Level 2+ ยท talent planning
Talent Points and Weapon Branches
Eight points, a level cap of twelve and thirteen exclusive choices decide your talents, and branch damage only counts while the matching weapon is drawn.
Eight points is the entire budget
The talent screen shows 84 trees and 241 selectable nodes, which reads like a system you could sink a hundred hours into. Your actual budget is eight points. The character level cap is 12, and points arrive at levels 2, 3, 5, 6, 8, 9, 11 and 12. Levels 1, 4, 7 and 10 grant none. A fully levelled character has spent eight points out of a few hundred available nodes, and that ratio is the single most important thing to understand about the system.
The dry spells matter as much as the payouts. Levels 4, 7 and 10 give you nothing on this screen, so if you were planning around "one more level and I can take that node", count the ladder first. Level 10 in particular is a long stretch that hands you skills but no talent, and level 11 and 12 arrive together at the top of the cap.
Because the budget is so small, most planning failures are not about picking the wrong node; they are about picking eight nodes that do not point at the same thing. Six good talents spread across three weapon branches lose to four talents stacked behind one action you press. Open the talent planner with a specific weapon in mind rather than browsing for interesting names.
Every current node is a single rank
There is no rank stacking in the current trees. Every node you can select is one rank, taken once, and it does nothing more if you look at it again. That removes a whole category of planning that other games train you into: you never need to work out whether a fifth point in a node is worth more than a first point somewhere else. You take it or you skip it.
Prerequisites do exist, but only on a minority of nodes; 49 of the 241 require a specific earlier node before they open. Those chains are where your budget disappears fastest, because a two-node chain costs a quarter of everything you will ever have. Before you commit to a deep node, trace the path back to the branch entry and count it.
Branches themselves gate on character level, with entry requirements ranging from 2 up to 12. What no branch currently does is demand a minimum number of points already spent, so you are never forced to buy filler to reach a gate. If you meet the level, the branch is open. That makes the trees much more permissive than they look, and it means a level 12 character can legitimately dip into a branch they ignored for the entire climb.
Branch damage follows the weapon in your hands
The rule that settles most arguments is stated plainly by the developers: increased damage from weapon-specific skills and talents only takes effect while you are wielding the corresponding weapon. Not while it is in your other slot. Not averaged across both. While you are holding it.
That converts a vague preference into an arithmetic problem. A Mercenary who carries Sword and Shield for gap closing and swaps to the Hammer for damage is a common and effective pattern, but the talent points only pay out during the half of the fight where the matching weapon is drawn. Splitting eight points between two weapon branches means each half of the fight runs on roughly four points of investment. Committing all eight to the hammer means the sword phase runs on none, which is fine if you only use the sword to close distance and block.
Both answers are defensible; the mistake is not knowing which one you chose. Decide up front whether the second weapon is a damage tool or a mobility and defence tool, and spend accordingly. The class pages list which weapons each class can hold, and the skills attached to each are covered in how weapons decide your skills.
Thirteen exclusive choices, and what they cost
Thirteen points in the trees present an exclusive choice: take one node and the alternative closes. Sword Edge or Shield Buildup. Fearless or Foresight. Stonebreaker Counter or Sudden Edge. Chant Guard or Masterful Technique. Measured Pace or Battle-fed. Shadow Stride or Frequency Shift. Battle Frenzy or Wound Buildup. Bloodthirst Frenzy or Shadow Dance. Reshape or Potent Wounding. Inexhaustible or Ferocity. Crime or Reprieve. Momentum Shift or Counter-Impact.
With eight points, an exclusive pair is rarely a painful loss on its own, because you were never going to afford both anyway. What it does change is the shape of a branch: an exclusive choice usually sits at the point where a branch decides what it is for, so picking one side tells you which of the remaining nodes are worth reaching. Read the pair, decide what the branch is doing for you, and let that decide the nodes below it.
Since talents can be reset between runs at no cost, an exclusive choice is not permanent; it is a per-run declaration. The one place you cannot change it is inside the training ground, which is exactly where you would want to test both sides, and it is a long-standing complaint. Plan the swap before you enter, not after.
The Withered Knight three-way choice
The question players ask most often about this class is whether one character can genuinely run both the Greatsword and Polearm and Shield. Official communication has pulled in two directions: one developer note describes a Withered Knight flexibly choosing skills and talents from both branches, while a reply about the Seer states flatly that Mace and Catalyst are separate branches and you cannot use both weapons and their abilities within a single run.
The talent data settles it for the Withered Knight. Greatsword Specialization, Endless Flow and Polearm and Shield Specialization form a single three-way exclusive choice; you may take exactly one of the three. So you can spread ordinary points across both weapon branches, and you can carry both weapons into a raid, but you cannot hold two specializations at once. Endless Flow is the middle option that exists precisely for players who refuse to commit to either weapon.
If you intend to fight most of a raid with one weapon, take its specialization and accept that the other weapon is a situational tool. If you swap on a timer, applying Wither with light strikes and consuming it with heavy strikes on one weapon before rotating out, the middle option fits that pattern. Compare the branch layouts on the Withered Knight page before you spend.
Two branches everyone shares
Two branches belong to all six classes rather than to one. Woodling Expert opens at character level 10 and holds three nodes; Way of Gylden Hunt opens at level 12 and holds four. They exist outside the weapon question entirely, which means they are the only talents that never switch off when you change weapons.
Their timing is the interesting part. Your last two points arrive at levels 11 and 12, exactly when these shared branches become reachable, so the endgame decision for every class is the same: do the final points go into the weapon branch you have been building all game, or into a shared branch that applies everywhere? There is no universal answer, and the effect magnitudes have never been published, so judge them by what the node text says it does for your routine rather than by any number you see quoted.
What is worth avoiding is treating them as automatic. A level 12 character has eight points; spending two of them on shared branches means the weapon branch you fight with runs on six. That is a real trade, not a freebie.
The trees are not the same size
Node counts differ noticeably by class. Seer has the most unique current nodes at 56, followed by Mercenary and Withered Knight at 49 each, Shadowstrix at 47, Sorcerer at 39 and Blackarrow at 36. That does not make Seer stronger. Every class still spends eight points. It does mean a Seer player has more genuinely different plans to choose between, and a Blackarrow player will find the same nodes coming up again in every discussion.
The internal split matters too. Blackarrow's ten branches are all built around the Bow, which reflects the fact that its second combat path is not a separate weapon family. Seer's branches divide along the Reverent and Blasphemer paths, matching the oath you switch from the skills panel, and that split is the reason the developer answer about Mace and Catalyst is so blunt.
Node magnitudes are not published anywhere reliable, so ignore any table claiming exact percentages. Judge a node by whether it touches something you do constantly. A talent that improves an action you press twice a raid is worth less than one that improves your opening sequence, whatever the numbers behind them turn out to be.
Turning this into a plan
Start from the weapon, not the tree. Write down the three or four actions that make up your normal fight: the opener, the main damage action, the escape and the defensive button. Then walk the branch attached to that weapon looking only for nodes that touch those actions. Eight points fill up faster than you expect once you apply that filter.
Then check the chain. If any node you want sits behind a prerequisite, add that prerequisite to the count immediately. If any of your picks sits on one side of an exclusive pair, confirm the other side is not serving a different plan you also wanted. If you carry two weapons, decide which one the points are for.
Lay the spend out in the talent planner, then check it against the gear and gems that have to agree with it; the whole chain is laid out in how the build system fits together, and the socket side of it in affix gems and socket rules.