Dungeons & Dynasties tracks four different ideas that look like one growth system: Current Ability, Potential Ability, individual attributes and Veterancy. Training feels broken when you watch the wrong number. Pick the target first; the assignment follows.
Want the numbers for a specific hero? The training calculatorshows ability per cycle, the age cap and when each star arrives.
Checked against v3.0.0
Primary target: Current Ability
A hero has room below Potential and needs broad growth.
Moves CA toward PA, with the points landing on the attributes the calling fights with. At the ceiling, Train refines: now and then a point moves from the least-weighted attribute to the most-weighted one.
Primary target: Chosen attributes
The hero is broadly capable but weak on one job-critical stat.
Works the chosen attributes up to the hone ceiling: the class ceiling plus the kindred leaning (a Dwarf’s Vitality goes three past a Human’s). Training orders queue stats with target numbers, top row first.
Primary target: Current Ability + teaching
A suitable veteran can teach a developing hero without wasting the match.
Grows CA with mentor quality and class or role fit; rising pupils can also gain Potential. A pupil delving beside their mentor learns 25% more.
Every at-home roster hero below Potential receives daily passive CA equal to0.12 × Training Ground level × staff training multiplier. Active work is separate: the yard begins with two assignment slots, every facility level adds two more, and each level makes assigned work 15% more effective. At the top level the wing takes a specialisation: the Sparring Ring adds two more benches (ten in all), or the Master's Yard lifts bench work from 1.45× to 1.7×. Work settles on the six-day cycle. Recover and Rehearse use no bench; Hone can use two. IfAuto-train idle is enabled, the game preserves manual choices, fills urgent recovery first, then assigns Train where CA has room below PA.
A missing passive Master-at-Arms reduces daily passive growth outside the Backwater, by 30%, 24%, 18%, 10% or 4% from the top division down. The assigned-work calculation floors the staff multiplier at 1, so that vacancy penalty does not reduce assigned work in the same way. A Master-at-Arms stops adding training speed at +90%; the Old Swordmaster guildmaster adds +40% to all Yard work on top.
Count the open active-work slots.
Assign the heroes whose target matters now.
Check automation before changing a surprise assignment.
A hero already near PA has little CA left to gain. A rising pupil can gain Potential through mentoring. Otherwise, check maturity and decline as well as PA before buying more training.
Growth pace is modified by the hero's stage and adaptability, so equal-looking assignments need not settle equally.
Veterancy rises to a 1.25× learning modifier at Veterancy 20. Diligent and Professional each add 1.1×, and a Late Bloomer learns at 0.8× early in their career and 1.25× later. Staff training quality contributes separately.
Hone is straightforward below 15. Every point beyond 15 costs 1.6× the last, and personal attribute ceilings still apply. Since 2.16 honing runs 50% faster overall, but growth from delves is smaller than it was before 2.12, so the Yard now carries more of the work.
A mentor's Veterancy and Leadership shape teaching quality. Matching the student's exact class is best; matching only the role is neutral; crossing into an unrelated role is much less effective. Kindred teaching modifiers and the Old-School, Broad Hand and Exacting traits also affect the pairing. Young Train pupils can receive automatic mentoring where teaching capacity is available. The student's own Veterancy still helps them learn. Mentor is therefore a roster pairing decision, not an automatic upgrade over Train.
Each tutor has a class size: one pupil at one or two teaching stars, two at three, three at four or five, and two more for whoever holds the Mentor post. A pupil whose tutor is full goes to the next best at home, and the Mentors tab in the Yard lets you pin a tutor. A mentor only teaches while home, and a tutored delver never learns slower than Train alone.
For the complete state thresholds behind recovery and roster availability, seeHero state & training.
