Where does Pokékipe's data come from? Sources, methods and reliability
Pokékipe is built on Smogon's public statistics, enriched with matchup scoring and historical context. This page explains exactly where our numbers come from, how we compute them, and what their limits are.
Where does Pokékipe's data come from?
The primary source: Smogon stats
All ladder usage data on Pokékipe comes from one public source: the monthly statistics published by Smogon at smogon.com/stats. We ingest the raw chaos JSON files (smogon.com/stats/{YYYY-MM}/chaos/{format}-{elo}.json) — the most detailed feed Smogon publishes. These statistics are automatically generated from rated battles played on Pokémon Showdown, the reference competitive simulator.
Pokémon Showdown is free and hosts millions of rated battles every month. Each rated battle contributes to the monthly statistics for its format and ELO bracket.
Smogon is not affiliated with Pokékipe
Pokékipe is an independent project. We use Smogon's public data in line with its free availability. Pokékipe is not affiliated with Smogon, Pokémon Showdown, The Pokémon Company or Nintendo.
How much data is analysed?
Battle volume per format
We're preparing a public dashboard of monthly volumes per format and ELO bracket. In the meantime, what you need to know is that volume varies dramatically by format and ELO threshold — high-traffic tiers like Gen 9 OU dwarf niche tiers, and low-ELO buckets dwarf high-ELO buckets.
Stats coming soon
Gen 9 OU monthly volume (threshold 0)
Stats coming soon
Gen 9 OU monthly volume (1695+ ELO)
Stats coming soon
Total battles ingested (all formats)
Volumes fluctuate with competitive news: a new patch, a regulation change or a new game release can meaningfully shift them. The public dashboard will surface these dynamics live.
Historical coverage
Pokékipe ingests data back to 2022 for priority formats, with partial coverage going back to 2014 for older tiers. The most recent month is always the most complete and reliable.
The ingestion pipeline
Behind the scenes, a polling watcher detects each new Smogon stat file as soon as it lands, parses it, computes deltas against the previous month, resolves Pokémon names against our reference dex, and writes a versioned snapshot to the database. Separate scrapers track suspect tests, Viability Ranking updates, VGC regulation changes, Smogon-tier tournaments, official Play! Pokémon events (via pokedata.ovh) and Pokémon Champions community tournaments (via the Limitless API).
Every number you see on Pokékipe is tied to a specific snapshot (format + month + ELO bracket), so historical comparisons are always apples-to-apples.
Which formats are covered?
Pokékipe covers every format for which Smogon publishes statistics. The priority formats ingested first are:
- Gen 9 OU — the reference Smogon Singles format (Scarlet & Violet)
- Gen 9 Ubers — no ban list, everything allowed
- Gen 9 UU / RU / NU / PU — lower tiers for underused Pokémon
- Gen 9 Doubles OU — Smogon's Doubles format
- Gen 9 Monotype — teams built around a single shared type
- Gen 9 Little Cup — first-stage Pokémon only
- Gen 9 National Dex — includes Pokémon not available in Scarlet & Violet
- VGC — current official Nintendo regulation (plus historical regulations)
Older generations (Gen 7, Gen 8 OU and variants) and special formats (Random Battle, etc.) are available but processed at lower priority.
How is usage rate calculated?
Precise definition
A Pokémon's usage rate is the percentage of battles in which that Pokémon appeared on at least one of the two teams. The formula is simple:
usage rate = (appearances / total battles) × 100
A 30% usage rate means the Pokémon was present in roughly 3 out of every 10 battles. This is not a win rate — a popular but mediocre Pokémon can still have a very high usage rate.
What usage rate does not measure
- It does not measure whether the Pokémon was actually sent in (in VGC you pick 4 out of 6).
- It does not measure whether the Pokémon wins or loses its battles.
- Usage rates of all Pokémon do not sum to 100% — they would sum to ~600% if every team ran the top 6.
How is the monthly delta calculated?
The delta (shown with ▲ / ▼ on Pokékipe) is the absolute difference in usage rate between the current month and the previous month for the same format and ELO bracket:
delta = usage_rate(month N) − usage_rate(month N−1)
A delta of +2.1% means the Pokémon gained 2.1 percentage points of usage compared to last month. Deltas are the fastest way to detect meta shifts: a newly published set, a suspect ban or a fresh counter shows up in the delta before anything else.
What do the ELO thresholds mean?
ELO measures player skill on the Pokémon Showdown ladder. Smogon publishes statistics at several thresholds:
| Threshold | Included population | Usage on Pokékipe |
|---|---|---|
| 0 (ALL) | Every rated battle, all skill levels | ALL mode — the full player base |
| 1500+ | Above-average players | — |
| 1630+ | Strong ladder players | — |
| 1695+ | Top ~5% of the ladder | ELITE mode (default) |
| 1760+ | Top ~2% of the ladder | — |
| 1825+ | Top 1% — high-level competitive play | — |
Pokékipe exposes two practical modes: ALL (threshold 0, every battle) and ELITE (the highest threshold available for the format — typically 1695, 1760 or 1825 depending on the tier and month). ELITE is selected by default because it reflects the actual competitive metagame.
The checks & counters score
Beyond usage, Pokékipe exposes a checks & counters score from 0 to 100 that represents how reliably Pokémon A beats Pokémon B in direct matchups on ladder. A score above 70 means a hard counter; above 55 means a soft check.
This score is derived from Smogon's checks and counters data and combines switch-in survivability, KO potential and lead-vs-lead interactions. It is not a single win rate — it is a synthetic reliability score.
Glossary of metrics used on Pokékipe
Usage rate
Percentage of battles in which the Pokémon appeared on at least one team.
Delta (Δ)
Change in usage rate compared to the previous month, same format and bracket.
Usage rank
The Pokémon's rank by descending usage rate in the selected format and snapshot.
Checks & counters score
Reliability of Pokémon A beating Pokémon B (0–100). >70 hard counter, >55 check.
Teammate rate
Among teams containing the focal Pokémon, percentage that also contain the teammate.
Snapshot
A full capture of a format at a given month and ELO bracket. Pokékipe stores every snapshot back to 2022.
Limitations and known biases
Player selection bias
Data reflects Pokémon Showdown players, not cartridge players. Showdown players tend to explore experimental sets, which can inflate the usage of creative picks that would never appear on Nintendo ladder.
Entry-gate bias
ALL mode (threshold 0) includes battles from complete beginners, which skews usage toward beginner-friendly Pokémon. ELITE mode corrects this but with lower total volume.
Time lag
Data published at the start of the month reflects the previous month. A very dynamic meta (post-ban, post-patch, post-regulation change) can be four to six weeks behind the current high-level reality.
Ladder vs tournaments
VGC ladder data does not perfectly reflect in-person tournaments. The ladder meta diverges from the tournament meta on item choices and niche-but-effective sets that excel in best-of-three series.