Bureau of Rift Statistics · Report No. 2 · Data
Every number in Report No. 2, with its range and source
Each number in the report comes from one row of this table: the claim, the value with its 95% range where it has one, and the file or page it was read from. Rows marked exploratory were added after the analysis plan was locked. The figures’ data is one file: Download the JSON (results by champion, day and Worlds year; never a player or a match). How it was counted is on the method page.
Free to reuse with credit: “Bureau of Rift Statistics, Report No. 2 (Grey Screen), 202,725 NA ranked games, 9 Sep to 25 Oct 2026.”
| # | Claim | Value | Source |
|---|---|---|---|
| 1 | Ranked games counted | 202,725 | census2.sqlite via r2load.py · kept matches: NA1, queue 420, 15 min+, gameVersion 16.18 to 16.21 |
| 2 | Champion appearances | 2,027,250 | census2.sqlite via r2load.py · 10 per kept match |
| 3 | Players drawn at random from the ladder | 18,523 | census2.sqlite accounts · main and top-up waves, all three rank groups, with or without games |
| 4 | Games dated | 9 September to 25 October 2026 | census2.sqlite matches.start_ms · last day 2026-10-25 (UTC) |
| 5 | Pick rate change after the Worlds-patch (26.20) nerfs, relative to unchanged champions | −14% (−17 to −10) | r2_out/part1_results.json amendment6.A1_all_nerfs_2620 · pooled, 4 champions: Ambessa 26.20, Ashe 26.20, K'Sante 26.20, Yunara 26.20; days -14..-1 vs 0..+13 |
| 6 | Win rate change after the 26.20 nerfs, relative to unchanged champions | −1.6 points (−3.4 to +0.3) | same · same |
| 7 | Pick rate change for champions that got a new skin and no balance change (26.19, 26.20) | +10% (+8 to +13) | r2_out/part1_results.json amendment6.A2_skin_only · 13 champions |
| 8 | Win rate change for the skin-only champions | +0.5 points (−0.4 to +1.4) | same · same |
| 9 | Pick rate change for buffed champions that were in the least-picked third | +25% (+20 to +30) | r2_out/part1_results.json amendment6.A5_buffs_low_third · 7 champions, 26.19 and 26.20 |
| 10 | Win rate change for those buffs | −0.9 points (−2.7 to +0.8) | same · same |
| 11 | Pick rate change for the other buffed champions | +18% (+16 to +20) | r2_out/part1_results.json amendment6.A5_buffs_rest · 13 champions |
| 12 | Win rate change for the other buffs | +0.4 points (−0.4 to +1.2) | same · same |
| 13 | Pick rate change, nerfs aimed at pros vs at everyone (26.19 and 26.20) | −13% (−17 to −10) vs −13% (−18 to −9) | r2_out/part1_results.json amendment6.A4_* · Riot stated audience, coded 21 Oct |
| 14 | Yunara after 26.20: pick and win change, relative to unchanged champions | picks −35% (−38 to −32), win −1.3 points (−3.5 to +1.0) | r2_out/part1_champions.csv · nerf; 8,615 games before, 5,474 after |
| 15 | K'Sante after 26.20: pick and win change, relative to unchanged champions | picks −1% (−11 to +10), win −5.3 points (−10.0 to −0.2) | r2_out/part1_champions.csv · nerf; 1,524 games before, 1,496 after |
| 16 | Janna after 26.19: pick and win change, relative to unchanged champions | picks +64% (+53 to +77), win +0.7 points (−2.7 to +4.3) | r2_out/part1_champions.csv · skin only; 2,331 games before, 4,078 after |
| 17 | Ashe after 26.20: pick and win change, relative to unchanged champions | picks −7% (−11 to −2), win −0.5 points (−2.7 to +1.7) | r2_out/part1_champions.csv · nerf; 6,212 games before, 5,723 after |
| 18 | Ambessa after 26.20: pick and win change, relative to unchanged champions | picks −6% (−14 to +1), win +0.8 points (−3.3 to +4.7) | r2_out/part1_champions.csv · nerf; 2,321 games before, 2,013 after |
| 19 | Pooled 26.20 nerf pick change on patch day | +5% (−8 to +20) | r2_out/S4_event_study_2620.json · daily abnormal pick change vs pre-window, B = 500 |
| 20 | on day 7 | −15% (−24 to −6) | same · same |
| 21 | on day 13 | −18% (−25 to −9) | same · same |
| 22 | Yunara pick change on patch day | −21% (−33 to −7) | same · same |
| 23 | Change in the share of experienced players on the 26.20 nerfed champions | −0.1 points (−2.6 to +2.3) | r2_out/S2_mastery_pooled.json · experienced = top third of mastery points before the patch; mastery read at pull time |
| 24 | How much adjusting for experience moves the win change (nerfs, buffs) | 0.06 and 0.06 points | same · crude minus adjusted |
| 25 | Second method: win change for the 26.20 nerfs (logistic model, champion and day effects) | −1.3 points (−2.6 to +0.1) | r2_out/S7_logit_2620_nerfs.json · at a 50% baseline; B = 200 player bootstrap |
| 26 | Worlds tournaments read | 8 (2019 to 2026) | part0/worlds_presence.txt · Leaguepedia Picks and Bans, Play-In + Main Event |
| 27 | Worlds games read | 833 | part0/worlds_presence.txt · sum over years |
| 28 | Champion changes coded in the three patches up to each Worlds | 414 | part0/changes_coded.csv · Riot stated intent |
| 29 | Worlds pick-and-ban presence, buffed minus unchanged champions, mean of 8 years | +1.6 points (−1.1 to +4.4) | part0/part0_results.json · bootstrap over champions within year |
| 30 | Years in which buffed champions showed up more than unchanged ones | 4 of 8 | same · same |
| 31 | Average Worlds presence of champions nerfed in the run-up, by year (range) | 28% to 43% | same · mean over the year's nerfed champions |
| 32 | Average Worlds presence of unchanged champions, by year (range) | 7% to 12% | same · same |
| 33 | Average presence of champions buffed before Worlds 2026, through 26 Oct | 3% | same · 51 games |
| 34 | Azir: Worlds windows with a change, and his presence in those years | 5 years; 2020: 25%, 2022: 72%, 2023: 65%, 2024: 0%, 2025: 82% | same · every one included a nerf |
| 35 | Ashe: Worlds windows with a change | 7 of 8 | same · the most of any champion |
| 36 | Champions changed in at least 4 of the 8 Worlds windows | 31 | same |
| 37 | K'Sante top on 26.19: average win rate, matchup spread, worst matchup | 48.9%, spread 2.3 points (0.0 to 9.3), worst 47.5% vs Kayle | r2_out/part3_results.json · 1,442 games; DerSimonian-Laird, Q-profile interval |
| 38 | K'Sante top on 26.20: average win rate, matchup spread, worst matchup | 44.3%, spread 6.6 points (0.0 to 12.9), worst 36.0% vs Teemo | r2_out/part3_results.json · 1,466 games; DerSimonian-Laird, Q-profile interval |
| 39 | Typical matchup spread for a champion in a role, after removing noise (median, by patch) | 2.1 to 2.6 points | r2_out/part3_spread.csv · champion-roles with 2,000+ games |
| 40 | F1 on the new games: engage-only vs no movement, time dead | +7.2% (+7.0 to +7.5) | r2_out/F1_r2_contrasts.csv · F1 plan, unchanged; B = 2,000 |
| 41 | F1 on the new games: real escapes vs no movement | +0.5% (+0.3 to +0.7) | same · same |
| 42 | F1, Diamond and up: real escapes vs no movement | −0.5% (−0.8 to −0.2) | same · same |
| 43 | Kog'Maw after 26.20 (no change to him): pick change, relative to unchanged champions | +17% (+7 to +27) | r2_out/S8_named_champions.csv · unchanged; 1,898 games before, 2,014 after |
| 44 | Kog'Maw mid on the Worlds stage | 3 games, all MVK Esports, 16 and 17 October | worlds/stage_picks.csv · Play-In; Leaguepedia, read in the browser |
| 45 | Share of Kog'Maw's solo-queue games played mid, each day in the two weeks before | 10% to 26% (average 15%) | r2_out/kogmaw_mid.json · kogmaw_mid.py; weighted; 2 to 15 Oct |
| 46 | Share played mid on the two stage days, and in the days after | 20%, then 16% | same · 16 and 17 Oct; 18 to 25 Oct |
| 47 | Worlds 2026 pick-and-ban presence of the five champions nerfed in 26.20, through 26 Oct | Cassiopeia 67%, K'Sante 51%, Ambessa 27%, Ashe 20%, Yunara 20% | part0/worlds_presence.txt · 51 games; presence = picked or banned |
| 48 | The 26.20 nerfed champions picked on the Worlds stage in the two weeks after the patch | 5 of 5 | r2_out/S6_stage_exposure.csv · Play-In, 15 to 18 Oct |
| 49 | Cassiopeia games in the two weeks before and after 26.20 (under the 1,000 the plan needs for a champion figure) | 920 and 789 | r2_out/part1_champions.csv · plan §5 minimum |
| 50 | K'Sante's record on the Worlds 2026 stage: Play-In | 0 wins, 5 losses | worlds/stage_picks.csv · Leaguepedia match history, read in the browser |
| 51 | K'Sante's record on the Worlds 2026 stage: Main Event through 25 Oct | 4 wins, 5 losses | same · same |
| 52 | MVK Esports' record with Kog'Maw mid | 2 wins, 1 losses | same · Play-In, 16 and 17 Oct |
| 53 | Worked example: Yunara in bottom lane, games per 100, two weeks before and after 26.20 (crude weighted rate) | 12.6 before, 8.3 after (about 1 in 8, then 1 in 12) | r2_out/worked_example_yunara.json · worked_example.py; illustration of the method; the locked estimate is row 14 |
| 54 | Worked example: Yunara win rate before and after, and the untouched bottom-lane champions | Yunara 49.8% to 48.6%; untouched 50.3% to 50.3% | same · same |
| 55 | Patch notes dates in the census window | 26.18 on 9 Sep, 26.19 on 22 Sep, 26.20 on 6 Oct, 26.21 on 20 Oct (every two weeks) | leagueoflegends.com patch-notes tag page · read in the browser 27 Oct |
| 56 | Yunara's 26.20 nerf | Q on-hit damage 5 / 10 / 15 / 20 / 25 to 3 / 8 / 13 / 18 / 23 (2 less per hit at every rank); W slow no longer stacks with other slows | leagueoflegends.com, League of Legends Patch 26.20 Notes (6 Oct 2026), read in the browser 27 Oct · Riot: "handshake hyperscaler do nothing"; avoid it becoming the dominant meta heading into Worlds |
| 57 | K'Sante's 26.20 nerf | W mana cost 40 to 60 up to 45 to 65; W damage reduction 30% / 75% down to 25% / 70% | leagueoflegends.com, League of Legends Patch 26.20 Notes (6 Oct 2026), read in the browser 27 Oct · Riot: aimed at his reliability as a blind pick in coordinated play; HEARTSTEEL K'Sante skin released the same patch |
| 58 | Ashe's 26.20 nerf | attack damage per level 3.5 to 3 | leagueoflegends.com, League of Legends Patch 26.20 Notes (6 Oct 2026), read in the browser 27 Oct · Riot: one of the premier marksmen in pro play |
| 59 | Yunara release date | 16 July 2025 | League of Legends Wiki, Yunara page, read 27 Oct |
| 60 | Share of Kog'Maw's solo-queue games played mid, by patch | 26.18: 9%, 26.19: 16%, 26.20: 14% | r2_out/kogmaw_by_patch.json · kogmaw_patch.py; weighted; descriptive |
| 61 | Kog'Maw's win rate in the mid lane, by patch (rough 95% range from the game count) | 26.18: 61% (±7, 197 games), 26.19: 53% (±5, 375 games), 26.20: 45% (±5, 363 games) | same · weighted rate; range is a binomial approximation on the unweighted count |
| 62 | Solo-queue win rate change for nerfs aimed at pros vs nerfs aimed at everyone (26.19 and 26.20) | −2.0 points (−3.8 to −0.2) vs −1.3 points (−3.4 to +0.8) | r2_out/part1_results.json amendment6.A4_* · pro-aimed: Ryze, Ambessa, Ashe, K'Sante, Yunara; everyone: Nocturne, Poppy, Vi |
| 63 | Caps in the Hall of Legends | inducted 2026 (patch 26.18 notes headline) | leagueoflegends.com, Patch 26.18 Notes (9 Sep 2026) |
| 64 | Last Worlds won by a Western (North American or European) team | 2011 (Fnatic, Season 1); teams from Asia have won every Worlds since (Taipei Assassins in 2012, then Korean and Chinese teams) | Leaguepedia, World Championship results · background fact |
| 65 | EXPLORATORY. The 26.19 nerfs (Nocturne, Poppy, Rumble, Ryze, Vi): pick change relative to untouched champions, weeks 1 to 4 after the patch | week 1 −9% (−13 to −4), week 2 −15% (−19 to −10), week 3 −12% (−17 to −7), week 4 −19% (−24 to −15) | r2_out/S4_return_2619_exploratory.json · s4_return.py; added after the plan was locked; B = 500 |
| 66 | EXPLORATORY. Vi, week 4 after 26.19 | −4% | same · point estimate only |
| 67 | EXPLORATORY. Win change on the four 26.20 nerfs: experienced players (top third of mastery) vs everyone else | −3.0 points (−8.1 to +2.4) vs +1.0 points (−2.4 to +4.3) | r2_out/S2_split_exploratory.json · s2_split.py; 900 sampled appearances per champion and side; inconclusive |
| 68 | Fearless draft | a champion played once in a series cannot be played again in that series; used at Worlds 2025 | esports.net, Worlds 2025 Fearless Draft; Sheep Esports, What is Fearless draft · read 27 Oct |
| 69 | Thebausffs | streamer known for one-tricking Sion; reached Challenger in Korea in under a month | Dexerto, League streamer Thebausffs hits Challenger in Korea in under a month · read 27 Oct |
| 70 | K'Sante's balance changes since his 2024 gameplay update | nerfs in 25.05 (passive mark damage 20 to 12; All Out true damage up to 80% from 100%), 26.13 (Footwork shield) and 26.20 (Path Maker); one buff in 25.09 (Path Maker damage) | League of Legends Wiki, K'Sante/Patch history, read 27 Oct · gameplay update in 14.19; bug fixes not counted |
| 71 | K'Sante's Worlds pick-and-ban presence by year | 2023: 55%, 2024: 38%, 2025: 38%, 2026: 51% | part0/worlds_presence.txt · 2026 through 26 Oct |
| 72 | Azir's Worlds pick-and-ban presence by year | 2022: 72%, 2023: 65%, 2024: 0%, 2025: 82%, 2026: 0% | same · 2024: buffed 14.16, nerfed 14.18; 2025: nerfed 25.18 and 25.20; 2026 through 26 Oct |
| 73 | Azir buffed in 2026 | 26.16 (Conquering Sands base damage) and 26.21 (mana growth) | League of Legends Wiki patch histories, read 21 Oct (pre-mortem §9) |
| 74 | Yunara's Worlds pick-and-ban presence | 2025: 88%, 2026: 20% | part0/worlds_presence.txt · 2026 through 26 Oct |
| 75 | EntreMed, May 1998 | New York Times front page Sunday 3 May 1998; stock about $12 at Friday close, opened about $85 Monday; research published in Nature in November 1997 | Huberman and Regev, Journal of Finance (2001), vault anchor articles/huberman-regev-contagious-speculation-cure-for-cancer |
| 76 | Check against Report 1 on the days both cover: champions within the margin of error; correlation | 158 of 173; 0.997 | r2_out/V5_summary.json · 29 Sep to 15 Oct |
Report No. 1
Bureau of Rift Statistics · Report No. 1 · Data
The data behind the figures
Every figure in Report No. 1 draws from one file. Here it is whole, and split into tables you can open in a spreadsheet. They hold the census's results by tier, champion and region, never a player or a match. How they were counted is on the method page.
Everything the figures draw, as JSON.
Download the JSONLoading the tables…
Terms
- Tier: Iron to Diamond, and Master+ (Master, Grandmaster and Challenger together).
- Effective appearances: a champion's appearances after weighting, the sample size its estimate rests on.
- Observed over expected: a champion's time dead over what its games predict. 1.25 means a quarter more grey screen than expected.
- Band: typical, or above or below what chance alone allows (the funnel's limits).
Free to reuse with credit: “Bureau of Rift Statistics, Report No. 1 (Grey Screen), 30,510 NA ranked games, 29 Sep to 15 Oct 2026.”
What this site keeps on your device
No account, no analytics and no cookies of ours. A few things are kept in your browser's storage, on this device only, so the burrow remembers you. None of it leaves your device:
| What | Why | Until |
|---|---|---|
| Your guess for each report (yes, no or inconclusive, and when) | So Ping knows what you called. You can change it in the den until you first head down the burrow; then it locks | You clear this site's data |
| That you've headed down past the den | It's what locks your guess | You clear this site's data |
| Your keepsake (gilt, bronze or plain) | Earned when you reach the verdict; it goes on the shelf in Ping's room next time | You clear this site's data |
| Your tier, sound on or off, light or dark | So the burrow opens the way you left it | You clear this site's data |
Not kept: the paint in Ping's room, where the rocks landed and where he wandered. Every visit starts clean. Your guess is never counted with anyone else's; when a shared tally comes, it will say so here first. The sign-up form is Kit's. Kit's script loads only when you start on the form, not with the page. Once it loads, it tells Kit the form was opened, keeps its own visitor id in your browser (ckid), and Kit's host may set a short-lived security cookie. What you send there is used only to email you the reports.
To forget all of it, clear this site's data in your browser's settings.