A Beginner's Guide to Theorycrafting - Including both Videos and Text Descriptions

Full video explantion (~20 minutes)

Brief video explanation (~7 minutes)

• Evasion

It is important to note that evasion applies to attacks only. Additionally, evasion is entropy-based. What this means is that if there were to be a single enemy attacking you, you would notice that there was a pattern to how frequently they were able to hit you (e.g., only every fifth attack actually hits you). Evasion is also capped at a 95% change to evade, so you will eventually be hit by an enemy--you can't evade everything.

It is important to note that there are two components that matter when determining if you will evade an attack--your evasion rating and the enemy's accuracy. When you look at your character tab in game and it shows your chance to evade, this is actually an estimate based upon enemies who are your current level. In practice, how likely you are to evade an attack depends upon the enemy. For example, you'll more frequently evade a regular, white mob's attack, but get hit more frequently by a rare or unique monster of the same level. This is because those enemies have a larger accuracy rating than the typical mob.

Movement speed and movement skills are also exceptionally useful for evasion, as they allow you to position yourself to avoid being able to be hit. Combining them with evasion helps you to maximize the power of your evasion by making your character hard to hit.

TL; DR - Evasion applies to attacks only, is entropy-based, capped at 95% chance to evade, is based upon the enemy's accuracy and your evasion rating, and movement skills and movement speed help you maximize the effectiveness of your evasion.

• Dodge

Unlike evasion, dodge can apply to either spells or attacks. Spell dodge is explicitly stated, where as if the word "dodge" is only present, it is assumed that this applies to attack dodge. It is also important to note that dodge is pseudo-random, and is capped at a 75% to dodge--you can't dodge everything.

• Blocking

Similar to dodge, blocking can apply to either spells or attacks. Blocks spells is explicitly stated. However, if this is not explicitly stated, it is assumed that the block change applies to attacks only. Block chance is probability-based, and similar to dodge, is capped at a 75% chance to block--you can't block everything. Block is acquired through based block chance (explicitly stated on shields and staves, and is 15% by default when dual wielding weapons). It can also be acquired by unique items such as Rearguard, and improved by picking up blocking passives as well as equipping unique items like The Anvil or Wings of Entropy.

TL;DR Block applies to spells or attacks, is probability-based, capped at a 75% chance to block, prevents all damage, and can be acquired through passives and items.

• Damage Mitigation

There are two primary types of damage that are dealt--physical damage and elemental damage (yes, chaos damage exists, but it behaves in the same way elemental does).

Physical damage is mitigated through a variety of sources, including armor, passives (e.g., Soul of Steel), endurance charges, and items (e.g., Basalt flasks). Some sources reduce physical damage by a flat amount (e.g., -X damage taken from attacks), a percentage (e.g., endurance charges reducing physical damage by 4% per endurance charge), or based upon the amount of physical damage being dealt (e.g., the armor mechanic).

Elemental damage is mitigated through resistances, which reduce the amount of damage by a parentage.

When an enemy attack or casts a spell at you, PoE checks the following, in order, to determine if you take damage, and how much you take.

1. Avoid being hit

PoE first checks to see if your character is hit by the spell or attack. There are two defensive mechanics that you can draw upon at this stage to help deal with this--evasion and dodge. Again, evasion applies to attacks only and dodge can apply to spells or attacks, depending upon which you have (see mechanics tab above).

2. Avoid the damage you're hit by

Well, maybe you end up being hit by the spell or attack. However, this does not necessarily mean that you're guaranteed to take damage--block is the mechanic that allows the spell or attack to hit you, but prevents all of the damage associated with that spell or attack. Again, spell block and block are different (see mechanics tab above).

3. Shifting the damage type to position your character to be able to mitigate it more effectively

If you're hit and you can't prevent the damage, then PoE checks to see if you're able to convert the damage from one form to another. One of the most dangerous types of hits in PoE are large, physical hits, as it is exceptionally difficult to mitigate them. However, converting part of a large, physical hit to part elemental and part physical makes the damage easier to mitigate. This is why items such as Lightning Coil, Taste of Hate, and Cloak of Flame are such powerful and valued uniques--they make one of the most dangerous elements of the game less dangerous.

Shifting one damage type to another is not unique to physical damage though. There are items, such as talismans than can convert a portion of damage from one type to another.

4. Mitigating damage

When your character is hit with damage, PoE then checks your characters defenses to see how much of this damage is mitigated. Elemental (and chaos) damage is checked against your elemental (and chaos) resistance. It is reduced by whatever percentage resistances you have. For example, if you were hit with 100 fire damage and you have 75% fire resistance, you would mitigate 75 of that 100 fire damage, allowing only 25 fire damage.

Physical damage varies depending upon the mitigation. Armor is the most commonly utilized physical mitigation defense, but how much it mitigates the physical damage you're hit with depends upon the size of the physical hit and your armor rating (see mechanics in the tab above). However, there are other ways of mitigating physical damage, including (but not limited to) endurance charges, the Soul of Steel notable on the passive skill tree, and basalt flasks. Stacking multiple sources of physical mitigation (e.g., using a basalt flask with a large amount of armor) can help you to effectively mitigate a large amount of physical damage.

5. Taking damage

After mitigating damage, PoE then checks to see how much damage will be applied to your "effective life". It first removes this damage from your energy shield, regardless if this energy shield is protecting your life or mana (aka Mind over Matter). If there still is any damage remaining to be applied, it then looks to see if there are any items or passives that indicate that a percentage of this damage should be removed from mana before life. Again, Mind over Matter is such an example. Any remaining damage is then removed from your life total.

TL;DR PoE processes damage by first checking to see if your character is hit (dodge and evasion mechanics), then if your character can avoid damage associated with the hit (block mechanic), then if you can shift damage from one form to another, then what damage is mitigated through physical and elemental (or chaos) defenses, and then this damage is applied to your "effective health".

Based upon successful characters I've created, studying other player's successful characters, and talking with other players about what makes their characters feel powerful, I've compiled the following lists of considerations and suggestions for helping you to plan your characters defenses. Keep in mind that these are just considerations--they are not absolutes or ways guaranteed to make sure your character is defensively prepared.

• It is more effective have a single, strong defensive mechanic that multiple weak mechanics

The central idea here is that despite there being multiple stages through which PoE checks to see how much damage is dealt to your character when they are attacked or have a spell cast at them, a single stage where there is a solid plan is more effective at dealing with damage than several weak plans in multiple levels. For example, a player might know that they will always eventually be hit, so they want to stack a lot of armor and resistances to mitigate damage that they are hit by. Having this large amount of armor and resistances makes it so when they are hit with damage, they can mitigate it effectively. This is in contrast to having a little bit of evasion, block, dodge, and armor.

• Several strong layered defenses are stronger than non-layered defenses

If you have a large amount of armor, say 12,000, then continuing to stack more and more armor will tend to yield diminishing returns. Rather, what becomes more effective is to begin building a second defensive layer, such as block. For example, max block Aegis Aurora characters focus on doing exactly this--obtaining a large amount of armor as well as obtaining max block. This creates a really tanky character because they prevent 75% of all the damage that they are hit by (aka, max block) and for the remaining 25% of damage that gets through, they have an effective mitigation mechanic--armor. This is why this particular type of character ends up being so strong.

• Account for random and entropy-based mechanics (i.e., evasion, dodge, block), as well as linear (e.g., resistances, endurance charges) and non-linear damage mitigation (e.g., armor)

One difficulty when planning defenses that are based upon entropy and/or randomness is that the stars can eventually align and they will all fail. This is one critique that is raised with evasion- and dodge-based characters--eventually these defenses will fail. This isn't to say that evasion- and/or dodge-based characters are "bad", but that there needs to be a contingency plan in place for when these defensive mechanics fail. One potential workaround of this eventual failing is to have a large life pool to be able to absorb any hits that eventually get though. So then the vast majority of hits are avoided, and the few that do get through can be absorbed before using health pots to get back to full health, so this process can be repeated.

• Look for items, skills, etc. that synergize well together

When creating your characters defenses, you want your choices to all work well together. The strongest characters have the largest amount of synergy between all of their item, skill, etc. choices. For example, an evasion-based character that uses enfeeble (especially on Blasphemy) and blind can feel exceptionally tanky, due to enfeeble and blind reducing the enemies accuracy rating (an a character's chance to evade is based upon its evasion rating and the enemies accuracy). Moreover, enfeeble reduces the amount of damage that enemies deal when they do eventually hit, which helps to circumvent one challenge evasion-based characters face--getting a lot of life to absorb hits that do eventually come through.

Full video explanation (~17 min)

Brief video explanation (~10 min)

In designing a build, there are two fundamental questions to be answered:

• Is this build a spell-casting or an attack-based build?

• Is this build a non-crit or crit build?

The combination of the answers to these two questions produces four major arch-types of builds: a) non-crit spell casters, b) crit spell casters, c) non-crit attack builds, and d) crit attack builds.

Disclaimer: Yes, I am aware this is not a perfect way to classify every possible build that could every be created in PoE (e.g., summoners don't fit nicely into this framework). However, for most beginning theorycrafters, this framework helps provide a decent classification of the major arch-types of characters that can be played. Additionally, skills such as traps and totems have been classified under the "spell" category, given that these skills are often (but not always) associated with the "spell" tag on the skill, and can be scaled as such.

Some of the benefits and drawbacks of each of these builds is described below.

Non-Crit Casters

- Spells are independent of accuracy--they always hit the enemy assuming the enemy is within rage of the spell's ability to hit.
- Non-crit builds provide the greatest gearing flexibility, given they do not need to account for accuracy, critical strike chance, or critical strike multiplier..
- The damage scaling for non-crit builds has a lower ceiling than for that of crit-based builds.
- Arguably the lowest potential damage compared to any other build arch-type. However, dependingupon the strength of the build and items being used, end-game content (i.e., top-tier, red maps) are still able to be done.

Crit Casters

- Spells are independent of accuracy--they always hit the enemy assuming the enemy is within rage of the spell's ability to hit.
- More potential damage than their non-crit caster counterparts.
- More item dependent than their non-crit caster counterparts.

Non-Crit Attackers

- Resolute Technique is nearly always used with these builds, providing more passive points that can be allocated for other uses on the passive tree (e.g., damage, life)
- Non-crit builds provide the greatest gearing flexibility, given they do not need to account for accuracy (if Resolute Technique), critical strike chance, or critical strike multiplier.

Crit attackers
- Crit attack-based builds have the highest potential damage scaling in the game.
- Crit attack-based builds have arguably the least gearing flexibility, due to the need for specialized items with the exact mods needed to achieve the highest damage.

One commonly misunderstood component of damage scaling is the different in language associated with the way PoE utilizes different damage scalars to compute damage. These often convoluted terms are "increased" and "reduced", along with "more" or "less".

The fundamental difference between these damage scalars is that increased and reduced modifiers are all added together, referred to as a sum, whereas all of the more and less damage modifiers are multiplied together, referred to as a product.

This distinction is absolutely critical, because increased/reduced modifiers follow additive/linear scaling, which is in contrast to more/less modifiers which follow multiplicative/exponential scaling. Unless the values being worked with are small, multiplicative/exponential scaling results in more damage growth compared to increased damage modifiers (see videos for an example illustrating this difference between multiplicative/exponential growth and additive/linear growth).

Multiplicative/exponential scaling with more/less modifiers is very desirable because it is the strongest modifier to improve your damage (or kill it in the case of a less multiplier). Yes, there are some extreme cases where this doesn't occur, but for nearly all builds, prioritizing stacking as many more modifiers is the most efficient method for scaling your damage, followed by then stacking as many increased damage modifiers.

Although DPS is often referenced, it is first important to identify how PoE calculates damage on a per hit basis. This occurs with the following formula:


Damage per Hit = (Base Damage of Skill or Item) * (Sum of all Increased/Reduced Damage Modifiers) * (Product of all More/Less Damage Modifiers)


It is worth noting that because multiplication is commutative, scaling by the increased/reduced modifiers sum or the product of the more/less damage modifiers can occur in either order. The result of this computation will still be the same.

In addition to addressing how damage it calculated per hit, damage per second (DPS) is also often referenced when describing a build's offenses. More specifically, it is often the average DPS that is referenced, not the low and top end of the range of values the DPS can fluctuate within. The average DPS calculation is determined as follows:


Average DPS = (Average Damage per Hit) * (Number of Casts/Attacks per Second)


Now, it is important to be careful with how DPS is interpreted, as more DPS doesn't necessarily mean a stronger offense. For example, skills such as Explosive Arrow or Caustic Arrow often have lower tooltip DPS because they are not scaled in the same ways as many skils. Explosive Arrow works to cap the number of fuse stacks on each enemy (5), so it's scaling its damage on a per fuse/hit (and often associated AoE overlap). Caustic Arrow often seeks to scale its damage through the DoT (Damage over Time) associated with the initial hit of the arrow. The tooltip DPS can currently only display this initial hit, but not the associated DoT of Caustic Arrow. Another example of this is for traps, which are often scaled on a per hit basis, as trap cool downs (and throwing speed) need to be accounted for.

The purpose of these examples isn't to discount DPS, but rather to serve as a caution to treating it as the most valuable or exclusive metric for measuring the strength of a build's offense.

Disclaimer: Yes, I am aware the formula given for the average DPS isn't exactly quite right. This is because it doesn't account for your chance to hit (assuming the build is non-Resolute Technique, or if for the times when a hit is a crit versus when it is a non-crit. It can be adjusted slightly to reflect this though using a weighted average. I purposefully didn't include this in an attempt to make damage calculations as accessible as possible for beginning theorycrafters.

• Stacking the largest number of "more" damage multipliers is critical to scaling your damage in an efficient manner. Similarly, avoiding "less" damage multipliers wherever possible is critical due to how much they kill your damage.

The reason it is often challenging to stack so many "more" damage multipliers is because of how strong they are in scaling your damage. These "more" multipliers are why skills, such as Righteous Fire, and the keystone Pain Attunement that is picked up on low life builds are so powerful (and why the combination of low-life, Righteous Fire is stronger yet still!). Simply put, scaling your damage to high levels in the most efficient manner possible requires the use of as many "more" damage multipliers as possible.

• Balance increased attack/cast speed with "more" and "increased" damage scaling modifiers

Depending upon the skill being used, and the plan to scale damage, you want to balance how slow, but hard, you are hitting enemies. This is in contrast to hitting enemies frequently, but for a low amount of damage. Beyond the thematic distinction between whether you want to hit slow but hard (e.g., burn-prolif builds like Flameblast, aftershock Earthquake builds) with those that hit fast but with lower damage values (e.g., Reave, Incinerate), an important consideration is whether or not you can support the increase in mana required to speed up the frequency of the hits. Although attack/cast speed is often exceptionally useful for scaling up a build's damage, and/or effective clear speed, the associated increase in mana required for the skill must be supported.

Another consideration related to this is that not every skill scales attack or cast speed in the same manner. For example, the skill Ice Crash has a 25% less attack speed modifier as part of the skill. Although attack speed is useful with this skill, you might first consider making sure you're killing mobs efficiently (white mobs in 1 hit, blues in 2-3 hits), given the high based damage of this skill. Scaling attack speed after this occurs then becomes quite efficient, as you're able to attack and move from pack to pack in quicker succession.

• Normal white mobs should die within one hit. Blue magic packs should die within two or three hits.

Although there is some variation in this, depending upon the skill being used and enemy being hit, these benchmarks are ideally what you want to try and aim for. If your character is needing two ore three hits to kill normal, white mobs and four, fire, or six hits to kill magic, blue mobs, then you'll want to pick up more "increased" and "more" damage modifiers to scale up the damage of each individual hit. If your character is one-shotting magic, blue mobs, then picking up increased attack/cast speed will increase your overall clear as you'll be able to move from pack to pack quicker.

• Account for a skills damage effectiveness if you're wanting to scale damage through added flat damage sources (e.g., gear, auras, heralds)

Another consideration for damage scaling is the effectiveness of added flat damage scaling. There are now many sources able to do this, ranging from items (e.g., added fire/cold/light damage to spells on wands, daggers, and scepters; added fire/cold/light/physical on rings, amulets, or gloves) as well as through Heralds, and auras (e.g.,Anger). Despite being able to be added to any appropriate skill, some skills are more effective for delivering this added damage, identified through a skill's damage effectiveness. Damage effectiveness is explicitly identified on each skill (e.g., any damage applied by Blade Vortex is applied at 30% of the damage's value). If this isn't explicitly stated on a skill, it is implied that the damage effectiveness of that skill is 100%.

It's worth noting that lower damage effectiveness values doesn't mean that they may not be efficient for scaling flat damage. Rather, it just means that each hit will apply the added damage at the indicated damage effectiveness value. So if the goal is to scale flat added damage, increasing the number of times the skill hits can be one method for scaling this added damage, even if it is using a skill with a lesser added damage efficiency. It is also worth noting that this added flat damage is scaled through other damage modifiers (e.g., increased spell damage, more elemental damage).

Full video explanation (~17 min)

Brief video explanation (~11 min)

Passive tree build demonstration

Before you begin planning your passive skill tree itself, there are several considerations and plans to already have laid. Each is identified below.

What is the "big idea" driving your build?

Prior to designing your passive skill tree, you need to have identified a "big idea" for your build. Essentially, what is the theme or flavor of your character. After looking across many builds and talking to a variety of players, I identified three common places where identifying this "big idea" seems to emanate from.

• A novel idea

Sometimes a build is designed around a new idea that hasn't been realized before. For example, realizing that The Taming's "10% increased Damage per Freeze, Shock and Ignite on Enemy" interacts with Emberwake such that you can stack up to 300 ignites on an enemy (and gain 3000% increased damage!) would be an such an example.

• An interesting item or skill

Sometimes the "big idea" for a build comes from a newly released item or skill gem. For example, the sword Varunastra is a unique that allows all one-handed melee weapon types (e.g., daggers, claws, maces, swords, axes) all to scale its damage. Given that there are several clusters whos damage scaling applies to multiple weapons (e.g., By the Blade, Deadly Dilettante), this can allow for some powerful damage scaling with minimal damage scaling investment.

• A new take or tweak on an existing build

Sometimes the "big idea" for a build isn't something as groundbreaking as the two components identified above, but is simply a re-envisioning of a current build with a new take or tweak on it. For example, Flame Totem builds have existed for quite some time. However, many appear to utilize one-handed weapons (e.g., wand, dagger, sceptre) and a shield, or dual wielding two, one-handed weapons. A character that planed to use a +3 fire staff and Empower as a means to scale the damage with the gem might serve as an example that falls under this third category.


What is the plan to scale your damage offensively?

Designing a passive skill tree also requires having a plan for how the build works offensively. Does it scale through physical or elemental damage nodes? Chaos damage? Attacks or spells? On-hit effects? Is the damage on hit or a damage over time? Having a plan for how you want to scale damage is critical as it informs what passives are to be acquired on the passive tree.


What is the defensive plan to keep your character alive?

Similar to have an offensive plan, it is critical to have a plan for scaling your character's defenses. Is the plan to avoid damage from hits (e.g., evasion, dodge, block), to anticipate being hit and mitigate damage from those hits (e.g., resistances, armor, fortify), or to just have a deep life pool and absorb hits and pot/leech through them? What components, if any, will be layered to make for a tougher, tankier character? You want to have a plan for what will keep your character alive as it will influence the passives you take on the skill tree.


What unique items, if any, are you planning on using in your build?

Considering any unique items that you plan to use in your build is another critical question to consider. This is because some items may influence what kinds of stats you are trying to acquire from the passive skill tree. For example, if you were planning to use Belly of the Beast, you might not include as many passive points on the skill tree into % increased maximum life, as this chest provides a 30% to 40% increased maximum life.


What skill gem and potential support gems are you planning on using?

It also helps to have considered which skill and support gems the character will be using prior to planning your skill tree. Being familiar with the tags on a skill and which components of the skill they apply to is critical to knowing what kind of damage you plan to scale. For example, Molten Strike is an attack that causes 3 projectiles to be launched following the melee attack portion of this skill. It has the tags of Projectile, Attack, AoE, Melee, and Fire. Yet all of these tags do not apply to all portions of the skill (e.g., the melee tag applies only to the initial melee hit, but not the projectiles then launched). So picking up damage scaling with the word "melee" will only scale the damage of the initial hit, but not any of the hits from the projectiles launched.

Similarly, this can also influence the support gems utilized. Which portions of the skill are being augmented or scaled? Additionally, does the support gem itself provide more, increased, or no damage scaling. The fewer more multipliers acquired from support gems (or gems that scale damage in a multiplicative manner), the larger amount of damage will need to be acquired from the tree. Greater importance is also then placed on acquiring "more" multipliers from the tree whenever possible (e.g., Elemental Overload, Point Blank).


TL; DR - Before you plan your passive tree itself, you need to have a "big idea" driving your build, have a plan for scaling your character's offenses and defenses, and considered any special items (e.g., unique items) and gems that you plan on using.

I have found three questions particularly helping to guide my passive tree planning. These questions are listed below.

• What stats are you trying to acquire?

Are you interested in % life or % ES? Are you looking to acquire % increased physical damage with claws, or % weapon elemental damage? Before you can plan your passive tree, you have to know which kinds of stats you're wanting to path towards and pick up on the passive tree itself.

This first question is also closely related to the second question, which is

• How much of each stat are you trying to acquire?

One issue I've commonly seen with responses to this question is that the response is "as much as possible." The reason this is a problem is that it doesn't indicate when you should stop trying to acquire that stat from the tree to focus your efforts on allocating points on other stats. For example, if you identify that you want to get % increased life from the tree, how much of this stat do you need? Is 100% enough? 150%? 200%? 250%? 300%? Without a target, there is not indicator for when you should stop spending points investing into a particular stat.

The third question is

• Which stats are non-negotiable, and which are strongly desired?

When identifying the kinds of and how much of the stats desired for a build, you may find that you're not able to acquire all of the stats and/or as much of these stats as desired. Therefore, you'll have to make some decisions about what stats to keep, and which stats to reduce and/or cut. For example, a keystone is likely a stat this is a non-negotiable, because a build wouldn't function in the manner designed if it was not acquired (e.g., Resolute Technique builds w/o this keystone would be pretty bad, due to the low accuracy marauders frequently have). In contrast, the stat % increased physical damage with maces might be only strong desired, given that it can be made up for in other places (e.g., a strong weapon, jewel stats).

The purpose of providing these benchmarks is to help give beginning theorycrafters targets to aim for when determining how much offensive and defensive stat scaling to acquire from the passive tree.

The benchmarks provided are NOT absolutes, and my vary depending upon who is providing the information (e.g., someone who builds super defensive characters versus someone who builds super offensive characters), the skill being used (e.g., Earthquake versus Ground Slam), gem links (e.g., 3-link versus a 6-link), and the gear being used (e.g., Kaom's Heart versus a rare chest).

It gets even messier when you begin to consider the variety of stats that can be acquired through jewels. With this being said, I've separated these potential benchmarks by offense and defense below.







It is also important to note that the use of stat benchmarks is not the only approach that exists to designing passive trees. A different example would be based upon the distribution of skill points into offensive nodes, defensive nodes, and stat/pathing nodes (e.g., 1/3 of the total available points into each category).

Another common question asked when designing a passive skill tree is whether or not to allocate points towards a particular jewel socket. Although this depends as a whole, I frequently use the following criteria to help decide whether or not it is worth spending points for the jewel.

• 2 passive points for the jewel socket

Yes, this is a jewel socket worth picking up. Even a blue two property jewel will often be more efficient in improving your stats than allocating the two points for the jewel socket elsewhere on the tree. This benefit is further amplified when three and four property jewels are acquired and can be placed into the socket. Needing to use a particular unique jewel (e.g., Rain of Splinters) also makes the two passive points the most efficient method for being able to use this unique jewel.

• 3 passive points for the jewel socket

Whether or not points are best spent for the jewel varies--sometimes yes, sometimes no. Typically, in order to make this skill investment worth it, the jewel needs to have three or four strong stats that improve your build, or be an essential unique jewel required to make your build's "big idea" work.

• 4 (or more) passive points for the jewel socket

In nearly every case, spending four or more points to acquire a jewel socket is not worth it. The four (or more) points can typically be allocated in different locations in the passive tree that would provide a stronger stat boost than on acquiring the jewel socket (not to mention that the jewel would have to be an insanely strong four property jewel). The only exception to this that I can currently think of is if you're trying to stack a particular statistic that is difficult to acquire on the passive tree (e.g., shock or freeze chance). However, it's probably worth trying to acquire this on gear rather than spending points to path into a jewel socket. Play testing can also be used to see if the marginal boost the jewel's stats would provide are worth it.

Essentially, the idea with pathing efficiency is to determine how much of each stat, on average, is acquired with each passive point allocated towards or into a cluster that is desired on the passive skill tree. This is calculated by determining the total amount of each stat that is acquired from allocating all points into the cluster, divided by the total number of passive points needed to be allocated towards and into that cluster. The purpose of this metric isn't to make the decision for you, but to provide you with information that you can use to help you make the most informed decision based upon the needs of your build.

To illustrate, consider the example below.

Example

Imagine you are playing a totem build that paths past the Ironwood and the Shamanistic Fury notable clusters. Both are totem clusters, and both require three points to spec into. Utilizing the idea behind pathing efficicency, we can determine how much of each stat, on average, we're acquiring by specing into each cluster.

• The Ironwood cluster provides 38% totem damage, 3% cast speed, 10% totem placement speed, and a few other stats (e.g., totem elemental resistances). Given that it takes three passives points to acquire these stats in total, each point spec'd into this cluster provides, on average: 12.7% totem damage, 1% cast speed, and 3.3% totem placement speed

• The Shamanistic Fury cluster provides a total of 48% totem damage and 15% totem placement speed. This translates to 16% totem damage and 5% totem placement speed per point invested into this cluster, on average.

Which cluster you take is dependent upon the stats you are trying to acquire for your build. For example, a Searing Bond build would be better served by the Shamanistic Fury cluster, as cast speed doesn't effect Searing Bond and the totem damage is a larger value on a per point invested basis (as well as totem placement speed). In contrast, a Flame Totem build might be more interested in the Ironwood cluster, depending upon the relative increase in DPS the 3% cast speed provides when contrasted with the 10% totem damage missed out on in the Shamanistic Fury cluster. To make this decision about which cluster is better for your build, you have to compare the stats provided by each cluster. Comparing them on a per point invested basis allows you determine the "efficiency" of pathing to and into clusters, as well as for making comparisons across clusters that may require different amounts of points to spec into.

In practice this an be a little messier, but I purposefully chose this example in an attempt to keep this concept as "clean" as possible. Sometimes it may be worth taking the slightly less efficient cluster because str/dex/int stats are needed that are acquired on the way to the cluster. Additionally, this inefficient pathing in the short term may serve the build better in the long run by positioning it to acquire other clusters and/or keystone(s) that make up for this short-term inefficiency.