (ignoring Strength bonuses) would be (the "+X" of the weapon)^0.9. So a +5 weapon has (5^0.9) = +4.26 effectiveness, and a +10 weapon has +7.94. Thus:

accuracy = 1.065^(net enchant) = 1.065^((nominal enchant)^0.9))

which, for +0 -> +5 -> +10 weapons, would be 100% -> 131% -> 165%. Using our troll example again, the odds of hitting are 40% -> 52% -> 66%. Or a gain of 12% and 14%, as opposed to our earlier gains of 15% and 20%.

• One more possibility to mention, though perhaps a very radical one, would be to change not how much enchanting affects weapons but rather what it affects. What if it it increased the weapon's max damage, instead of its accuracy? (Unless I'm mistaken, weapon damage is fixed by type and can currently not be boosted by enchants or anything else.) The average damage would increase by a flat amount per enchant - but even if you chose to not enchant your weapon, you could still at least land nearly the same number of hits on monsters as if you did. They'd just be weaker hits.

[tmewett]

Thanks for the great comment.

You're right that we are trying to solve two issues here - they are split into two lists in the original, but grouped under armor as we initially expected armor would be the way to solve them. The way I see it, we're focusing on (1) and (3) in your points.

Point (2) is far less of a problem imo, hence I never thought much about your suggestion of adjusting the enchantment curve of weapons. Lowering defense and boosting HP seems very similar to the % damage reduction idea, except is a larger change, seeing as HP affects all types of damage but armor DR would affect melee combat only. Making weapons scale less strongly i suppose would even out the differences between high- and low-enchanted weapons, which would devalue enchanting them - however if the game mechanics don't support this well (the current hypothesis) then i don't see this being a big win.

The fixed minimum hit chance given by a d20 crit is an interesting solution for (3), but I fear 5% is still too low to make a substantial difference here. But maybe not.

I don't really have much more analysis on this right now - I'm more interested in experimentation, with DR or otherwise, to at least see something real. And sorry that I didn't reply to all your points, I think they're all decent possibilities; you've definitely got the right idea of the problem.

[tmewett]

The most glaring problem I see is how hit effects (acid, weakening, knockback) should work. Absorption is a pretty big nerf if they work as normal.

[RojjaCebolla]

Reprisal and absorption are funny ones to scale alongside armor value. Armor prevents hits. Reprisal and absorption trigger on hits.

Reprisal triggering when they miss is much more likely to be worthy to enchant. And that works thematically!

Absorption just isn't good, as far as i know. Maybe larger numbers, or a side effect, could do it.

Reflection is more valuable when your stealth is bad - it should probably be on a heavy armor. (unless we are decoupling armor weight from stealth?) It's not scale/chain wearers who are getting sniped by spells from across a level-wide chasm.

Aside from these bikeshedding nits, i do think the armor rework sounds interesting and good. A nice counterpart for attack patterns.