WARFRAME Wiki:Stat Comparison

Design document for automation of weapon stat comparison on Characteristics sections of weapon articles. Discuss here: WARFRAME Wiki talk:Stat Comparison.

Last updated: Sun, 14 Jan 2024 22:08:58 +0000 (UTC) by User:Cephalon Scientia


  1. To provide a consistent and well-defined methodology on how stats are determined to be "very low", "low", "average", "high", or "very high" across the wiki.
    • What makes a weapon "good" or "bad" is on reader's discretion. That judgement should be left to reader's interpretation as the wiki should strive to be as objective as possible. So stat comparison/ranking ≠ gameplay meta. Though obviously stat comparison/rankings can be used as tools to define what's meta, but that onus is on readers.
  2. To standardize the language used to describe relative stats (e.g. "average" should describe statistical mean, not "standard").

  • When a well-defined standard is established, we can automate the stat categorization process since we already store weapon stats on the wiki in Module:Weapons/data. No more going through and manually update multiple weapon articles whenever a new weapon is released (roughly at least one every 3 months with Prime Access; about 20-30 weapons a year) or when the developers go through a major balance pass on certain weapon classes or all weapons:
    • Update 32.2 (2022-11-30) semi-automatic weapon changes, mostly lowering fire rate and buffing damage/crit stats to make up for DPS loss and to make theoretical DPS more achievable without macros
    • Update 32.0 (2022-09-07) modular weapons gilding changes (full stats are available pre-gilding)
    • Update 32.0 (2022-09-07) ammo changes with explicit Ammo Pickup stat and Max Ammo adjustments
    • Update 29.5 (2020-11-19) and Update 30.5 (2021-07-06) glaive changes
    • Update 27.4 (2020-05-01) Railjack armament changes with Railjack Revisited (Part 1)
    • Update 27.2 (2020-03-05) Warframe Revised changes
      • Self-damage removal and AoE changes
      • Changes to shotgun multishot interactions with status effects with status chance buff per individual pellet
    • Melee 2.9999 in Update 26.0 (2019-10-31)
      • (Re)introduction of heavy attacks
      • Base damage, range, combo duration buffs
      • Almost every Melee weapon that originally dealt a single elemental damage type has been converted to Physical + Element
      • Removal of Channeling mechanic
    • Update 25.7 (2019-08-29) bow stat changes
    • Update 24.2 (2018-12-18) separate Archgun stats with release of "Atmosphere" mode + some weapon changes
    • Hotfix 22.13.3 (2018-02-21) and Update 22.14 (2018-03-01) Continuous Weapon changes
    • Update 21.0 (2017-06-29) introduction of dual-wielding glaives and single secondary feature also included some buffs
    • Update 18.5 (2016-03-04) Conclave weapon changes
    • Update 17.0 (2015-07-31) removal of Stamina (melees don't use stamina now)
    • Update 15.1 (2014-11-05) heavy melee weapon adjustments
    • Update 11.0 (2013-11-20) Damage 2.0 overhaul of damage system
    • Update 7.2 (2013-03-22) Weapon Balance Changes (Update 7.x)
    • Etc.
  • If you want to do mass numerical weapon stat changes on the wiki, you do not need a dedicated bot account to (semi) automate your edits. Editors just need to update the respective data store listed on Module:Weapons/data.
    • Lowers the bar of entry for wiki editing; editors still need to be willing to learn how to edit JSON-like data structures if they are not familiar with them already
  • To easily allow for new raw/derived stat comparisons on demand. In other words, automating stat comparisons opens up the possibility to add more stats by changing a few lines of code.


Categorization is the process of organizing data into multiple bins/groups for analysis of how a particular datum relates to the overall population. These bins are classified as such:

  • Very low
  • Low
  • Below average (as in "just below the norm")
  • Average (as in "norm")
  • Above average (as in "just above the norm")
  • High
  • Very high

Raw Value Categorization[]

  • Easy to create, just pick a contiguous range of values for what you consider to be "very low", "low", "average", "high", or "very high".
  • Some stats like Riven Disposition already have its own categorization that can be adopted by the wiki:
    • ●○○○○ (0.5-0.69) = very low, ●●○○○ (0.7-0.89) = low, ●●●○○ (0.9-1.1) = average, ●●●●○ (1.11-1.3) = high, ●●●●● (1.31-1.55) = very high
  • Very subjective. There is no mathematical consistency in how one determines whether or not a stat is high or low. For example, two people can say two different numbers for the "average" critical chance based on different metrics (note that sample values are made up for the sake of demonstration):
    • 15% critical chance is the "average" because most weapons have that critical chance (mode).
    • 25% critical chance is the "average" because that is the numerical mean when accounting for all weapons' critical chance.
    • 20% critical chance is the "average" because that is "good" enough critical chance to perform well against tough enemies after modding (say Sortie level). A critical hit occurs once every 5 shots on average.
    • 20% critical chance is the "average" because that is the minimum critical chance (for primary rifles) to reach 50% crit chance after equipping Mod TT 20px Point Strike. Having a critical hit occur once every 2 shots on average is a "good" standard for crit builds (i.e. crits are not rare/uncommon enough to warrant not building towards crits).
    • 22.73% critical chance is the "average" because that is the minimum critical chance (for secondaries) to reach 50% crit chance after equipping Mod TT 20px Pistol Gambit.
  • Prone to extreme recategorization when new weapons (also new mods and more sources for a particular stat bonus) are added since these bins are hard coded.
  • Relies on other people's "game sense" which differs based on how long they have played the game and how they play the game (i.e. "gut feeling" or instinct).
  • "Eyeballing" method can make quick estimates but bins are volatile since each person's bins may be different in size and range.
  • Some stats can only be practically organized into two bins ("high" and "low"). Multishot and Noise Level are some examples of binary categorization (i.e. "low" and "average" multishot is pretty much equal to 1 regardless of your method of categorization). This means we cannot use the same categorization scheme for all stats based on raw value.


Sample reformat https://warframe.fandom.com/wiki/Panthera_Prime?diff=2224262

From User:Misdirected on 2022-08-22:

I looked at the total comparison table per category
So right now I'm doing all auto-rifles, so I compared auto-rifles, and got a relatively good grouping on what stats are normative

Stats (Very Low/Low/Average/High/Very High):


[High/Low] Base Damage: 0-15/16-20/21-40/40-60/60+
[High/Low] Critical Chance: 0-5%/6-10%/11-15%/16-20%/21-25%+
[High/Low] Critical Damage: 1.5x/1.6-1.9x/2/2.1-2.5x/2.5x+
[High/Low] Status Chance: 5-10%/11-15%/16-20%/21-25%/26-30%+
[High/Low] Accuracy: 0-10/10.1-20/20.1-30/30.1-40/40.1-50+
[High/Low] Recoil: This I have no damn clue on, using what's pre-existing.
[Fast/Slow] Fire Rate: <=3/3.1-6/6.1-10/10.1-15/15.1-20+
[Fast/Slow] Reload: 5-3.5/2.5-3.4/2-2.4/1.5-1.9/0-1.4
[Large/Small] Magazine Size: <30/30-45/46-60/61-75/76-90+
[Large/Small] Ammo Capacity: Largely conditional based on the weapon. Average is 540, very high is probably 800+.

Not all of these are linear, like damage or fire rate, but most tend to be fairly linear
I didn't bother adding an "extremely" category as of now because I felt like that can runaway really fast, but it may be needed for some categories, like magazine

Standard Deviation-Based Categorization[]

  • Mathematically consistent in how categorical bins are calculated. Uses the fundamental statistical concepts mean and standard deviation to determine the boundaries of bins.
  • Bins automatically adjust when new weapons are added.
  • Prone to extreme outliers that shift the mean (applying a bias towards either extreme, like an asymmetrical bell curve).
  • Categories may not follow a binomial distribution of weapon count. For example, say 50% of melee weapons have a critical chance less than the average 15%. Some people would argue that 15% is "average" since there is more weapons with less than that critical chance.
  • For stats that have a low range of values (e.g. multishot), this method is not good for determining whether or not a stat is "low", "average", or "high" since in most cases, the result will be binary, tending towards either extreme of the spectrum
  • More difficult to understand for the common reader. Concepts like "value is one standard deviation from the mean" are abstract in nature.


From User:Misdirected on 2022-08-22:

Auto-rifle base damage:

11, 12, 12, 15, 16, 16, 18, 23, 24, 24, 24, 25, 25, 25, 27, 27.5, 28, 28, 28.75, 29, 30, 30, 30, 31, 32, 32, 35, 35, 37.5, 38, 39, 40, 40, 79, 46, 46, 52, 57, 58, 60, 66, 75, 132, 120, 100, 30, 33

Mean: 39.186170212766
Standard Deviation: 25.709478966657

25.7/2 = 12.85

39.2 +/- 12.85 = 26.35 - 52.05

39.2 - 25.7 = 13.5

39.2 + 25.7 = 64.9

Using this math, we define "average" as any value within MEAN +/- HALF_STDV
High and low are anything from outside average but within one standard deviation
Very high and very low fall outside of that.

Auto-rifle damage:
[High/Low] 0-13.4/13.5-26.3/26.4-52.1/52.2-64.9/65+

The short of it is, it produces pretty reasonable results compared to what I had prior
[High/Low] Base Damage: 0-15/16-20/21-40/40-60/60+

Percentile-Based Categorization[]

  • Know exactly how well a weapon stat is relative to other weapons. This can be expressed in every day language: "15% critical chance is higher than 50% of all weapons".
  • Value of weapon stats are independent from how bins are categorized. In other words, bins will always be consistent.
    • Thus, extreme outliers have no effect on bins.
    • New weapons will have no effect on bins also.
  • Percentile ranges are subjective and hard coded.
  • Categorization bins don't account for minor stat differences. In other words, it is not nuanced enough to compare stats in some cases.
    • For example, say the breakpoints for average/below average/low critical chance is 20%/19%/17% respectively. If a weapon has 19.5% critical chance, then it would be labeled as having "below average critical chance" even though it is just a mere 0.5 percentage points off from being "average". The gap between an "average" critical chance and a "low" one is a 3 percentage points difference (which may or may not be significant depending on factors like mod availability or the magnitude of bonuses associated with a stat).


Percentiles categorization:
* 90-100% - "Very high"
* 75-90% - "High"
* 60-75% - "Above average"
* 45-60% - Average, doesn't display anything
* 30-45% - "Below average"
* 15-30% - "Low"
* 0-15% - "Very low"

Data Clustering[]

Data clustering is the process of grouping objects such that objects in the same group are more similar to each other than other groups (like grouping based on closeness of values).

  • Can get fine-tuned and more distinct groups that other methods cannot achieve.
  • One of the more complicated algorithms to implement.
  • Clustering leaves numerical gaps between groups so bins are not contiguous. If one wants to make bins contiguous when using this method then the bins will be likely be subjective. Someone has to make the opinionated decision on where to draw the boundaries between groups.

Weapon Sampling[]

Like how there are different ways to categorize weapon stats by, there are many different ways to choose which weapons to compare stats against. In order of increasing specificity:

  1. All weapons
  2. By slot
  3. By class
  4. By class and trigger type
  5. By class, trigger type, and some other arbitrary trait (like unique mechanics)

Sampling All Weapons[]

A particular stat of a weapon is compared against the same stat from all weapons in WARFRAME.

  • Easiest method
  • Does not account for differences in:
    • Weapon gameplay role (e.g. sustained DPS, burst DPS, status proccer/primer, utility, etc.)
    • Weapon usage (e.g. close-ranged vs. long-ranged)
    • Mod availability and compatibility (some weapon slots have access to better mods for a particular stat, making stat comparisons misleading)
    • Mechanics (e.g. melees have stances that have force procs so comparing status chance between melees and primaries is meaningless to players)

Sampling by Slot[]

A particular stat of a weapon is compared against the same stat from only weapons in the same arsenal slot (e.g. primary, secondary, melee, arch-gun, etc.).

  • More specific sampling
  • Easy to group weapons by slot, just look at what arsenal slot a weapon can be equipped on:
    • Primary
    • Secondary
    • Melee
    • Arch-gun
    • Arch-melee
    • Exalted weapon
    • Companion weapon
    • Railjack Armament
    • Railjack Ordnance
  • There is a more clear gameplay divide between slots; each weapon in each slot is usually meant for a specific gameplay purpose based on the weapons and mods available in that slot:
    • In general, primary weapons are used for long-ranged, sustained engagement OR crowd clearing. The "main" weapon.
    • In general, secondary weapons are used for short to medium ranged, burst engagement, usually against single targets. The "backup" to primary weapons or the "sidearm".
    • In general, melee weapons are used for close-ranged engagement, finishing off enemies with high-risk/high-reward due to proximity to enemies.
    • In general, Exalted weapons and arch-guns (on the ground) are used as temporary heavy artillery for dealing against tanky enemies or crowds. The limited-use "special" or "exotic" weapon.
  • Does not account for differences in weapon classes within the same slot (e.g. gunblades which are more for ranged encounters vs. heavy blades which are more close-ranged). There is a big range of gameplay diversity between weapons in the same slot.

Sampling by Weapon Slot and Class[]

A particular stat of a weapon is compared against the same stat from only weapons in the same weapon slot and class. There is no one standard for classifying weapons, so in the end it would be a mix of official classification, classes from mod compatibillity, classes from skin compatibility, and player opinions:

  • Primaries
    • Rifles
    • Shotguns
    • Sniper rifles
    • Bows
    • Crossbows
    • Launchers
    • Arm cannons
    • Spearguns
    • Exalted weapons
  • Secondaries
    • Single pistol
    • Dual pistols
    • Single shotgun
    • Dual shotguns
    • Crossbows
    • Throwing knives
    • Exalted weapons
  • Melees
    • Each melee class with a stance
    • Exalted weapons
  • More finer grouping of weapons to match shared functions and gameplay role (e.g. primary launchers are for AoE clearing while sniper rifles are for single target precision shooting).
  • Some combinations of weapon slot and class have a low population of weapons (e.g. Ghoulsaw Ghoulsaw is the only assault saw melee as of Update 31.3 (2022-03-28)) so stat comparisons will tend towards binary results. Stats will either be "very low" or "very high".
  • Some weapons have unique mechanics that changes how the weapon is used, so simple comparisons are not a good metric for whether or not a stat is "high" or "low" compared to other stats in the same group since they don't account these nuances. Theory ≠ practical

Sampling by Weapon Class and Trigger Type[]

A particular stat of a weapon is compared against the same stat from only weapons in the same weapon class and with the same trigger type (e.g. semi-auto rifles, automatic rifles, burst rifles, etc.)

  • The highest level of granularity allows us to group the most similar weapons in function and gameplay role for a "fair" comparison.
  • Some combinations of weapon class and trigger type have a low population of weapons (e.g. TenetEnvoy Tenet Envoy is the only automatic launcher as of Update 31.3 (2022-03-28)) so stat comparisons will tend towards binary results. Stats will either be "very low" or "very high".
  • Some weapons have unique mechanics that changes how the weapon is used, so simple comparisons are not a good metric for whether or not a stat is "high" or "low" compared to other stats in the same group since they don't account these nuances. Theory ≠ practical


See these discussions on this issue:

  • "What the heck are these ridiculous comparisons? "Advantages over other auto-trigger pistol-type secondary weapons"? "Tied for second highest critical multiplier with Azima, Furis"!? So like, 2.0x which most laymen would regard as a very average crit damage multiplier? Who edits and writes these ridiculous assessments? Practically no relevant information is being imparted here that isn't already pointed out in the sidebar outlining the stats. I understand sometimes making comparisons within weapon classes can help to enumerate the relative strength of a weapon, but when you're being so specific, it starts to lose all meaning. For example, within the duplex selective shotgun trigger group of weapons, I'd warrant Tigris Prime rates fairly highly! Just complete nonsense."

What qualifies as an advantage/disadvantage?[]

A statement that describes an advantage must have an equal and opposite statement that describes the disadvantage counterpart. For example, if "high damage" is an advantage, then "low damage" is the disadvantage statement.

Likert Scale to judge statements by
Disadvantage (generally negative trait) Neutral/normative/gray area (general trait) Advantage (generally positive trait)
Large disadvantage Medium disadvantage Small disadvantage Neither disadvantage or advantage Small advantage Medium advantage Large advantage
"Very low" "Low" "Below average" "Above average" "High" "Very high"
-3 -2 -1 0 +1 +2 +3


  • Fundamental numerical weapon stats:
  • Derived weapon stats:
    • Average shot damage
    • Effective fire rate
      • For charge trigger attacks, fire rate is not the only factor in determining the rate at which attacks occur; need to account for both fire rate and charge time stats
      • For burst trigger attacks, fire rate is not the only factor in determining the rate at which attacks occur; need to account for burst delay (time between individual shots in a burst) and burst count (number of shots in a burst)
      • For semi-auto trigger attacks, assume that the player reaches the maximum possible fire rate as determined by the weapon's fire rate stat
    • Burst DPS
    • Sustained DPS
    • Average proc count per shot
      • A "shot" is arbitrary defined as some unit of attack that consumes some number of ammo. For continuous weapons it is a single damage instance. For shotguns, it includes all pellets from a single attack input. For burst weapons, it is just a single bullet, not the entire burst.
    • Average proc count per second
    • Average crit count per shot
    • Average crit count per second
    • Number of shots per magazine
    • Time to deplete magazine before reloading
    • Reload speed (as inverse of total reload time, measuring reloading progress as a percentage per second)
      • For example, if a weapon has a total reload time of 2s, then its "reload speed" is 1/2 = 0.5 = 50% reload progress / second
    • Archgun Deployer cooldown per round (for atmosphere Archguns)
      • Cooldown time for every shot fired
    • Archgun Deployer cooldown per second (for atmosphere Archguns)
      • Cooldown time for every second of continuous firing accumulates
  • Subjective, non-numerical weapon stats (cannot be automated):
    • Recoil
    • Accuracy (accuracy value in-game does not really mean much quantitatively, but we can somewhat categorize weapons into buckets based on their accuracy number)
      • Compare minimum and maximum spread values instead as those are the raw stats that produce the derived Accuracy stat.
    • Ammo efficiency/economy (don't have a method for finding a numerical value; some metric of max ammo, magazine size, and fire rate)
      • Also depends on player performance so there may be variances in ammo economy experience (e.g. hit rate % when using the weapon and gameplay style)
  • Stats that only (dis)advantageous relative to a specific weapon class:
    • Noise Level
      • Only count as a disadvantage for alarming level for weapon classes that are silent (e.g. bows)
      • Only count as an advantage for silent weapons for weapon classes that are alarming (e.g. launchers)
    • Presence of self-stagger effect
    • Whether or not a weapon is hit-scan or projectile
      • Only count projectile as a disadvantage amongst weapon classes that are mostly hit-scan
      • Only count hit-scan as an advantage amongst weapon classes that are mostly projectile-based

Normative Traits[]

We typically don't need to explicitly state normative traits that are common among certain weapon classes or types. Most readers are reading specific weapon pages to learn about unique mechanics, interactions, and stats (aside from farming locations and the like) that make that weapon standout from other weapons under the framework of "opportunity cost" or theorycrafting. Questions that can be answered by the reader based on the information provided include:

  • "Is this weapon better than the ones I currently have?"
  • "Should I then farm this weapon if it's good?"
  • "What role does this weapon play in my build?"
  • "Does this weapon build for crit, status, or both?"
  • "What Warframes/playstyle benefit most from this weapon?"

As such, stating normative traits is often moot and unproductive. For example, since hit-scan is the most common projectile type, we do not need to put a note in the Characteristics section saying that "this weapon is hit-scan". However, if the weapon shoots a projectile that travels to its target, then we can say "this weapon's projectiles have travel time".

Note that normative traits are subject to change as the game evolves.


The "average" or "basic" gun (e.g. Braton Braton):

  • Is hit-scan (for bows and crossbows, this would be projectile instead)
  • Has multishot of 1
  • Uses the ammo type of its weapon slot (e.g. primary ammo for primary rifles)
  • Has an Alarming noise level (for bows and crossbows, this would be Silent instead)
  • Has the default zoom level
  • Has the default recoil behavior
  • Has the default ammo pickup amount:
  • Has the default max/reserve ammo:
    • 540 for primary rifles and primary crossbows
    • 120 for primary shotguns
    • 72 for sniper rifles and bows
    • 210 for secondaries
  • Has no Punch Through
  • Has Damage Falloff after 300m
  • No innate polarities (except in Exilus slot)
  • Can deal additional damage on headshots and trigger "On Headshot (Kill)" conditions
  • Needs to reload after depleting magazine
    • Can interrupt reload animation to continue firing if magazine is not empty
      • No partial reloads if reload animation is interrupted early
    • Can force reload by pressing "Reload" key (PC default R )
  • Has no Alternate Fire mechanic


The "average" or "basic" shotgun (e.g. Strun Strun or Boar Boar):

  • Is hit-scan
  • Has multishot greater than 1
  • Has low accuracy and high spread
  • Uses the ammo type of its weapon slot (e.g. primary ammo for primary shotguns)
  • Has an Alarming noise level
  • Has the default zoom level
  • Has the default recoil behavior
  • Has the default ammo pickup amount:
  • Has the default max/reserve ammo:
    • 120 for primary shotguns
    • 210 for secondaries
  • Has no Punch Through
  • Has a short Damage Falloff
    • Numbers vary per shotgun
  • No innate polarities (except in Exilus slot)
  • Can deal additional damage on headshots and trigger "On Headshot (Kill)" conditions
  • Needs to reload after depleting magazine
    • Can interrupt reload animation to continue firing if magazine is not empty
      • No partial reloads if reload animation is interrupted early
    • Can force reload by pressing "Reload" key (PC default R )
  • Has no Alternate Fire mechanic


“It's taking longer than I calculated.”
This page is actively being worked on and may not be completely correct. Please assist in making this page accurate. See WARFRAME Wiki:Research on ways to perform research on this game. Click here to add more info.
Do AoE's inherit base projectile's Noise Level stat?

The "average" or "basic" area-of-effect attack:

  • Has two independent damage components, direct hit and radial hit
    • These can independently deal damage, crit, and proc status effects
    • Both components deal damage immediately on impact
      • There is no delay ("embed delay") between direct hits and the corresponding AoE component
    • Damage components may have different stats
      • Usually AoE component inherit/share critical chance, critical multiplier, and status chance from direct hit component
  • Has a limited radius
  • Has Damage Falloff that starts from the center of the spherical area-of-effect and ends on a point of the sphere's surface
  • Radial hits ignore surrounding terrain and does not require line-of-sight (i.e. enemies can be hit behind terrain, objects, and other enemies)
  • Has no Punch Through
  • Has a 1x headshot multiplier and cannot trigger "On Headshot (Kill)" conditions
  • Inflicts self-stagger or knockdown if player is close enough to AoE radius
  • Does not have a minimum "arming-distance" for AoE component to occur; AoE can happen at any distance between player and point of impact


The "average" or "basic" melee (e.g. Skana Skana):

  • Has basic Melee Combo behavior
    • Starts at 1x combo (0 hits)
    • Combo duration of 5 seconds
    • Heavy attacks consume all combo (i.e. 0% Heavy Attack Efficiency)
    • +1 combo per hit, multiplied by stance's damage multiplier
    • +1 combo per attack blocked
    • Caps at 220 hits for 12x heavy attack multiplier
  • Has Silent noise level
  • No innate polarities (except in Stance slot)
  • Has Follow Through, meaning each successive enemy hit in a single melee sweep will receive less damage
  • Can deal additional damage on headshots if melee attack's hitbox hits overhead
  • Default sweep radius of 0.25m (radius of obround hitbox).

Not qualified[]

  • Dealing a particular damage type (e.g. "High DmgImpactSmall64 Impact damage, effective against Shield, Machinery, and Proto Shield"). Type advantages are nuanced and are usually inconsequential in general gameplay. Players usually mod for a particular damage type so original damage distribution is moot. There is an argument where players may desire less/more of a particular damage type for proccing applications so having less/more damage as an "advantage/disadvantage" varies player to player.
    • I believe this content was holdover from Damage 1.0 and early Damage 2.0 system where there was a more clear "rock-paper-scissors" strategy in picking and choosing what damage types to use against enemies.
    • Modded elements will always deal more damage than base physical/primary elemental types because of how elemental mods apply multiplicatively to base damage stats. For example, modded DmgCorrosiveSmall64 Corrosive damage will most likely always deal more damage than DmgImpactSmall64 Impact even if enemy Damage Type Modifiers from their health/armor classes are involved. Whether or not the original base Impact damage is high or low relative to other weapons is insignificant and is practically trivia.
      • Logically speaking, one would think physical damage bonuses work the same way as elemental damage bonuses in WARFRAME, but they are only multiplicative to the specific base physical damage type that the bonus is for. For example, a weapon with no base DmgSlashSmall64 Slash will not benefit from mods like Mod TT 20px Fanged Fusillade or Mod TT 20px Heavy Trauma. It is probably misleading to state that a weapon does "High Impact/Puncture/Slash damage, effective against X health/armor class" because it implies players should mod for physical damage bonuses when in fact they would probably benefit more from adding elemental mods instead.
    • There is the issue of status distribution where sometimes it is advantageous for a damage type to be low to allow for other more desired procs to occur (e.g. low DmgPunctureSmall64 Puncture is good when weapon has high DmgSlashSmall64 Slash damage for DmgSlashSmall64 Slash procs). Would having no DmgPunctureSmall64 Puncture damage be an advantage in this scenario?
      • This problem is more prevalent when you have innate elemental damage types fighting for proc dominance. For example, say a weapon has low base DmgSlashSmall64 Slash and high base DmgViralSmall64 Viral. Typically speaking, for a DmgSlashSmall64 Slash proc build, you want Slash to proc more often than DmgViralSmall64 Viral since DmgViralSmall64 Viral caps out at 10 procs on an enemy for its max damage bonus and has opportunity cost associated (the damage difference between 0 and 1 DmgViralSmall64 Viral proc is larger than the damage difference between 1 and 2 DmgViralSmall64 Viral procs). So would having high DmgViralSmall64 Viral damage be a disadvantage in this scenario?
    • You could also argue that there is not an opportunity cost when a weapon deals a desired damage type for status effects (like having innate DmgViralSmall64 Viral procs for damage bonus) since you don't have to use up one or two mod slots to create a secondary elemental type.
      • Though if the presence of an innate primary/secondary elemental type if an advantage, then the opposite statement would be the lack of a primary/secondary elemental type is a disadvantage. While it is true to say that KuvaBramma Kuva Bramma doesn't deal DmgViralSmall64 Viral damage, is that really a "disadvantage" when there is player choice in modding?
      • Could also interpret the presence of an innate primary/secondary elemental type as a "bonus", "luxury", or "quality of life" stat since it is not typical for weapons to have an innate elemental type.
    • There is also the unique effect of DmgToxinSmall64 Toxin damage ignoring Shields and deals directly to health (exception of Bosses and special enemies). In most cases, this is a net advantage since you are ignoring a health class that has higher Enemy Level Scaling than traditional health.
    • Stances often times have forced DmgSlashSmall64 Slash and DmgImpactSmall64 Impact procs, making innate physical damage distribution moot for melee builds that rely on DmgSlashSmall64 Slash and DmgImpactSmall64 Impact procs.
    • Ultimately, editors have to make the distinction whether or not an "advantage" or "disadvantage" is referring to an element as a damage type or its status effect. Otherwise the conversation is too muddy to draw any useful conclusions.
  • Having a particular polarity or not. For example, having an innate Vazarin polarity may mean that a player needs one less Forma to finish their build, but this is only an "advantage" for those that make use of that polarity. If having a polarity is an advantage then that means the inverse must be true, but having no polarity is neither an advantage or disadvantage.
    • If having a polarity is an advantage (to halve the cost of equipping a mod of a certain polarity), then the disadvantage would be to have an "anti-polarity" or a polarity that doubles the cost of equipping mods regardless of its polarity.
    • Only time a slot polarity would really be an "advantage" is if it matches a particular Stance's polarity. But people play-styles vary so one may have one stance preference over another (and stances may have different polarities). Though one can approximate average damage/second using a stance's damage multipliers and animation times to make a more "objective" judgement on what stance is better.
  • Having Punch Through or not. If "having punch through" is an advantage, then the disadvantage would be "no punch through" logically speaking, but most weapons have no punch through so there isn't really a solid basis for comparison. Punch through seems to be more of a feature/bonus stat rather than a fundamental stat for comparison.
    • If you reword the statement to "has 2m punch through", then the disadvantage would be "has -2m punch through" which doesn't exist in the game (What's the opposite of piercing more than one enemy with one bullet?).
    • If you reword the statement to "has high punch through", then the disadvantage would be "has low punch through" which poses an issue for weapons with no punch through at all (How low of a value is 0m punch through? Infinitely low since it doesn't exist?).
    • Though there is an argument that positive punch through allows for situational DPS increases since a single bullet can damage more than one enemy, potentially doubling or tripling the damage output. In these situations, having punch through would be a clear advantage.
      • Though value of punch through matters since 0.2m won't even go through a single enemy while infinite punch through on bodies will give a significant DPS boost.
  • Presence of any zoom levels beyond default zoom. High zoom is situational and can hinder weapon handling despite giving additional bonuses for snipers. It is difficult to label zoom levels as (dis)advantages when most weapons just use the default zoom. Any higher zoom magnification can be treated as a bonus stat.
  • Ammo consumed per shot. A majority of gun attacks consume only 1 point of ammo per shot. Any attacks that consume more than 1 point of ammo will logically be classified as having "high ammo consumption per shot" if we do direct numerical comparisons. Conversely, most attacks will have "low ammo consumption per shot" if this were to be implemented which is a meaningless statement since by default gun attacks consume only 1 point of ammo per shot. We can treat ammo consumption as a general knowledge, putting them in the same category as unique weapon mechanics.
    • Number of shots per magazine before reloading (not the same as magazine size necessarily) may be a better metric of comparing this sort of behavior.
  • How easy/hard is it to farm/obtain the weapon or how high/low are weapon part(s) drop chances. This experience varies from person to person and despite being able to calculate the average number of runs to get parts (assuming RNG produces numbers in a uniform distribution), mission difficulty differs from player to player as well as mission completion times.

"Neutral" Knowledge[]

Notes that do not fit under the binary definition of advantage/disadvantage and are mostly informational (things that are nice to know, may share scope with Notes section):

  • Damage type distribution
  • Punch Through
  • Ammo consumed per shot
  • Innate polarities
  • Stance compatibility (e.g. stance slot has a Madurai polarity which fits Mod TT 20px Cleaving Whirlwind)
  • Forced procs
  • Exclusive mods (weapon augments)
  • Mod compatibility (e.g. can equip Mod TT 20px Firestorm and Mod TT 20px Cautious Shot)
  • Additional zoom magnification and zoom bonuses
  • Unique mechanics
  • Riven compatibility
  • Weapon passives
  • Signature Weapon bonuses
  • Innate upgrades (e.g. Progenitor bonus from Liches)

Edge Cases and Exceptions[]

For very specific builds or loadout compositions, there may be cases where a stat with a high value may be a "disadvantage" and vice versa, contrary to player intuition due some interesting interactions or mechanics. In simple terms, big number may not always be "good" so don't treat the wiki's classification of stat advantage/disadvantage as gospel when comparing weapons. We can label these scenarios as "situational (dis)advantages".

Here are some examples of this nuance:

  • For weapons that use Mod TT 20px Internal Bleeding for DmgSlashSmall64 Slash procs, having low Fire Rate (below 2.5) is a positive to maximize the mod's bonus.
  • Laetum Laetum can benefit from having a low critical chance stat since it has an Evolution perk that gives a "50% chance to deal 2000% damage on non-critical hits.". In this case, a low critical chance is an "advantage".

Comparison Language[]

If possible, try to reframe the comparison so that the bigger the numerical value, the more "desired" or "advantageous" a trait is. For example, instead of comparing reload time of multiple weapons, compare the inverse of reload time ("reload speed") instead so it is programmatically easier to implement.

Stat Disadvantage Advantage
Damage Low High
Multishot Low High
Status chance Low High
Critical chance Low High
Critical multiplier Low High
Fire rate Low (referring to raw value)
Slow (referring to rate)
High (referring to raw value)
Fast (referring to rate)
Charge time Long (referring to raw value)
Slow (referring to rate)
Short (referring to raw value)
Fast (referring to rate)
Magazine size Small Large
Max/reserve ammo Low (referring to raw value)
Small (referring to quantity)
High (referring to raw value)
Large (referring to quantity)
Projectile Speed Slow Fast
Disposition Low High
Reload speed (as inverse of reload time) Slow Fast
Reload time Long (referring to raw value)
Slow (referring to rate)
Short (referring to raw value)
Fast (referring to rate)
Recharge rate Low (referring to raw value)
Slow (referring to rate)
High (referring to raw value)
Fast (referring to rate)
Range Small (referring to raw value)
Close (referring to distance)
Large (referring to raw value)
Far (referring to distance)
Radius Small Large
Damage falloff reduction High Low
Average shot damage Low High
Burst DPS Low High
Sustained DPS Low High
Average proc count Small Large
Average proc count per second Low High
Ammo economy Low (referring to some arbitrary measurement)
Bad (referring to trait's desirability)
High (referring to some arbitrary measurement)
Good (referring to trait's desirability)
Recoil High/Uncontrollable Low/Controllable
Accuracy Low High
Spread (inverse of accuracy) High/Wider Low/Tighter


See these discussions on this issue:

Current Implementation[]

As of 20:02, 6 November 2022 (UTC):

Known Limitations[]

Many weapons have additional mechanics that may cause their performance to not match theoretical damage output or behavior. Readers are expected to make their own judgments on weapon performance using the objective information the wiki provides as a starting point of reference. The best way to judge weapon performance is to test them against enemies in actual gameplay instead of relying solely on the wiki, using metrics such as kills per minute.

The following are known limitations in the way we perform stat comparisons on the wiki:

  • Does not accurately access performance of weapons with Infinite Body Punch Through.
    • Actual DPS and procs/second values will be higher than theoretical values in the presence of multiple enemies that are grouped up.
  • Does not factor in modding. Only compare stats as is.
  • Does not factor in Riven Mods that have the potential to improve a weapon's performance. Due to the randomized nature of Rivens, they are a wildcard so we cannot expect consistent results from one player to another.
  • Does not factor in Incarnon upgrades.
  • Status procs and damage over time effects are ignored in DPS calculations. More focus on raw output.
  • Ignoring Damage Reduction and Damage Attenuation on enemies. Stat comparisons solely focus on individual weapon stats, but in actual gameplay, enemy resistances and mechanics need to be accounted for to get the full picture of the damage system.
  • Does not factor unique Multishot interaction with most Continuous Weapons.
  • Does not factor Stance multipliers, forced procs from stances, and combo animation speed for Melees.
  • Does not factor Follow Through mechanics when using melees against groups of enemies.
  • Does not factor area-of-effects in damage calculations. All damage calculations assume single target performance.
  • Does not factor player-dependent performance when using guns which may be affected by weapon Accuracy/Spread and/or Recoil mechanics. A weapon may have high paper DPS, but if the player cannot hit anything with it or use it effectively (e.g. cannot control recoil or spread), then actual DPS will be lower.
    • Semi-auto trigger weapons may have fire rates unachievable by the average player (especially when modded or with bonuses), requiring some sort of macro or modified keybinding to reach theoretical DPS.
  • Assumes player only uses that weapon and that weapon only for simplicity. No status primers, debuffers, or mixing of weapon usage (e.g. melee + secondary).
  • We treat all attacks equally. There is currently no real differentiation between main attacks, alt-fire, area-of-effects, additional effects, etc. In other words, we don't have a weighting system for attacks, a way to denote "importance" for an attack, a filtering system for attacks, or a way to omit an attack from stat comparison.
    • For example, a weapon may deal 0 damage on direct hits, but has an AoE component. That 0 damage attack will also be included in determining percentile bins despite being an outlier.