Scoring sources

How these tests are scored

Dialed Color is a game benchmark, not a diagnosis. Some measurement units come from color science, audio engineering, timing research, geometry overlap metrics, and testing standards. The final 0-10 round score is a product-calibrated game curve so results feel readable after five rounds.

Important Generated targets are deterministic by seed. A fixed seed gives the same target set; it is not a clinical sample, population norm, or hidden personality measurement. Reports can explain observed run behavior, but they must not overclaim beyond the measured task.
color

Color memory

Reference implementation

The audited reference color game uses HSB values, converts them to RGB and CIE Lab, then scores CIEDE2000 Delta E with a hue-sensitive product curve. Dialed Color matches that measurement model for the random HSB color game; the reference site's separate logo/character Color2 dataset is a curated-content task, not a random target task.

Target source

Solo and challenge targets are seed-generated pseudo-random HSB colors. Daily targets use date-based fixed palettes or date hash offsets, so the same seed gives the same colors.

Measurement

The app converts target and submitted colors into RGB and CIE Lab coordinates, then measures color difference with the CIEDE2000 Delta E formula.

Score curve

The round score maps CIEDE2000 color difference onto a 0-10 assessment score with a reference-style product curve. Delta E is standardized; the 0-10 curve is not a CIE standard.

Report use

A report can discuss hue drift, saturation/brightness drift, average Delta E, and consistency across rounds. It should not claim clinical color vision, brand-color literacy, or long-term memory ability from one five-round run.

Limits

Screen calibration, brightness, color profile, ambient light, and the user's display all affect results.

sound

Sound memory

Reference implementation

The reference sound task is a pitch-recall game: play a browser-generated tone, let the user drag vertically to choose a remembered tone, then score perceptual pitch distance. Dialed Color uses the same task framing with local Web Audio synthesis, cents, and normalized ERB-rate distance.

Target source

Targets are seed-generated pseudo-random browser tones inside a fixed frequency range. The same seed and difficulty reproduce the same target tones.

Measurement

The app compares target Hz and submitted Hz with cents plus ERB-rate distance, so pitch error is measured on logarithmic and perceptual scales instead of raw hertz alone.

Score curve

The round score maps ERB-rate distance onto a 0-10 assessment score with a product-calibrated curve. It is not a medical hearing score.

Report use

A report can describe average cents error, whether guesses trend sharp or flat, and whether low/mid/high tones are harder. It should not diagnose hearing acuity or pitch disorders.

Limits

Headphones, speaker quality, volume, browser audio behavior, background noise, and hearing range can all change performance.

time

Time memory

Reference implementation

The reference time task uses deterministic intervals, visual hints on easy mode, and a log-ratio timing score. Dialed Color now uses the same interval ranges and log-ratio curve, while keeping the brand and report language separate.

Target source

Targets are seed-generated pseudo-random time intervals. Easy rounds draw 1000-3500ms intervals; harder rounds draw from 500-6000ms with long intervals spread across the run.

Measurement

The app compares target duration and submitted duration using absolute error, relative error, and Weber fraction-style signed error.

Score curve

The round score maps log-ratio timing error onto a 0-10 assessment score. The log/ratio basis follows timing psychophysics; the exact 0-10 curve is product calibration.

Report use

A report can discuss early/late bias, average absolute error, relative error, and Weber-fraction-style consistency. It should not claim attention disorder, neurological status, or reaction-time diagnosis.

Limits

Input device latency, browser frame timing, attention, audio/visual lag, and refresh rate can change timing results.

shape

Shape memory

Reference implementation

The reference shape task compares the user's recreated geometry against a generated target with overlap, rotation, position, and scale behavior. Dialed Color uses a raster-overlap and symmetry-aware rotation model to make those errors visible.

Target source

Targets are seed-generated circles, polygons, and stars from the reference side sequence, with controlled position, rotation, and scale per difficulty.

Measurement

The app compares target and submitted geometry with a 64x64 raster overlap score, then applies symmetry-aware rotation and scale penalties.

Score curve

The round score maps combined geometry error onto a 0-10 assessment score with a product-calibrated curve.

Report use

A report can describe outline overlap, center drift, scale drift, and rotation drift. It should not infer visual-spatial disorder or motor impairment.

Limits

Pointer precision, touch-screen size, motor control, viewport size, and visual attention can affect results.

SBTI

SBTI

Reference implementation

The SBTI task is a structured self-report assessment layer: questions map answers into 15 SBTI dimensions, a nearest archetype label, and Big Five-informed OCEAN tendencies. This requires transparent wording and strong non-diagnostic disclaimers.

Target source

Questions and type labels come from the authorized sbti-test.com content package, not from random generation.

Measurement

The app maps answers into SBTI dimension patterns and directly scores Big Five-informed tendencies from selected answer evidence. It is not an official IPIP, BFI, NEO, or clinical personality instrument.

Score curve

The match score maps answer patterns onto a 0-10 style assessment score and share result. It is not a psychometric validity score.

Report use

A report can summarize answer patterns, strongest dimensions, weakest dimensions, nearest SBTI archetypes, and Big Five-informed OCEAN tendencies. It should not make clinical, hiring, school, or relationship decisions.

Limits

Self-report mood, joke wording, social context, and intentionally exaggerated answers can change the result.