# Lamp rating and anti-rating

Alpha Research Project · DRAFT · Working reconstruction · 16 September 2026

This section brings together source data, the principles of the rating and anti-rating, and a reproducible technical selection. The rating will evolve with measurements, testing and feedback. The next stage is an individual ranking for each person and situation.

## Recovered context and new calculations

Alpha Research Project working materials dated 26 May 2026 refer to a positive list of 10 entries, an anti-rating of 5 entries and a database of 4,431 records. The historical PDFs are unavailable for verification. Their exact membership, order and numeric thresholds remain unconfirmed. The tables below were recalculated on 16 September 2026 from the current LampTest export using the explicit editorial rules below. This reconstruction is for discussion; equivalence to the May edition is not claimed.

## 1. Source data and provenance

Measurements: LampTest, Alexey Nadezhin. Analytical framework: Alpha Research Project. The original Windows-1251, semicolon-delimited export is preserved unchanged. Retrieved 16 September 2026. Each sample has its own test date. Identical model names or barcodes can cover different batches or constructions.

The dataset contains 4,325 LED records, 82 incandescent, 28 compact fluorescent and 3 metal-halide records. Current tables include LED sources with a specified socket; base=no, other and empty values are excluded. Strips, some fixtures and other technologies remain in the source file. Comparison concerns measured samples.

Source: https://lamptest.ru/led.csv

Snapshot SHA-256: 70212ebff29af7052f2035426f24e22c983b1b89620a93338b53c91102d2ef4b

## 2. Positive rating: first 10 samples

| No. | Sample and date | CRI / Ra | R9 | Flicker, % | Flux, lm: measured / declared | Flux, % | LampTest, 0–5 |
| --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | [Navigator Supervision NLL-A60-9-230-4K-E27-FR-SV, ID 3959, 29.11.2021](https://lamptest.ru/review/03959-navigator-supervision-nll-a60-9-230-4k-e27-fr-sv/) | 98.4 | 97 | 0.3 | 957 / 900 | 106.3 | 5 |
| 2 | [Наносвет L293 LE-GX-D-6/GX53/940, ID 3761, 10.03.2021](https://lamptest.ru/review/03761-nanosvet-l293-le-gx-d-6-gx53-940/) | 98.3 | 98 | 0 | 571 / 540 | 105.7 | 5 |
| 3 | [Navigator Supervision NLL-A60-9-230-4K-E27-FR-SV, ID 3958, 29.11.2021](https://lamptest.ru/review/03958-navigator-supervision-nll-a60-9-230-4k-e27-fr-sv/) | 98.3 | 96 | 0.3 | 951 / 900 | 105.7 | 5 |
| 4 | [Наносвет L012 LH-MR16-50/GU5.3/940, ID 3759, 10.03.2021](https://lamptest.ru/review/03759-nanosvet-l012-lh-mr16-50-gu53-940/) | 98.2 | 98 | 0.6 | 481 / 500 | 96.2 | 5 |
| 5 | [Наносвет L018 LH-MR16-50/GU5.3/940/60D, ID 3756, 10.03.2021](https://lamptest.ru/review/03756-nanosvet-l018-lh-mr16-50-gu53-940-60d/) | 98.1 | 97 | 0.8 | 512 / 500 | 102.4 | 5 |
| 6 | [Фотон LED FL P45-C 7W E14 4000K, ID 3985, 22.12.2021](https://lamptest.ru/review/03985-foton-led-fl-p45-c-7w-e14-4000k/) | 98.1 | 96 | 0.8 | 687 / 560 | 122.7 | 5 |
| 7 | [Фотон LED FL P45-C 7W E14 4000K, ID 3984, 22.12.2021](https://lamptest.ru/review/03984-foton-led-fl-p45-c-7w-e14-4000k/) | 98.1 | 95 | 0.7 | 683 / 560 | 122.0 | 5 |
| 8 | [Наносвет L095 LH-GLS-100/E27/940, ID 3732, 10.03.2021](https://lamptest.ru/review/03732-nanosvet-l095-lh-gls-100-e27-940/) | 98 | 95 | 0.3 | 1117 / 1100 | 101.5 | 5 |
| 9 | [Наносвет L092 LH-GLS-75/E27/940, ID 3735, 10.03.2021](https://lamptest.ru/review/03735-nanosvet-l092-lh-gls-75-e27-940/) | 98 | 95 | 0.4 | 916 / 900 | 101.8 | 5 |
| 10 | [Navigator Supervision NLL-G45-6-230-2.7K-E27-FR-SV, ID 3953, 29.11.2021](https://lamptest.ru/review/03953-navigator-supervision-nll-g45-6-230-27k-e27-fr-sv/) | 98 | 91 | 0.4 | 542 / 540 | 100.4 | 5 |

## 3. Anti-rating: first 5 red-zone samples

| No. | Sample and date | CRI / Ra | R9 | Flicker, % | Flux, lm: measured / declared | Flux, % | LampTest, 0–5 |
| --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | [APIS G4 3W 220V, ID 805, 20.05.2016](https://lamptest.ru/review/00805-apis-g4-3w-220v/) | 65.3 | -49 | 98 | 70 / 300 | 23.3 | 0.1 |
| 2 | [APIS G4 5W 220V, ID 804, 20.05.2016](https://lamptest.ru/review/00804-apis-g4-5w-220v/) | 63.1 | -57 | 98 | 132 / 500 | 26.4 | 0.2 |
| 3 | [noname Corn 5.5W, ID 28, 04.02.2014](https://lamptest.ru/review/00028-corn-55w/) | 62.4 | -51 | 72 | 133 / 450 | 29.6 | 0 |
| 4 | [noname Yaloo E14-120SMD3528WW, ID 27, 04.02.2014](https://lamptest.ru/review/00027-yaloo-e14-120smd3528ww/) | 64.3 | -48 | 75 | 221 / 700 | 31.6 | 0.2 |
| 5 | [noname Corn COB 8W, ID 24, 04.02.2014](https://lamptest.ru/review/00024-corn-cob-8w/) | 66.3 | -50 | 97 | 267 / 800 | 33.4 | 0.2 |

## 4. Methodology of this working edition

### Original ARP framework

Earlier work considered measured and declared flux, power, CCT, CRI/Ra, R9, available Rf/Rg, flicker, practical suitability and specification accuracy. Weak flux, high flicker and poor colour rendering informed the anti-rating. The complete historical formula requires checking against the PDFs.

### Comparison unit

One LampTest record with its ID, model, test date and production marking. Repeated names remain separate tests. Different sockets and forms appear in the general technical selection; individual selection will require compatible products.

### Explicit positive rules in this reconstruction

CRI ≥ 90, R9 ≥ 50, flicker ≤ 5%, measured flux ≥ 90% of declared; all four values known. These editorial thresholds are open for discussion and do not establish universal suitability.

### Positive ordering

CRI descending, R9 descending, flicker ascending, flux percentage capped at 100 descending, then numeric LampTest ID ascending. Displayed flux is uncapped. No weighted ARP score has been introduced.

### Red-zone rule and order

At least one known flag: flux <50% of declared, flicker >40%, CRI <70 or R9 <0. Sort by flag count descending, flux percentage ascending, flicker descending, CRI ascending, numeric ID ascending. Unknown sorting values go last.

### Calculation and missing values

Flux percentage = 100 × lm / lm_l. Missing or nonpositive declared flux gives an unknown ratio. Empty and nonnumeric entries remain null; zeros remain zero. Negative R9 is retained. Source flicker above 100% is preserved and requires checking the metric definition and measurement method.

### LampTest score

The 0–5 column reproduces the source rating field and is not used to order this ARP selection. Read the separate LampTest scoring methodology at the linked source.

### Additional parameters

Power, CCT, Rf/Rg and flicker at different voltages are retained in JSON. Socket, dimensions, beam, dimming, driver and thermal conditions require source and installation checks. High efficacy or high CCT alone does not establish a place.

### Evidence limits

This export does not supply a full measured spectrum for every row, eye-level mEDI, PstLM/SVM, long-term degradation tests, electromagnetic tests or biological outcomes. Declared lifetime does not establish tested durability. Table position does not predict sleep, health or performance.

### Updates and corrections

Keep the date, source snapshot, calculation rules and change history for each edition. Match challenges to the sample ID and test report. New measurements may change order and membership. Commercial participation must not determine ranking.

## Fields used to reproduce the calculation

| CSV field | Meaning | Use |
| --- | --- | --- |
| no, brand, model, date, prod | ID, brand, model, test date, production marking | Sample identity |
| cri, r9 | General and red colour rendering | Selection and flags |
| flicker | Source summary flicker percentage | Threshold and order; not PstLM/SVM |
| lm, lm_l | Measured and declared flux, lm | 100 × lm / lm_l |
| p, power_l; color, color_l | Measured and declared power and CCT | Context and sample record |
| Rf, Rg; flicker230/220/207 | Additional rendering; flicker by voltage | Retained where present |
| rating | LampTest overall score | Shown separately |
| type, base | Technology and socket | Selection scope |

## 5. From general to individual ranking

The general rating describes measured source properties and supplies the initial choice set. An individual ranking will be built for each person, task and situation. The same lamp may occupy different places in different contexts.

### Source

Spectrum, colour rendering, flux, flicker, power, optics, durability and data quality.

### Installation

Geometry, reflections, daylight, eye position, glare, dimmer and actual operating settings.

### Person and situation

Visual task, time, duration, sleep and work schedule, age group, preferences and observed outcomes.

### Compatibility first

Socket, voltage, size, optics, fixture and controls constrain the candidate set.

### Specify the scenario

Evening reading, daytime work, checkout, retail, school and cabin contexts need different conditions. Fix weights and mandatory limits before ranking.

### Measure the installation

Check task illuminance, eye-level light, spectrum, flicker and glare in the actual operating mode. Keep unknown values visible.

### Gather feedback

Observations and verified measurements update the model. Explain changes to the individual ranking.

A general technical draft is available now. The questionnaire below collects context and a measurement plan. Personalised ordering of individual lamps is not yet implemented.

Use the CSV snapshot and calculation JSON with this method. Preserve sample IDs, dates, unknowns and reconstruction status. Individual recommendations require context and measurements. Medical claims require separate research evidence.

[Original CSV snapshot](https://xpower.agency/academy/light/data/lamptest-2026-09-16.csv)

[All records and calculations JSON](https://xpower.agency/academy/light/data/arp-rating-2026-09-16.json)

[Methodology Markdown](https://xpower.agency/academy/light/materials/arp-rating-en.md)

[Reproduction code](https://xpower.agency/academy/light/data/rebuild-rating.mjs)