The rules behind the K-Pop Arena ranking, live counters you can check, and every correction we make.
POST /api/votes/{groupId} with a device ID (a random UUID created in your browser and kept in local storage) or your account ID when you are signed in. Writes are limited to 15 per minute per IP address.HTTP 202 “not counted” and nothing is written. Google’s user-triggered fetchers are not blocked: there are real people behind them.HTTP 429 with the next vote time. The check runs again inside the database transaction, so two simultaneous taps cannot both count.A device is not a person: someone with several devices can vote from each. The ranking measures fan engagement on K-Pop Arena, not a census of every fan.
Live from /api/vote-transparency · last 14 days (UTC) · updated 2026-10-11 17:28 UTC
| Day (UTC) | Counted | 24-h limit | Bots | Counted share |
|---|---|---|---|---|
| 2026-10-11 (in progress) | 587 | 222 | 107 | 64% |
| 2026-10-10 | 782 | 187 | 270 | 63% |
| 2026-10-09 | 894 | 205 | 231 | 67% |
| 2026-10-08 | 869 | 183 | 291 | 65% |
| 2026-10-07 | 1,050 | 237 | 366 | 64% |
| 2026-10-06 | 1,022 | 299 | 450 | 58% |
| 2026-10-05 | 1,227 | 403 | 404 | 60% |
| 2026-10-04 | 982 | 500 | 194 | 59% |
| 2026-10-03 | 788 | 414 | 440 | 48% |
| 2026-10-02 | — | — | — | — |
| 2026-10-01 | — | — | — | — |
| 2026-09-30 | — | — | — | — |
| 2026-09-29 | — | — | — | — |
| 2026-09-28 | — | — | — | — |
These counters went live on Oct 3, 2026. Days before that show “—” (no data), not zero. “Refused” counts attempts, not people: one device retrying three times adds three. “Counted share” = counted ÷ all vote attempts that day.
Counter totals from Oct 3, 2026 (UTC): 8,201 counted, 2,650 refused by the 24-hour limit, 2,753 blocked as bots. “Votes in the ranking” is the season table after corrections.
Community polls and the Weekly Challenge show how much a result can be trusted, not just who is ahead:
The main ranking is different: it is not a sample but the exact count of the votes cast on K-Pop Arena under the rules above. Intervals describe how much a small poll could still move; they do not mean votes were miscounted.
We found that crawlers (mostly Meta's link-preview crawler) were triggering automatic votes from shared vote links. We blocked them and removed 1,747 crawler votes from Season 2.
1,391 votes came from Meta's link-preview crawler, 354 from other bots identified by their user agent (for example ionbot, TikTokSpider, Baiduspider and Googlebot) and 2 from Meta's network.
The removal touched 115 groups, in both the Season 2 table and the all-time totals. The values before and after were saved, so the correction is reversible.
Votes that arrive through Google's user-triggered fetchers are not blocked: they come from real fans using a Google service, and about half of them hit the same device's 24-hour limit.
K-Pop Arena vote data, methodology v1, accessed Oct 11, 2026. https://bestkpopband.com/methodology
Press and data questions: apps@brainystack.co · Press kit
Each step a vote goes through on the server, with the real endpoints and live counters. 84 s, no AI footage.
K-Pop Arena is a fan project, not affiliated with any artist or agency. If you find a problem with the numbers, write to us: confirmed problems go into the corrections log.