A guide on how to proportion vigor, endurance, and damage stats with objective math rather than softcaps, with additional information on other stat properties.
 
 

Starting classes

Any excess levels in unused stats will pit you against players that many levels higher in the PvP matchmaking. Any low stat can be leveled up, but a high stat can’t be leveled down, so classes are chosen not for their strengths but their minimums. If you want to have a class for every build, but with the fewest overall save files, your best option is to make a vagabond, a hero, and a prisoner: this will allow you to minimize intelligence, faith, and arcane separately or any combination of the two, if not all three in the vagabond’s case. These three stats are the caster stats, which nearly every build will minimize at least one of, since no item in the game utilizes all three. The only reason to use any other class is to minimize strength or dexterity, but this becomes less necessary the more levels you gain, especially once you have enough in total to take off radagon’s soreseal. The stat that is your absolute lowest priority will almost always be one of the last three.
 
 
Chart detailing certain stat sums for each class, highlighting minimums:
 
 

 
 
 

Leveling and soreseal talismans

The first thing to understand about build optimization is the difference between nominal and proportional gain. Nominal gain is the flat return you get for some amount of investment, like 21 HP for one level, while proportional gain is that return measured in proportion to what you already had, like 21/1793= another 1.17% health for one level.
 
 
Proportional measures are important because unlike nominal measures, proportions can be compared against one another. Nominal health and attack rating (AR) can’t be compared since they don’t share units, but a build with X% more health will be evenly balanced against a build with X% more damage. We know this because any build’s hits-to-die equals their total health divided by their opponent’s damage per hit: if each side of this fraction is increased by X%, the value will remain the same, since X/X=1 and anything times 1 equals itself. Whichever version of your build has the highest product of health and damage is therefore the one which deals the most damage-per-life, and will always kill other versions of itself in fewer hits than they kill it.
 
 
Most guides only tell you to hit softcaps, but softcaps only mark changes in nominal gain. You might get a fixed 5 AR per level between two caps, but each time you add it, it represents proportionally less than before since you’re now adding to a larger base. Within certain level ranges, the ideal place to stop leveling a damage stat might even be below the first softcap, and not all caps are alike. You might think that after hitting 40 vigor the next thing to do is to grab your first damage cap, but the next 20 levels of vigor add 31% HP whereas 20 levels of damage won’t typically increase your AR by an equal percent even with S scaling. Going from 20 to 40 arc on an S-scaling ripple crescent halberd only adds 29.8% AR for example, though this will be a bit more damage in practice due to how base defenses work.
 
 
This does come with some complications. Since vigor only increases your base health, this means it won’t increase the hits each healing flask represents. With 14 flasks included in your healthpool, vigor’s last 20 levels go from adding 31% health in a duel to as low as 3.5% outside them. On the other hand, leveling endurance for armor presents no such issue and includes a number of other benefits like poise and stamina. This means for any set amount of offense, we can only calculate the maximum vigor and minimum endurance that might be proportional, with the former being traded for the latter based off of how many flasks you expect to drink, how likely they are to be clipped by your max health, and how important endurance’s additional benefits seem. It takes about 45 vigor to have have double a flask’s worth of health though, and the crimson amber medallion won’t add more than five level’s worth of value until that point.
 
 
Generally speaking, going from 28 to 40 units of armor weight at medium load with favor equipped takes about 10 levels of endurance. Doing so raises your physical damage negation from a quarter to a third. Negating a quarter of the incoming damage is equal to having another third more HP, since that would be your fourth third, while negating a third of the incoming damage is equal to having another half more HP, since that would be your third half. If we divide our new three-halves by our old four-thirds, we’re left with nine eighths, or rather a one-eighths increase in effective HP for the ten levels added. The more armor you add, the less efficient it becomes, and there’s an endurance softcap at 25 too, but the last ten levels of vigor before the cap at 60 add only 11.5% HP instead, closer to a ninth. This means that you should have at least 40 units of armor weight by the time you reach 60 vigor. Medium load with the banished knight set and 5.9 units of talisman weight takes at least 18 endurance with the favor talisman on. The bull goat’s talisman or great jar’s arsenal will only multiply endurance’s effectiveness further. For elemental defense you’re only going from a fifth to a quarter though.
 
 
This logic also applies to talismans. Most commonly, people ask about radagon’s soreseal. Wearing a soreseal gives you 20 extra levels to push around, but also makes you take 15% more damage, and takes up a talisman slot too. If you replace the soreseal with a crimson amber medallion, you’re going from a 15% penalty to 8% more hp. If we divide 108% by 1/115%, we’re left with 124.2% our prior effective HP, or about a one-quarter increase. If your worst twenty levels aren’t giving you a quarter more health, damage, or a fifth more damage negation, they’re worth removing to replace the soreseal. Any opponent with a quarter more damage-per-life than you can kill you in four hits with any weapon you can kill them in five with, so I emphasize that the soreseal is not worth using by level 80 to 100. You could argue that taking damage is optional, but in that case, don’t level vig, and you’ll be able to cap multiple offense stats by that point anyways.
 
 
Ultimately this leaves one ideal leveling path:
 
 

• put the soreseal on
• hit the minimum requirements for what you need
• get 45 vigor or more with the soreseal
• take the soreseal off there or by the point you hit 60
• get a middling amount of your damage stat and at least 18 endurance by that point also
• take the soreseal off and get the rest of what you need
• level mind whenever you feel like it, but not past 38 until you’ve got everything else

 
Mind is the only stat that can’t be objectively compared to all the others, but a blue flask recovers 220 FP, and you’ll hit 221 by 38 mind, so 38 serves as a practical maximum. Some builds might not use it at all, but intelligence builds don’t need to worry about pushing it below the prisoner’s 12. This is because anything that reduces your total FP (like the hierodas crown, the lightest armor piece in the game to add 2 endurance, giving it nearly negative weight) will bring you down to 77 FP total, the exact amount you’d want to cast law of regression and law of causality back to back. Without any penalty, the vagabond’s 10 will give you 78, but the vagabond spends a point more on arc than the prisoner spends on faith. Personally I think 1 level for 2 mind is worthwhile, and you might want less strength. Alternatively a faith build might need 23 mind to cast all its buffs back to back.
 
 
Altogether I think this makes level 125 a much better PvP meta than 150. Level 125 gives you enough statpoints for:
 
 

• 60 vig
• at least 50 in your main stat
• a net 16 in your lowest two stats (9 and 7)
• a net 30 in two other stats (12/18, 16/14), or just 20 with 10 more on your main
• a net 48 mind and endurance, (38/10, 23/25), or more in your damage stats

 
Go any lower and people might not finish leveling health, but go any higher and they’ll only increase their damage, FP, or poise. I would rather stop where health stops.
 
 
Chart comparing a 40 unit set to the most similar 28 unit alternative:
 
 

 
 
The physical average is slightly worse than what 10 levels of health would add (contrary to my envelope math), but the overall average is slightly better than what a crimson amber medallion would (8%), making the great jar’s arsenal charm a priority over the medallion if you’re gaining at least 12 units of equipload that way (though the amber medallion can be replaced once your health has been lost, giving it more flexibility than great jar). With this charm the comparison moves overwhelmingly in endurance’s favor.
 
 
 

Attack rating and base defenses

Attack rating (AR) is not exactly proportional to damage. This is because of how base defenses work, which are the defenses gained from leveling up, not the ones gained from wearing armor. How these defenses work is that the lower the incoming AR is, the larger a percentage of it they remove. This might make them sound flat, but adding 85 fire damage with fire grease will still make a difference even against players with 85 fire defense, so they aren’t: they’re just curved.
 
 
Either way this carries some implications. First, it means that any X% increase in AR will represent slightly more than an X% increase in damage, since you’re both increasing your AR by that much and decreasing the percentage negated. Second, it means that split damage AR values will add up to less than their total suggests, since each damage type is treated as a separate, smaller attack. The compound effect all this is best observed through dark bead from Dark Souls 1. Dark bead fired a seven-projectiles spread where each projectile was split between magic and physical damage, for a fourteen-way split overall. This put the spell deep into the defense curve, so much so that stacking the bellowing dragoncrest ring, red tearstone ring, crown of dusk, and power within for a net 302.4% damage boost on paper could result in over 600% more damage in practice. Critical modifiers work like any other damage boost too, so people will favor split damage weapons for ripostes sometimes, since split damage weapons tend to come with higher total AR to compensate for their split in any other case. Even an R2 has a modifier that makes it stronger than an R1, so an affinity that splits your damage can weaken your R1 while strengthening your R2 sometimes.
 
 
Conclusively what this means is that if you really want to test damage, your best bet is whacking the humanoid mimics by night’s sacred ground. Short of that, AR is a decent approximation if you can watch for splits. A two-way split weapon will usually do around three-quarters what an equal amount of pure AR would, or more depending on your total AR or the size of the attack you’re using. A three-way split weapon is… not worth using, if you can help it.
 
 
As a final note, ashes of war can greatly vary in their scaling. If at any point your ash hits the enemy with your weapon, it’ll use your weapon’s AR, but projectile and spell-type ashes ignore everything about your weapon besides its upgrade level. These ashes have their own internal scaling based off of whatever affinity they would normally prescribe (i.e. without a whetblade), but despite apparently being spells they do not benefit from the graven mass or flock’s canvas talismans.
 
 
 

Status buildup and resistance

Compared to attack rating, status scaling is far more simple. If your weapon has blood, poison, madness or sleep buildup (even if through grease), then the amount of that status applied per hit will increase as you level arcane by an amount based on your weapon’s arcane scaling. If you have no scaling, arcane has no effect. The final buildup number is the amount of buildup applied per hit (prior to modifiers: R1s, L1s, and R2s will not be the same), while your opponent’s resistance is the amount they can receive. There is no sort of defense stat between these: just buildup and health.
 
 
I highly recommend having a mottled necklace, a stalwart, clarifying, and immunizing horn charm, and even a prince of death’s cyst on you at all times to switch on when you see someone running a status-focused build. Ripple weapons with sleep grease and high arcane can put you to sleep in two hits, for example. A clarifying horn charm +1 should give you about 50% more meter though, which bumps their hitcount up to three. Since taking more hits to be slept also means you take more time to be slept, this means you also have more time to let the buildup drain between activations, so in practice it might be more like four hits due just to a 50% bonus. This also makes consuming boluses much safer, since now even if your consumption is punished, you won’t be put back into immediate danger of being slept yet. All of the focused status resistance charms can be stacked with the mottled necklace too, which by itself should be enough to prevent a pulley crossbow from knocking you out in a single triple-shot sleep bolt volley (which can be comboed into from a colossal sword poke then combo into giant’s hunt and so on).
 
 
Don’t try this with the dragoncrest talismans as their effects are reduced to a quarter their value in PvP. Some spells like golden vow are reduced to half effectiveness in PvP, but are still quite good: the dragoncrest talismans are the only items which I think are ever reduced this much.
 
 
Also, status affinities replace your weapon’s buildup rather than augmenting it. Don’t use the blood affinity on a blood weapon, etc; use occult if you want your innate status to scale with arcane.
 
 
 

Stat requirements

Suppose you’re running 18 faith for vyke’s spear. Dexterity scales with it better, and vigor generally scales better than damage, so ignoring everything else, you’d think to level vigor, dex, and faith in that order. However, vigor and dex won’t to give you access to incantations like faith does: having 20 more faith on the other hand gives you access to blessing of the erdtree, which recovers 1080 HP over 90 seconds. Over the course of a 90 second fight, that’s another 50% health or so, for just 20 levels which do still do give you damage, just slightly less. If you stack protection of the erdtree on top, you can negate 30% of all elemental damage too, or 60% of one type with more specific barrier spells instead. This with 50% more health totals out to 375% effective HP, and you can still stack golden vow on top to boot. Sure, an intelligence build can wipe all your buffs out by casting law of regression once, but that takes 37 int, only a point shy of blessing’s faith requirement. Point being, you can’t just think about damage scaling; you have to also think about stuff. Stuff is what builds are made out of.
 
 
Hitting important stat requirements can be an order of magnitude more important than scaling your stats efficiently, even just for buffs.
 
 
 

Cast speed

Dexterity can improve cast speed, and though this doesn’t sound at all comparable to damage or health, it can be far more important. Though cast speed only reduces the startup of spells and never the time between chained casts, a spell like bestial sling can be thrown faster by alternating between left and righthanded seals to loop the startup in place of using the normal followups you’d get casting from one hand. Silly as it sounds, this lets you throw slings a more frequently than your opponent can roll above about 50 dexterity’s worth of cast speed. Combine that with the fact honed bolt can have its startup entirely skipped when cast after a sling, and you’ve got a setup that’s both very hard to approach and still dangerous to walk away from. Sling flinches just about everything and even outranges glintstone pebble. You’ll have to throw around a dozen slings before you get a whole sling-ahead of your opponent’s roll count, and this setup is pretty much shut down by any decent shield, but all this again points to the fact that having more tools can easily take precedence over damage efficiency.
 
 
Cast speed caps at 70 dex, and scales linearly. Cast speed only divides the startup time of spells, where a sigil is shown, and does not improve running or rolling casts. Radagon’s icon adds 30 dex-worth of cast speed, while holding azur’s staff adds 40: these effects stack, but are still limited by the cap.
 
 
Video by Amir on cast speed:
 
 

 
 
 

Poise and hyperarmor

Lastly, poise. Poise lets you take hits without flinching, but it can be confusing for a number of reasons. Firstly, you’re building around your opponent’s weapon instead of your own, without even knowing what it is yet. Secondly, some of your own weapons might have hyperarmor frames in their animations, which can alter the equation. What we do know is that, in general, it’s always worth having 40 units or so of armor weight for the defense alone once you’re high enough level to hit 60 vigor, at least if you have a talisman improving your equipload per level. With just the banished knight set, this comes out to 56 poise, enough to tank any twohanded straightsword, curved sword, katana, spear, axe, or twinblade R1, or a onehanded heavy thrusting sword R1. If we instead wear the veteran’s chestplate and greaves (nerf these) with the briar’s helmet and gauntlets though, we actually save a half unit of weight, and get a special effect on our rolls, and end up two poisebreaks higher. These are the poisebreaks you should expect people to hit.
 
 
61 poise lets you tank:
 

• two of the smallest attacks in the game in a row (30+30)
• the smallest onehanded R2
• a whip or small thrusting sword onehanded jumping R1
• any paired straightsword, curved sword, katana, spear, axe, or twinblade L1
• any twohanded heavy thrusting sword R1
• any onehanded hammer or flail R1

 
67 poise lets you tank:
 

• two of the second-smallest attacks in the game in a row (33+33)
• one of the smallest with one of the third-smallest in a row (30+36)
• any paired heavy thrusting sword L1
• any onehanded attack 61 poise could tank, but twohanded
• anything 61 poise could tank, but with cragblade added (60+10%)

 
The next highest break that’s actually important is probably 91, since this lets you tank a twohanded jumping R1 from a heavy thrusting sword. This is important since this move will overwise combo into a standing R1 afterwards, so it can save you from a lot of damage. More often people will go for the next highest break, 93, since it’s only two points away and also lets you tank onehanded halberd and reaper R1s, which are the next heaviest R1s after hammers and flails (and thrusting halberds are just really good weapons). This all assumes you aren’t using hyperarmor though.
 
 
Elden ring’s hyperarmor equation sucks. How it works is that, in hyperarmor frames, your poise gets averaged with some arbitrary value based on the weapon you’re using. For greatswords, this value is only 60, so if you have any more than 60 poise, greatsword hyperarmor actually causes your poise to be reduced. Luckily 60.5 poise is still enough to tank any 60-poise attack, so the ever-important 61 poisebreak still works, but beyond that hyperarmor hurts you.
 
 
The only way around this is the bull goat’s talisman. Despite what it says on the stat page, the bull goat’s talisman does not increase your poise by a third: it reduces your opponent’s poise damage by a quarter. Outside of hyperarmor, there’s no difference between multiplying your poise by 4/3 or dividing your opponent’s damage by that amount, but, by going last in the equation, the bull goat talisman escapes from getting averaged down to half its value in hyperarmor. This means greatswords aren’t trapped at 61 poise: they’re just trapped at 61 armor poise, and can still go to 81 (82) with the bull goat’s talisman before their hyperarmor turns negative.
 
 
The upside to hyperarmor is that attacking still provides a partial poise reset like in Dark Souls 3. Most ashes of war will provide a 100% reset, but standard attacks will still bring you back to 80% if your poise was below that value. This means an attack can be poised through indefinitely if you can do so with just that much of your total. To find out how much you need, just divide the poise damage by 80%, then have more than that. Make sure not to add the bull goat’s talisman to this, but rather to multiply by 3/4ths afterward.
 
 
This means that to consistently tank a 66 poise attack, you need over 82.5 poise; the bull goat’s talisman lowers this to 61.875 armor poise, and since hyperarmor averages your poise with 60, you’ll need an extra 1.875 too, which comes out to 63.75 total. Therefore it takes just 64 poise on your armor to consistently hit the 67 poisebreak in hyperarmor with a greatsword. On the stat page though, it’ll say you have 86.
 
 
To consistently tank a 60 poise attack, you need over 75 poise. The bull goat’s talisman lowers this to 56.25 armor poise. Since this is below 60, averaging it with 60 actually can save you 3.75 poise. Therefore it takes just 53 poise on your armor to consistently hit the 67 poisebreak in hyperarmor with a greatsword. On the stat page though, it’ll say you have 71 poise.
 
 
Spreadsheet by Basscanon90 with all the game’s poisebreaks (these are all 1 above their respective attack’s poise damage):
 
 
https://docs.google.com/spreadsheets/d/1pHudihNsTW3LNP9-AqPKybhwyftl5QI1pD09kl986Ok/edit?usp=sharing – [google.com] 
 
 
 

Personal tierlist and the shortest banlist

SS TIER: LACKS COUNTERPLAY
 

• Bloodhound’s step: consecutive BHS uses have one punishable frame between their invulnerability periods, but for an attack to land, you cannot be invulnerable on your screen OR your attacker’s, meaning that if your screens are displaced by even one frame, you are not punishable
• Royal knight’s resolve/determination: these ashes are direct upgrades to endure; in DS3 endure (perseverance) was a useful tool that could grant poise to let one R1 combo into another for 100% more damage, but RKR adds 80% AR to any one hit and can be cast in advance
• Seppuku: with the right setup (occult) seppuku can proc status in two hits even on people wearing the correct horn charm. Bleeding in two hits with the ability to punish any bolus attempt means every bolus attempt will only put the user back where they started, and on a heavy rapier you’ll almost always be able to land phantom hits with running r2s on people who run away.

The rest of this list will assume these are banned, as it would be entirely different if they weren’t. If you’re invulnerable except while trading damage, all you really need is a colossal sword rolling attack to win practically every trade, short of your opponent using RKR over BHS. As a matter of taste, the jar shard talismans should probably be fourth in line on any banlist.
 
 
S TIER: DEMANDS COUNTERPLAY
 

• Greatshield poking: the fingerprint shield used to be able to hit 100 (100%) stability (stamina damage negation) with the greatshield talisman, making its guard unbreakable, but this has since been fixed and ice spear exists as an almost decent counter to greatshields
• Double bestial sling: alternating between left/right bestial slings can be faster than chain casting them from one hand, to the point slings come out slightly faster than consecutive rolls with enough cast speed, but this can be countered by any decent physical shield
• Placidusax’s ruin: the spell as a whole isn’t great but the initial shockwave comes out in 18 frames (at 60fps) even without investment cast speed, making it particularly hard to react to with anything besides quickstep/bloodhound’s step, which unlike normal dodges will activate on press instead of release (not that you would even know your opponent has ruin)
• Status builds: paired bleed weapons are absurdly good, especially with seppuku or when paired with frost, and with the right setup sleep can be applied in two hits or even just off a colossal sword crouch poke via an offhanded pulley crossbow; this can be prevented by wearing the right talismans and using boluses, but forcing your opponent to switch gear makes these setups S tier
• Paired spears: paired spear L1’s are what elevate 61 poise from “practical” to “mandatory” on any build that attempts to engage in melee
• Heavy rapiers: heavy rapier twohanded running R2s are basically the only standard attack in the game that can keep up with people sprinting away, and it takes 91 poise for their twohanded jumping R1s to not combo into another R1 on the ground
• Having over 93 poise: this is enough to tank a onehanded halberd R1 and is very likely to force your opponent to play differently; it’s enough to bump any weapon in the A-F range up a tier against the right weapons

 
A TIER:
 

• Rivers of blood, moonveil (L2), hand of malenia, nagakiba (length)
• Greatspears
• Thrusting halberds (can combo off colossal sword crouch pokes)
• Swordspear halberds
• Light rapiers (with a shield, paired, or in the offhand)
• Unpaired spears
• Fire strike, glintstone pebble, and bestial sling generally

 
B TIER:
 

• Any katana that isn’t A tier (besides wakazashi)
• Any halberd that isn’t A tier
• Straightswords
• Greatswords (with a consistent 61 poise in hyperarmor)
• Colossal swords (by themselves, below 93 poise)
• Curved swords
• Curved greatswords
• Twinblades
• Reapers
• Whips

 
C TIER:
 

• Reduvia and black knife
• Hammers
• Axes
• Fists
• Claws

 
D TIER:
 

• Daggers besides reduvia and black knife (including wakazashi)
• Greathammers
• Greataxes
• Colossal weapons
• Flails

 
 
 

Typical 125 builds (TL;DR)

54 str build: 60/10/35/54/13/9/16/7 (13 dex for estoc/crks: flip int/fai for ruins gs)
 
57 dex build: 60/10/25/18/57/9/18/7 (18 faith for treespear, vykes, black knife)
 
quality build: 60/10/25/34/50/9/9/7 (the quality affinity sucks but bh fang is okay)
 
 
52 int build: 60/22/25/(net 30)/52/6/9 (52 int for comet and lusat/azur staffs)
 
50 faith build: 60/23/25/(net 30)/7/50/9 (blessing+barrier+vow takes 23 mind)
 
50 arc build: 60/21/25/(net 30)/7/11/50 (flies takes 11 faith)
 
 
pure wizard: 60/38/10/8/12/60/7/9 (trade mind/vig for int as needed)
 
pure priest: 60/38/8/11/10/7/60/10 (trade mind/vig for fai as needed)
 
lean tear: 60/9/20/16/12/7/8/72 (80 arc silver tear mask and sleep grease ripples)
 
 
str/int build: 60/12/33/50/14/20/6/9 (20 int for fallingstar jaw, demihuman cap)
 
int/str build: 60/12/23/26/18/50/6/9 (royal greatsword favors int even twohanded)
 
hierodas/haima crowns will still let you cast regression and causality back to back
 
 
tribuff arcfaith: 60/23/10/11/10/7/38/45 (38/45 dragon seal = 9/74 erdtree seal)
 
bandit version: 60/23/10/9/13/9/38/42 (12 int with sellen’s crown for swiftshard)
 
arc catalysts improve status buildup and dragon seal beats pure faith below 75
 
 
tribuff strfaith: 60/23/16/38/13/9/38/7 (38 fai for blessing, 38 str for 57 twohanded)
 
tribuff dexfaith: 60/23/16/16/35/9/38/7 (vyke’s spear, talismans max cast speed)
 
better to use weapons that scale in both stats than casting four or five buffs usually
 
 
dexarc build: 60/19/23/16/35/7/14/30 (morgott cs, talismans can max cast speed)
 
dex only affects the physical half of rivers/poleblade damage: they’re for pure arc
 
strarc wants 14 dex for mohg spear or marias sword but they don’t scale the same
 
 
int/fai catalysts tend to blow even though lots of the spells are nice
 
int/arc doesn’t even get thorn spells (they take faith), they just get clayman bubbles
 
int/dex has war sickle, death rite weapons etc, but azur’s staff exists for free speed
 
 
most builds here with endurance below 25 can be improved by trading up to 10 vigor for endurance if you’re wearing favor or the great jar’s arsenal, and vigor might be less valuable in pve where a large fraction of your health comes from flasks (you still want 45 to sip two)
 
 

 
 
Hope you enjoy the Guide about ELDEN RING – How to Level up and soreseal talismans, if you think we should add extra information or forget something, please let us know via comment below, and we will do our best to fix or update as soon as possible!