A little science.
A little magic.

Your measurements, everyday abilities, and choices become the foundations of your character.

Strength combines grip and lifts

Grip and lifting capture different parts of Strength. When you provide both, lifts contribute 70% and grip contributes 30%. The grip score sets a floor. A grip score of 10 and a lift score of 18 produce Strength 15.6.

With a lift and behavioral answers, the lift contributes 90% and the behavior average 10%. A lift by itself supplies the full score. Grip by itself uses its measured reference. The sheet keeps each measurement and its calculation in your field notes; the grip percentile stays attached to grip.

Higher-repetition sets count

Enter any positive whole number of repetitions. A single uses the recorded load. Sets of 2–10 use the Epley estimate: load × (1 + reps/30). Above 10 repetitions, the multiplier increases smoothly with diminishing returns, approaching twice the recorded load:

Multiplier = 4/3 + 2/3 × (1 − exp(−(reps − 10)/30))

The high-repetition calculation is a scoring rule for the app. It keeps higher-repetition sets useful while emphasizing load as sets get longer. Exercise factors translate squat and bench records into the same scoring table as deadlift. Your field notes show the recorded load, reps, adjusted load, and resulting contribution.

A scale centered on 10

Reference percentiles convert to character points using a continuous scale, with a population midpoint of 10 and three points per standard deviation:

Score = 10 + 3 × inverse normal(percentile ÷ 100)

Scores extend beyond 18 and retain one decimal place. Each percentile uses its named reference population. NIST describes the normal distribution and inverse function.

Game modifiers use floor((round(score) − 10) ÷ 2). Reference endpoints use bounds such as ≥17; the original measurement stays visible.

Your evidence

Measured reference: a supported measurement compared with a research sample. Reported norm: an existing percentile or standardized cognitive result, supplied with its source. Estimate: points calculated from behavior observations, activity records, or a combination of measurements. Percentiles appear beside their specific reference measurements.

For Constitution and Intelligence, a sourced percentile takes precedence over activity and behavior points. A confirmed standardized cognitive result supplies Intelligence when there is no reported percentile. Unanswered abilities remain open.

The six abilities

Strength: muscular force. Dexterity: coordination, precision, balance, and timing. Constitution: sustained effort. Intelligence: reasoning and learning. Wisdom: awareness, judgment, and reflection. Charisma: communication, connection, and influence.

Each ability has four behavioral observations. Their five choices map to 6, 8, 10, 13, and 16; answered observations are averaged. Optional reaction and reasoning tasks contribute 20% when other game evidence exists.

Measurement references

Grip strength: 10,346 complete adult tests from NHANES 2011–2014. We sum the best of three trials per hand and pool examination weights across both cycles. The comparison includes adult participants across ages and recorded sex. The 1st and 99th percentiles set the displayed tail bounds. CDC protocol and data, second cycle, and survey weights.

Endurance: existing 5 km times and VO₂ max results contribute through activity point tables. With behavioral answers, the activity average and behavior average contribute equally. A percentile from your own assessment can supply a reference comparison instead. FRIEND publishes laboratory cardiorespiratory reference standards.

Cognitive scores: a standardized mean-100 result uses its reported standard deviation: 10 + 3 × (reported score − 100) ÷ SD. At SD 15, scores of 100, 145, and 160 become 10, 19, and 22. The source and test ceiling remain attached to the result. Publisher extended-norm documentation.

Field tests: reaction timing uses the median of five valid trials. The three reasoning puzzles use your first answers and contribute points from 6 to 15.

From your life to your legend

Eight dilemmas explore order and care. Four choices inform each axis, with values from −2 to +2. A mean at least 0.6 or at most −0.6 selects the corresponding pole; the middle stays neutral. Each axis needs two answers.

Your work, interests, appearance preferences, and choices inspire the AI’s class, origin, talents, equipment, ideals, bonds, flaws, and quests. Age and body measurements supply context, and an optional photo guides the portrait. Ability scores come from the assessment; the AI builds the character and combat details around them.

Every sheet records its scoring version. The repository includes the point tables, source data, and calculation tests.