- Hanwha Life Esports#1 · 79.3Zeus
ROSTER · DOSSIER
韩国 '26 数据核查the Korea '26 fact-check
对 proposed 韩国队名单做数据反查;候选池不是手填,而是从每个赛事名单里抓 KR 选手。统计窗口:An adversarial audit of the proposed Team Korea draft against the KR candidate pool, judged solely on tape from 2024 Spring → 2026 Spring (reviewable).
- Gen.G#5 · 60.2Canyon
- Hanwha Life Esports#2 · 61.9Zeka
- T1#7 · 45.6Faker
- Hanwha Life Esports#8 · 60.2Gumayusi
- T1#3 · 64.2Keria
公开计算Open Math
每个选手先按赛事分区抓原始数据,再按出场局数做加权平均;最后只在同位置候选池内做 min-max 标准化,再乘以该位置的指标权重。Every player is first aggregated from per-tournament raw rows using games-weighted averages; then each metric is min-max normalized inside the same-role pool and multiplied by role weights.
champion_pool = MAX(champion_pool_split)
例子:LCK 常规赛 40 局会比 Worlds 9 局权重大,因为样本更大;但 Worlds 不会被丢掉。Example: a 40-game LCK regular split carries more weight than a 9-game Worlds run, but Worlds is still included.
score = SUM(norm × role_weight) / SUM(role_weight) × 100
所以 70 分不是主观打分,而是这个选手在同位置所有韩国候选人里,各项指标标准化后的加权百分位。So a 70 is not a manual grade; it is the weighted percentile-like score after normalizing that player's metrics against Korean candidates in the same role.
以 Zeus 的 Top Lane 为例,前五个最高权重如下。完整权重已写入 Excel 的 Scoring Weights sheet。Using Zeus's Top Lane role as the example, the top five weights are below. Full weights are in the Excel Scoring Weights sheet.
明确排除(不会写入计分数据表):Out of scope (per the brief, NEVER aggregated into the score): MSI 2024 · MSI 2025 · MSI 2026 · First Stand 2025 · First Stand 2026 · Esports World Cup 2024 · Esports World Cup 2025 · EWC 2026 Online Qualifier - Korea。这些赛事不会参与任何最终分数。. These rows are simply never written to the database.
Top Lane 档案dossier
Side-lane carry weight, prio under pressure, scaling vs. tank flexibility. Solo-kill rate and CSD@14 indicate lane dominance; DMG% and DPG show carry conversion.
- 01ZeusHanwha Life Esports· 待核查· proposed3183.1457.8%22.9%6001.4920.26.50.3679.330
- 02KiinGen.G3054.2758.9%21.8%6141.5312.05.60.2778.336
- 03DoranT13203.0657.0%22.1%6001.516.71.20.3468.632
- 04SiwooDplus KIA1852.6051.8%23.2%6271.62-1.70.90.3765.830
- 05kingenNongshim RedForce2832.6952.2%23.3%5911.53-2.1-0.60.2457.932
- 06RichKiwoom DRX1662.4955.2%24.1%6291.68-15.1-7.20.3056.228
- 07DuDuDN SOOPers2392.4454.0%23.0%5571.454.43.10.2556.226
- 08CastingHANJIN BRION722.7752.1%23.5%6301.67-1.9-3.20.1754.624
- 09ClearBNK FearX2392.5052.8%21.6%5541.45-8.8-4.00.2744.428
- 10MorganOK BRION1742.3657.5%22.9%5321.41-7.7-2.30.1243.426
- 11PerfecTKT Rolster2643.0449.3%20.3%5161.39-10.2-2.60.2035.622
- 12DnDnNongshim RedForce551.7245.6%20.5%4341.15-15.6-3.70.2723.220
- 13RascalDRX621.6055.1%19.4%4071.11-18.3-9.60.1513.124
Jungle 档案dossier
Pathing impact and early skirmish presence. KP%, GDPM, CSD@14 and FB participation form the impact triangle. Vision/min reads as map command.
- 01PeanutHanwha Life eSports2544.5673.1%2.0-1.46.2440%1.682462.9—
- 02KanaviHanwha Life Esports633.1070.3%9.42.07.1343%1.811861.5—
- 03LucidDplus KIA3103.9672.6%4.51.06.1738%1.723061.3—
- 04CuzzKT Rolster2713.7374.0%3.22.16.5832%1.852860.6—
- 05CanyonGen.G· 待核查· proposed3055.5971.2%4.52.36.7131%1.662860.2—
- 06CrocoOK BRION553.3475.4%0.31.76.8938%1.442259.6—
- 07OnerT13354.3370.3%6.63.66.7236%1.642859.3—
- 08WillerKiwoom DRX1123.1572.1%-3.01.36.5335%1.903053.0—
- 09SpongeNongshim RedForce2332.3473.9%-5.1-1.26.1337%1.603248.6—
- 10PyosikDN SOOPers2493.4874.1%-7.4-3.26.2235%1.712847.7—
- 11GIDEONHANJIN BRION1923.2272.4%1.20.76.3734%1.692445.5—
- 12RaptorBNK FearX2063.0674.3%-7.7-3.96.2231%1.593041.6—
- 13YoungJaeOK BRION353.2075.0%-19.2-7.06.3031%1.812230.4—
- 14HamBakOK BRION322.4373.1%-10.2-8.16.1834%1.472025.3—
- 15SylvieNongshim RedForce842.4472.0%-11.3-5.95.6333%1.632424.0—
- · EnemyJG·min and Counter-Jungle CS require Riot Match-v5 timeline post-processing and are not available on public stat sites; columns show — until a timeline pipeline is added.
Mid Lane 档案dossier
Wave-and-roam tempo with team-fight ceiling. DPM/DMG% measure damage threat, DPG efficiency, CSD@14 lane control, solo kills the matchup edge.
- 01ChovyGen.G3056.1568.1%28.1%7871.7222.39.20.2993.732
- 02ZekaHanwha Life Esports· 待核查· proposed2735.3266.0%25.9%6721.5613.11.30.2461.928
- 03BddKT Rolster2864.6970.7%26.8%6701.625.51.50.1760.330
- 04ShowMakerDplus KIA3104.2467.0%25.6%6491.582.6-1.90.2355.134
- 05FisherNongshim RedForce643.0464.5%28.3%6991.68-0.6-2.80.2253.228
- 06ScoutNongshim RedForce613.8966.7%24.4%6901.66-0.20.40.1950.430
- 07FakerT1· 待核查· proposed3374.1562.3%24.8%6651.59-1.7-1.20.2045.626
- 08ViclaBNK FearX1402.9965.4%25.0%6871.72-7.7-1.40.2145.426
- 09BulldogDN Freecs1662.8568.7%28.3%6751.69-9.2-2.50.1044.622
- 10ClozerDN SOOPers2443.4165.8%26.1%6271.53-3.81.90.1644.330
- 11KyeahooDRX883.1968.1%27.1%6311.58-9.1-3.00.1941.824
- 12CallmeNongshim RedForce362.0772.5%29.2%6171.55-8.8-5.00.1439.318
- 13UcalKiwoom DRX1273.1365.8%24.2%6391.56-9.0-2.20.1235.528
- 14RoamerHANJIN BRION343.5865.3%22.4%6471.64-17.4-5.30.2131.820
- 15CalixNongshim RedForce732.8761.1%22.9%5811.44-7.9-1.50.2625.822
- 16SeTabDRX341.9165.1%27.0%5711.45-20.0-5.80.1521.720
- 17KarisOK BRION572.2666.2%26.0%5431.38-17.9-2.10.1219.420
Bottom 档案dossier
Late-game ceiling and damage-per-gold efficiency. Duo proximity (when available) maps to bot-side synergy.
- 01ViperHanwha Life eSports2546.2971.0%30.0%7731.719.99.60—81.328
- 02PeyzT11747.1670.0%27.7%7671.6117.29.48—77.128
- 03RulerGen.G1886.0968.8%27.3%8021.6314.79.98—74.724
- 04AimingKT Rolster2936.3172.6%29.0%7371.5712.99.98—73.826
- 05SmashDplus KIA904.5867.7%26.6%7521.5116.610.39—66.130
- 06DeftKT Rolster1054.4869.8%29.0%6951.635.78.76—64.932
- 07DiableBNK FearX1583.5270.1%27.7%7731.657.29.75—61.424
- 08GumayusiHanwha Life Esports· 待核查· proposed3034.8967.6%27.1%7241.589.19.43—60.226
- 09TaeyoonBNK FearX552.8464.9%28.0%7911.647.79.74—60.024
- 10TeddyHANJIN BRION1933.3770.0%28.6%7011.600.69.49—54.826
- 11deokdamDN SOOPers1793.8369.3%27.5%7101.63-23.79.00—49.228
- 12JiwooKiwoom DRX2162.9668.1%28.0%7011.63-12.99.01—46.626
- 13HypeOK BRION814.1669.2%27.0%6841.59-12.99.43—45.722
- 14BerserkerDN Freecs883.6869.0%26.4%6511.54-0.09.52—44.330
- 15HenaFearX913.7474.2%27.4%6211.50-18.58.91—41.222
- 16LeaperKwangdong Freecs442.9066.8%26.5%6391.51-33.99.60—31.822
- 17EnvyyOK BRION642.0663.2%28.3%6151.55-25.49.15—28.124
- 18BullOK BRION533.2570.2%23.2%5491.37-16.98.33—21.930
- · DuoProx requires Riot timeline positional data (not in public scrapes); the column is shown for transparency and will populate when a timeline pipeline is wired.
Support 档案dossier
Vision command and KP. Low gold share with high KP / VSPM signals roaming efficiency; duo & jungle proximity describe map coverage.
- 01EffortKT Rolster483.9880.8%4.26-4.531%20—75.8—
- 02DelightHanwha Life Esports3014.3775.4%3.666.835%34—67.0—
- 03KeriaT1· 待核查· proposed3374.5272.3%3.6215.133%38—64.2—
- 04CareerDplus KIA664.0574.0%3.91-1.333%36—62.7—
- 05DuroGen.G2254.6973.6%3.750.832%28—60.2—
- 06NamgungHANJIN BRION502.7568.1%4.244.632%26—51.8—
- 07KellinBNK FearX2483.3974.1%3.63-3.031%32—49.7—
- 08PeterDN SOOPers1873.1474.9%3.73-3.534%22—49.5—
- 09LehendsNongshim RedForce2693.1568.7%3.452.638%26—41.0—
- 10BeryLDplus KIA2243.0571.5%3.367.523%32—40.5—
- 11MohamDplus KIA352.5072.6%3.79-15.929%14—35.2—
- 12AndilKiwoom DRX2322.3672.4%3.36-10.230%32—33.1—
- 13PleataNongshim RedForce992.4873.3%3.49-8.025%30—32.2—
- 14ExecuteFearX542.2272.2%3.34-9.141%26—31.8—
- 15PolluKT Rolster1662.3872.0%3.47-13.837%24—30.2—
- 16LifeDN SOOPers981.9473.0%3.54-5.327%28—29.2—
- 17QuantumDN SOOPers301.7970.5%3.58-13.917%12—12.2—
- · DuoProx & JGProx require Riot timeline positional data (not in public scrapes); columns are placeholders.
方法说明how the score is built
候选池The candidate pool is 不是手工挑人not hand-curated。每个 in-scope 赛事都会先抓队伍列表,再进每个队伍的 roster 页,把 player_id、名字、国籍、位置、队伍写入 SQLite。来源:. For every in-scope tournament we fetch the team list from gol.gg。韩国队资格过滤就是 SQL 条件 , walk each team's roster page, and persist every participant's (golgg_id, name, country, role, team) into SQLite. Eligibility for Team Korea is the SQL filter country = ‘KR’ — which,所以去 LPL 的韩国选手仍然 eligible;非韩国选手不会进候选池。 which correctly includes Korean players who have moved to the LPL and excludes non-Koreans.
80 Korean players cleared the ≥30 in-scope games threshold (≈ one LCK split) across 559 per-tournament stat rows. Stats are games-weighted across every contributing tournament — large samples (LCK regular season) naturally outweigh small ones (Worlds 4-game runs) without throwing away any data. The 0–100 Score column is a role-weighted, min-max-normalised composite over the user’s metric list (KDA, solo kills, KP, DMG%, DPM, DPG, GDPM, 14CSD, gold%, VSPM, champion pool; +FB% for jungle). Highest cell in each column is gold.
Three metrics in the brief — jungle time spent invading the enemy jungle, counter-jungle CS, and bot / support proximity rates — are NOT published by gol.gg. They require Riot Match-v5 timeline post-processing. Those columns render — and are excluded from the composite rather than fabricated.
dataset built 2026-05-22 05:54 UTC · 15 tournaments contributed · hit ↻ above to re-ingest (or run npm run ingest on the host).