Sunday, March 8, 2015

D&D 5e weapon analysis and suggestions

UPDATE: I have made a website of the simulator (, which addressed some issues raised here.

SEEN ALSO: Description of battle simulator and some analysis, the GitHub code for the encounter simulatoranalysis of dice equivalence
The weapons list in 5e PHB p. 149 is rather small —even 4e had a longer list. Consequently, I started studying and looking into it. I was planning on writing a spatbook but it got nowhere as there are too many problematic elements and it was taking too long.

In summary, there is a lot of imbalance in the weapons list and it is not simply based on playing style (shield, heavy weapon, two handed weapon combat, finess etc). Ideally simulations would be amazing… but I don't have time for that unfortunately. The major suggestions here are (some not fleshed out):
  • that weapons could be made by changing a simple 1dn damage die to 2d(n/2) or 1d(n–2)+1 dice as it would not change the average, only the probability distribution.
  • New dynamics could be introduced, such as maintenance (whetstone…), defensive (trident…), bipod etc.
  • Class specific weapons, especially weapons that improve fighter's special abilities, but not ridiculous items such as the harp bow (magical item for Bards…)
  • Better weapons must be very expensive weapons. Unlike armor, the price of weapons is really negligible. If the setting is a technomagic/steampunk one, mercurial swords (sigh…) and recurve or even compound bows could be preset at a high cost.

Balance. There are many analyses online that look into the unbalanced nature of the weapon table in page 149. The most obvious imbalance is the trident, which is a martial weapon that costs 5 times more than the spear, which is a simple weapon and has the same stats. Here more weapons that could be added are discussed keeping such points in mind, along with other considerations, in order to be balanced, something that plagues splatbooks, especially homebrew ones. Consequently, before building an analysis is needed.

Damage amount. One big difference is the amount of damage. Swords come in a variety of lengths, so make the best illustration.

  • d4 —bad. dagger (one handed, light, finesse, thrown)
  • d6 —poor. shortsword (one handed, light, finesse, martial)
  • d8 —good. longsword* (one handed, martial)
  • d10 —very good. bastard sword** (two handed, martial) 
    d12 —awesome. Greatsword*** (two handed, heavy,
Note 1. daggers and shortswords do piercing damage, while longswords, bastard swords and greatswords do slashing. There is a piercing 1d8, the rapier (finesse), but it is different kind of short.
Note 2. There is no bastard sword anymore as the longsword can be used two-handed (versatile) and deals the same damage.
Note 3. The greatsword technically does 2d6, which is something that baffles me due to the different distributions a 1d12 (equal) and a 2d6 give (bellcurve) and the fact that 1d12 weapons exist. Parenthetically, the greatsword can be seen as a claymore (cf. Equivalence table 5.2 (page 75) of the 3.5 Book of Sword and Fist) or as a manga sword. In 3.5 it was pictured as a claymore, while Pathfinder as manga. Also, in my opinion heavy is not a very scary penalty nor are the weights too taxing —6 lb for a greatsword? a sleeping bag is heavier at 7 lb!
The axes are nearly identical, except that they all do slashing, the handaxe is better and the greataxe does 1d12:

  • 1d4 —
  • 1d6 —handaxe (light, finess, thrown)
  • 1d8 —battleaxe (one-handed, martial)
  • 1d10 —waraxe (now 2 handed battleaxe)
  • 1d12 Greataxe (two-handed, heavy, martial)
The main commonality is that for as damage dice increase, the extra perks decrease. Simple, light, fitness, thrown give way to martial, two-handed, heavy.

Damage type. There are three times of damage: Bludgeoning, piercing and slashing. Different monsters have different resistances and vulnerabilities (e.g. a skeleton can only be bludgeoned), but it is generally believed they are equivalent.

Weapon proficiency. Simple and Martial (no exotic for now). Almost everyone can use a simple weapon, while martial are available to the larger HD classes. So does it mean anything in terms of value, given that the low HD don’t normally use weapons? Looking at the list it actually does, which means that martial weapon options are generally better. The best way to find out is to do maths…

Maths. Therefore, based on the above, if points were to be given I would say:
  • max dice x/2
  • thrown +2
  • light –1
  • finesse +2
  • two-handed –2
  • heavy –1
But that is my feeling of the worth of the features, heavy is way more. What happens when I transcribe the table and solve the system of equations by least squares?
Calculations: I put most of these values as binary, while dice, gp and weight and range are numeric from 0 to 1, where 1 is the maximum value (d12 (zeroed on d4), 75 gp, 18 lb, 150’).
solve(qr(Arms, Lapack=true),Ones) in R.
Try 1. Fail. Ammo and melee are mutually exclusive, so are messing up the table and point out an underlying flaw to this approach to find how these values are weighted.
Try 2. Fail. Weight and gold make no sense, regardless of tweeks, so I deleted them.
Try 5. Fail. Changing the bad traits to negatives values (0 to –1).
Try 6. Fail. Heavy and range are negative? Oh, I give up. These are the results.
There is no rhyme or reason.

Unsual dice. The dice used in D&D are Platonic solids, so there are no odd numbers, d2, d14, d16, d18. However, recently Impact miniatures have started selling d5 and d7. A d9 and d11 are available solely printed ($$$). I really like the d7 and others, but I don't like the asymmetry of d5 (two values on faces, three on corners). These dice could be added to fix some incosistencies.

Bell curves. Rolling a d12 (ignoring a 1) and rolling a 2d6 are different as the latter will give a bellcurve as is clearest in Settlers of Catan. Parenthetically, it’s not a Gaussian curve as that would need 12 0–1 coin flips + 1. Anyway, there is a cool website that shows it.
The difference is in consistency. On average a d12 will roll 6.5 and a 2d6 a 7 (robber!), so there is only 0.5 difference on average, but the 2d6 will be much more consistent (smaller standard deviation) than a d12, which could just as easily roll a 6, 12 or 1. So a 2d6 weapon does the same damage as 1d12, but is more consistent. This is a really cool thing and could result in an interesting dynamic, but WotC doesn’t really use it like so. Multiple dice appear in the scaling by a 3/2 factor for large weapons as no dice exist.

Medium Weapon Damage
Large Weapon Damage
Ideal Large Weapon Damage
1d4 (4.5)

In D20 modern and in Dragonstar, the guns do a ridiculous amount of damage in the form of many d6 or d8. You would have expected that axes deal damage of the 1dx series, while swords 2d(x/2). Obviously that can’t be fixed, but there are lots of other weapons. For example, Japanese swords could be more precise:

  • 2d2 —Tantō (one handed, light, finess)
  • 2d3 —Wakizashi (one handed, light, finess, martial)
  • 2d4 —Katana (one handed, martial)
  • 2d5 —Kodachi or katana (two handed, martial)
  • 2d6 —Nodachi (two handed, heavy, martial)

Or an expanded series based on the above, but Elven. The Elven curveblade in 3.5 was a one-handed d10 (=bastard sword) and all elves got proficiency at it…

  • 2d2 —Elven knife. Sigil (one handed, light, finess, thrown)
  • 2d3 —short Elven curveblade. “magoleg” (one handed, light, finess, martial)
  • 2d4 —Standard Elven curveblade. magol (one handed, martial)
  • 2d5 —Standard Elven curveblade. magol  (two handed, martial)
  • 2d6 —Great Elven curveblade. lhang (two handed, heavy, martial)
A clear conceptual example of an Elven curveblade is Arwen’s sword in LotR, here showing the versatile trait:

Crossbows overview. As was calculated before, the ammo feature is nearly as expensive as dice increment —which makes sense as melee needs to be better than ranged.
  • 1d4 —dart (finesse, not light)
  • 1d6 —Hand crossbow (light, loading)
  • 1d8 —Light crossbow (twohanded, loading, simple)
  • 1d10 —Heavy crossbow (heavy*, twohanded, loading)
  • 1d12 —Great crossbow (3.5 Book of Swords and Fists, “large”, 15 lb.)
Note 1. Heavy crossbow used to weight 8 lb, now weights 18 lb. I sense a typo.

Great crossbow. In 3.5 the loading times differed, a heavy crossbow took a full action to load (winch it says, correctly called a cranequin), a light crossbow a move action (lever). The splatbook great crossbow is a full action too. The loading time resulted in the annoying issue of players claiming to always walk around with loaded crossbows. 5e “Loading” is all good, but upgrading the great crossbow will need more. I would say it takes a move action and must be fired within 5 turns. It used a windlass.
In d20 modern large weapons were two handed, while huge weapons (e.g. M60) needed a bipod. So if I were to make a Great crossbow I would add the bipod rule.
Great crossbow, 15 or 27 lb.,1d12 Piercing, 1 move loading, heavy, bipod, Range (150/600).
The two dice increment comes at the cost of bipod (move to set up) and move loading.

Repeating crossbows. These in 3.5 were expensive exotic weapons that made the loading time not an issue. In a steampunk/technomagic realm, would they be balanced?

Bow overview. There are two bows: shortbow 1d6 (two-handed, simple) and longbow 1d8 (two-handed, martial —yes, dice increase for martial).
Recurve bows also exist and (in fibreglass and steel) are the modern Olympic ones as they are more efficient than longbows. They have been around since antiquity and the Mongols were formidable with composite recurve bows. Composite refers to the use of different woods.
A more recent variant is the compound bow, which has winches (called cams) to store energy and is really nice, albeit weird (if one’s accustomed to a recurve). Compound crossbows are common in modern hunting.

Recurve bow. So what is the difference between bows and crossbows?
shortbow (2H, simple, 80’) vs. hand crossbow (1H per RAW, martial, light, load, 30’)
longbow (2H, M, 150’) vs. light crossbow (2H, simple, load, 80’).
Long range is rarely important in my experience, so I believe the longbow is underpowered.
So introducing:
recurve bow: 1d10P, 2-handed martial heavy.
One needs more strength to fire a longbow than a recurve, but let’s ignore that.
On youtube there are videos of a chap sequentially shooting three arrows in 1.3 seconds while riding pillion on a motorcycle (he holds the arrows), so loading isn’t really an issue.

Composite bow. In 3.5 the rules for composite bow were a tad unusual. However, with 5e dex is added to the damage roll, so composite bows would be dex + str? Best avoid them.

Compound bow and crossbow. A cool thing for a steampunk/technomagic realm, but the rules would depend on the other weapons in order to not be broken. Clockwork could be temperamental and jams on 1.

Garotte is a great weapon on paper as it is a grapple attack: it never worked effectively.
Sabre is a western scimitar: I feel like it should have a use.
bayonette is just a dagger in 3.5, but it makes one always armed.

A pavise is a shield with a spike at the bottom to place in the ground therefore giving cover to crossbowmen. A mantlet is a movable wall. Both would provide cover, but not shield AC.

1 comment:

  1. Fascinating, thank you! Did the Heavy Crossbow really gain 10 lbs between 3.5e and 5e? Great work (on the sim, too)!