Some thoughts on sword balance

About three pounds give or take depending on the kind of blade is what I vaguely recall being quoted as the accurate weight of a sword. The massively heavy swords that got imbedded into the public consciousness were if I recall correctly based on medieval era wall hangers and ceremonial blades that were never meant for use in combat.

3 Likes

As a begginer Destreza practitioner, I can tell you the major misconception is that daggers vs swords vs sabres vs two-handed swords has mainly to do with damage, instead of being a matter of strategy.

In Destreza, we practice mainly with three types of swords (there are more, but my center is small): Ropera (Rapier), Mano y media (between longsword and bastard) and Montante (would be a zweihander).

The first thing it surprises you is that all swords, no matter the class, weight more than 1 kg and less than 3 kg. An average rapier weights about 1.15 kilograms, an average hand-and-half 1.6kg (mine is a bit heavy at 1.860, but is because it got the pommel replaced by the previous owner, as it was a bit “cabezuda” -balance too forward-), and the montante weights, on average, 2.3kg.

The second thing is that while we spend a lot of time practicing wide sweeps (specially the montante guys), the killing blows are invariable penetrating thrusts. Here’s a link of a video of a free-combat practice, so you can see how boring we are. Notice on the video how invariable the assaults usually end with one sword through the neck or the armpit of the opponent (opponents usually wore cuirass and vambraces, so neck and armpit were less armored).

On the hand-and-half and montante, we use the cutting sweeps to control the area around us (especially with the montante, given that it was a tool to break pikemen formation). Even if they are not directly lethal blows, nobody wants to risk breaking a limb by directly facing a nasty sweep. Even a glancing blow usually gets you reeling back, and that’s with the armor. All sword techniques are designed to keep the opponent at the ideal range of your weapon, which usually means at least as long as the blade is (so 1.2 meter for a hand-and-half, a bit more than 2 meters for the montante). Once they get inside your control zone, though, you have much less options, as the longer blade becomes a nuisance. Once they made me duel against a more experienced guy using a knife (well, you probably would call it a dagger, given that the blade was almost 30cm long), and basically all the encounters ended with him getting inside my area and shanking me before I could pommel him.

The rapier is a bit of a different beast, as it was designed as a purely offensive tool. Rapiers have no parry, and cannot sweep. But they are more flexible, so usually a well placed rapier thrust was enough to penetrate the armor (not by piercing it, but by flexing around the edges). A great duelist weapon, but not a good tool of war… or as someone put it: “a weapon for more civilised times”.

Finally, the third thing I think is important on sword combat is that, against humans, to win is not to kill the opponent, but to avoid injury. Incapacitating your opponent is almost as good as killing him when using swords. After all, once is incapacitated, you can dispose of him. Also, you expect your opponent to behave similarly to you. Nobody wants to get hurt as “there are no healing spells IRL”.

So, given what I know about swords… I would never use one against zombies: all the strategy goes to hell if your opponent doesn’t fear death and seeks hand to hand combat even when you threaten it to cripple its limbs.

Machetes, on the other hand, are good weapons against zombies (at least the basic ones). They are short, so they won’t be a nuisance when zombies swarm you as would be a longer sword. They do cutting damage, as the balance is towards the front of the blade, making it easier to cut bone. If you get one with hand guard you can even punch with it. An axe would also be a good weapon.

But a longsword? I don’t think so.

3 Likes

Averaging between 4.9–6.2 lb

The two-handed claymore was a large sword used in the late Medieval and early modern periods. It was used in the constant clan warfare and border fights with the English from circa 1400 to 1700.

I suppose I was not clear enough. I agree with you that swords were “light” but the weight of the weapon is directly linked to the wielder and is explicitly linked to how much damage it can inflict. My experience is physics and a minor history buff.

Also on the weight of the weapons “most” swords from medieval Europe were 1.5-3lbs with 6lbs being the max as stated by these guys What Did Historical Swords Weigh?. If you read through this article they admit the weight of a weapon directly affects its swing velocity which can be shown with some calculations. Mind you the weight is quickly limited by the wielder. Overall there is a maximum to both the weight and velocity of the weapon and once these are approached technique then takes over and is always omnipresent.

With your other points I do not think anything I said contends with it at all. My point is that a “sharp” weapon would be directly effected by its edge making it harder to swing as it becomes damaged. Strike something orthogonal to the edge and voila you have a wrecked edge. This edge would then catch and get stuck on things which would increase the energy used (only if it hits) to use it again making what I said valid.

Also as a TLDR to this whole thing, swinging around a 3lb stick sure hurts like hell if you get hit by it, and something even with a flat edge WILL cut through something if it exceeds it’s puncture resistance. There are more scientific names for this but it does get the point across.

Also on your point of not using them to fight zeds, I am in agreement with that statement.

TLDR: If I again am not being clear enough let me know. Again my point is a more damaged blade/weapon requires more effort to use.

EDIT: Forgot to mention to TheKobold, the bevel will directly change the ability of the sword because it changes the surface area of the blade. A 1 micron thick, 3lbs blade, with the same swing parameters as a normal sword, will in fact cut through just about anything you have ever encountered because the contact area is so small, though this magical blade cannot exist it still proves the point I am trying to make. Weight is directly linked to the damage maximum a weapon can have as this is directly tied to physics as is the blade’s contact area when it strikes something.

EDIT 2: To be as clear as possible from a pure physics standpoint. Any object has x mass when swung uses y energy which imparts part of that to the object. This energy is what is causing the damage. This energy can be directly assessed as a joule, this energy is then transferred to the object it connects with, in whole or part depending on the object’s contact area.

Depending on this contact area you will see vastly differing outputs in damage and effects.

So all in all

Simple Version:
A “sharp” weapon: mass * velocity / surface area = damage
A “dull” weapon: mass * velocity / surface area = damage

Now as we can see it is the same. What is not the same is the surface area or contact area. So the same weapon depending on it’s state will inflict differing amounts of damage output. This also means that all objects with the same parameters, weight, length, etc, will impart identical damage. This damage is directly effected by the contact area of what is hit and it’s resistance to this damage.

A damaged weapon becomes harder to use and should increase in use of stamina. The damage is effected but that will only show itself if you get down into the nitty gritty of how the surface area changes vs. the resistance of the contact material.

For the game-play aspect, a damaged weapon should require more energy if you hit the object.

1 Like

To the people that know what you’re talking about, don’t worry about the misinformation, the dev team can tell the difference.

4 Likes

I can’t say that I do. The spreadsheet seems solid as a benchmark for determining where similar melee weapons should stand.

I have no idea how you would factor the techniques like rapid strike into a mathematical equation for setting up the average “value” of the weapon. But maybe techniques should be where the main difference is between weapons. Not raw damage, but just how you’re expected to use them.

The balance is out the window when zweihander steps in. Then again, it should do a really high amount, being that it’s a person-sized sword being swung around.

Rapid strike is actually fairly easy to calculate. For every 100 hits you land, N of them is going to be rapid strikes that do 2/3rds damage in 1/2 the moves. Assuming your normal damage before armor is D, the target’s armor is A, and your normal move cost is M, your expected average damage per move for your 100 hits is:
N * max(0, 2/3 * D - A) / 0.5 / M + (100 - N) * max(0, D - A) / M
critical hits make that a little more complicated, because you have to split the critical hits between the rapid strikes and the non-rapid strikes, but that’s just more algebra.

But a sledgehammer’s brutal strike can either win the fight (when it smashes back a brute that was about to slam you) or do nothing (when it stuns a child zombie that you were going to kill anyway). Same with wide strike - it can triple your damage, or do nothing. Those are harder to evaluate.

There’s also the question of the value of block - a baseball bat has medium blocking ability, a quarterstaff has high blocking ability but a slower attack speed for the same damage. How much is that high blocking ability worth? Enough to make up for that lower DPS? I really am not sure.

1 Like

Man, this is why I stopped at statistics. They don’t make you do this stuff in biology.

It’s debatable as far as blocking. I personally find it very useful, especially early game, since you can severely reduce the damage of attacks, and with half decent armor nullify them completely. But you can also get block from some of the arm guards, and since they don’t stack it invalidates the block of the weapon entirely. Does that change how much it’s “worth”?

Sure, the fact that weapon block doesn’t stack with arm guard block can reduce the perceived value of weapon block. But arm guards increase arm encumbrance and weight, and you may not want that either. So really, I don’t know.

My personal experience is that I’d rather have a bat (which my current formula puts at 36 average damage/100 moves) than a quarterstaff (27 damage/100 moves), even though the staff has better blocking. So BLOCK2 over BLOCK1 isn’t worth 9 points of damage to me. Am I alone in that? Is that a good trade-off? Not really sure.

I’m also not factoring in DURABLE_MELEE, which doesn’t matter much in my experience, or FRAGILE_MELEE, which really does.

Anyway, currently I’m averaging damage against a soldier zombie, a smoker zombie, and a survivor zombie. That might be giving too much benefit to piercing weapons, but soldier zombies and survivors are the zombies that tear up early-game survivors so I think there’s some value in biasing the formula towards good armor penetration. Smokers have decent dodge of 4 and no armor but there might be better candidates.

Zombie dogs are another early game killer. They’re faster than most survivors, have good damage, and are fairly hard to hit, which makes the damage and hit bonus more important than armor penetration. I also feel like they’re more common than feral runners, so they might be good to factor in if that’s what you’re looking for.

You’re not alone, imo block on weapons is only really good early game. After that I don’t just think it lacks value but see it as an active detriment since blocking with it opens it up to taking more damage and reducing its effectiveness.

Another thing to note is the advantage of fast hits.

Slow weapon could in worst case scenario almost kill a monster in single hit but has do do a second slow hit to finish it off. This almost halves the “real dps”.
A fast weapon could in ideal situation deal about half of the monster HP per hit and finish it in two efficient hits.

DPS number doesn’t show this at all.

Maybe a “expected time to kill monster X” calculation that does the calculations by using damage per hit and time per hit instead.

I’m currently sticking with the average injury past armor/100 moves idea, because trying to figure out actual time to kill is really, really complicated. That said, here’s where I’m currently going

Avg Weapon Acc Reach Moves Bash Cut Stab
33.11 monomolecular blade 2 1 78 0 33 0
29.87 naginata 2 2 133 5 31 0
29.86 lucern hammer 3 2 178 28 0 18
29.56 glaive 2 2 140 11 28 0
29.52 ji 2 2 177 10 36 0
29.49 halberd 2 2 177 15 32 0
29.06 war flail 0 2 200 58 0 0
28.69 battle axe 2 1 138 21 32 0
27.27 mace 2 1 110 37 0 0
26.93 nodachi 2 1 164 6 48 0
26.87 katana 2 1 115 4 36 0
26.77 zweihänder 2 1 169 18 40 0
26.46 fire axe 2 1 139 20 30 0
26.35 war hammer 2 1 107 20 0 20
26.06 longsword 2 1 139 13 35 0
25.35 arming sword 1 1 119 10 34 0
25.30 morningstar 1 1 112 35 0 6
25.23 war scythe 2 2 167 6 32 0
24.30 jian 1 1 110 6 33 0
24.23 broadsword 1 1 111 8 32 0
23.53 cutlass 1 1 100 7 29 0
22.81 loaded stick 2 1 107 30 0 0
22.58 makeshift glaive 2 2 143 8 24 0
22.57 scimitar 1 1 115 8 31 0
22.03 cavalry saber 1 1 100 6 28 0
21.78 sword bayonet 1 1 108 7 29 0

Approximate rankings are:
30 - combat polearms
28-26 - two handed, combat weapons
26-24 - “full size” one handed, combat weapons
24-22 - “short” one handed, combat weapons
22-20 - combat knives

inferior versions of combat weapons, quality makeshift or improvised weapons, and dangerous tools being used as weapons are probably going to have 3/4ths the value of the equivalent combat weapon. fakes, low quality makeshift or improvised weapons, and less dangerous tools are going to be about half, at best. This seems to match the guidelines in game_balance.md.

The difference in values is because this system tries to account for armor, at least a bit.

I was starting to worry about steel showing up in mines.

Well, as a designer myself, the fact that you’re having trouble coming up with a one dimensional array to describe the ranking of the weapons is a decidedly good thing, as it means that there are enough factors to consider that weapon choice is interesting, and scenario based.

In order to really describe your weapons, you’d need a multi-dimensional ranking chart, with separate axes for target armor and dodge, at a minimum.

I often wade into combat with generic zeds with a standard cutting weapon like a katana simply because of its high damage and fast strikes. It seems to consume a bit less stamina for the DPS I put out, and against targets with low/no armor it does excellent damage. Unfortunately the fast attack thing is an active disad against armored enemies.

So when armored enemies show up, I just pull out a spear and start with the stabbing.

As a note, I really wish there were some kind of feedback on how much stamina various attacks/defenses and so on use, or at least a reference to look it up. Stamina expenditure is now one of the most important aspects of combat, and it’s pretty heavily obscured from the player.

Like does the stamina consumed by an attack relate to its attack speed? Or the weight of the weapon? Or is it a totally flat value? How much stamina does blocking or dodging take by comparison? Etc.

For me the most frequent decisions I generally have to make in combat now are generally stamina related - do I engage the approaching 20 zombies with 3.5 stamina left, or do I need to disengage and catch my breath? What combat style should I go with to maximize my Damage Per Stamina against a given opponent? I have learned to ballpark many of these decisions, but even after many, MANY hours of play, I still can’t realistically answer some of those basic questions about stamina costs.

1 Like

I agree that a multi-dimensional array would be great, except I need a number that an NPC (who has a fairly limited set of AI routines) can use to decide whether to use a katana or a sledgehammer or a stone pot. The NPC is not going to figure out to use the sledgehammer against soldier zombies and the katana against smokers, but I’d like him to have 1 number so he doesn’t end up stashing his katana in his belt and using his stone pot.

Currently, my best guess for that 1 number is average DPS against a heavily armored foe, an armored but dodgy foe, and a fairly dodgy foe. I’d love to have suggestions for a better set of criteria that can be programmed and evaluated in reasonable amounts of time.

To reinforce this one point, this is not the one and only number used to drive weapon balance, this is a “good enough for now” mechanism for NPCs to use to decide which weapons they should hold onto and which ones to discard. In the future we’d LOVE to replace it with a much more nuanced evaluation where the NPC can assemble a kit based on the properties of gear they have available, then attempt to use the best weapon for the job.

Also, the code I used to generate this table is a unit test which I have lying around but haven’t PRed, which can assert a weak ordering on weapon goodness according to the NPC evaluation function. Just as soon as we have meaningful assertions to make we can add it to the test suite.

Ah, ok, I missed the context that this discussion was around what items an AI should choose to wield as a weapon. I was just thinking in terms of basic game balance and player experience.

Given that I had some chucklehead NPC come after me with a Pasta Extruder just the other day, I agree that the AI could use some guidance here. :smiley:

I’m pretty good at doing statistical analysis work in Excel, so if you were to point me at a reasonably condensed version of the combat resolution code, I could probably generate some useful data for pre-seeding weapons with value weights, or possibly come up with some reasonable estimation formulae that the AI could use on the fly to estimate the value of weapon vs. target.

FWIW though, as a ‘good enough’ solution the basic formula and weights you’ve produced already look like they’d work fine for NPCs AI in most circumstances.

I do think the current weapon evaluation code has some issues, because it doesn’t deal with armor protection and severely overvalues rapid strike. There’s just a flat assertion that rapid strike is +32% damage, and for a lot of weapons, that just isn’t true.

Which is why I’d like some discussion of a better formula for evaluating weapons, with the caveat that I’d prefer one that I can write as a google spreadsheet because those are slightly easier for me to think about.