Pro meta.
The latest pro meta (16.14), next to what our model predicts for the live patch (16.14).
Meta now Pro meta Over time
Our edge · model calls not yet in pro play · 16.14

Champions climbing in high-elo soloq that our model expects in pro drafts next, before they're picked. Historically about 58% of these early calls go on to reach pro play (the named ledger lives on the Track record).

Pyke
100% model
utility · not yet drafted
Qiyana
100% model
jungle · not yet drafted
Olaf New
100% model
top · not yet drafted
Taliyah New
99% model
middle · not yet drafted
Alistar
99% model
utility · not yet drafted
Gangplank
99% model
top · not yet drafted
Skarner New
99% model
jungle · not yet drafted
Kalista New
99% model
bottom · not yet drafted
Pantheon New
99% model
utility · not yet drafted
Hecarim
99% model
jungle · not yet drafted
Varus New
99% model
top · not yet drafted
Zeri
99% model
bottom · not yet drafted
Latest pro meta · 16.14 · sorted by presence
Champ Presence Prev patch Model Read
Poppy 100% 63% ▲ 100% model backs it
Nocturne 83% 55% ▲ 99% model backs it
Cassiopeia 83% 52% ▲ 99% model backs it
Shen 67% 13% ▲ 100% model backs it
Jayce 67% 68% 100% model backs it
Jarvan IV 67% 38% ▲ 100% model backs it
Vayne 67% 46% ▲ 100% model backs it
Syndra 67% 37% ▲ 100% model backs it
Naafiri 67% 37% ▲ 100% model backs it
Vi 67% 72% ▼ 100% model backs it
Orianna 67% 65% 99% model backs it
Locke 67% new 99% model backs it
Akali 67% 52% ▲ 98% model backs it
Camille 50% 48% ▲ 100% model backs it
Ezreal 50% 37% ▲ 100% model backs it
Bard 50% 55% ▼ 100% model backs it
Viktor 50% 21% ▲ 99% model backs it
Kai'Sa 50% 10% ▲ 99% model backs it
Rumble 50% 46% ▲ 99% model backs it
Ryze 50% 37% ▲ 99% model backs it
Leona 50% 7% ▲ 95% model backs it
Ambessa 33% 25% ▲ 100% model backs it
Nautilus 33% 52% ▼ 100% model backs it
Lucian 33% 31% ▲ 100% model backs it
Gnar 33% 49% ▼ 100% model backs it
Renekton 33% 32% 100% model backs it
Rell 33% 20% ▲ 99% model backs it
Galio 33% 10% ▲ 97% model backs it
Ziggs 33% 37% ▼ 95% model backs it
Anivia 33% 23% ▲ 93% model backs it
Lee Sin 17% 38% ▼ 100% model backs it
Dr. Mundo 17% 8% ▲ 100% model backs it
Jhin 17% 20% ▼ 100% model backs it
Ahri 17% 25% ▼ 100% model backs it
Caitlyn 17% 32% ▼ 100% model backs it
Sylas 17% 21% ▼ 99% model backs it
Mordekaiser 17% new 99% model backs it
Yone 17% 18% 99% model backs it
Yorick 17% 18% 99% model backs it
Annie 17% 28% ▼ 99% model backs it
Xin Zhao 17% 25% ▼ 99% model backs it
Karma 17% 23% ▼ 97% model backs it
K'Sante 17% 18% 97% model backs it
Rakan 17% 20% ▼ 97% model backs it
Milio 17% 11% ▲ 97% model backs it
Kled 17% 3% ▲ 96% model backs it
Twisted Fate 17% 14% ▲ 96% model backs it
Elise 17% 6% ▲ 92% model backs it
Wukong 17% 24% ▼ 91% model backs it
Trundle 17% 25% ▼ 84% model backs it
Neeko 17% 11% ▲ 80% model backs it
Tahm Kench 17% new 74% model backs it
Pro presence vs. model prediction
emergingconfirmed metaunsupportedpro presence →model score →Lee Sin*Dr. Mundo*Pyke*Qiyana*Olaf*Jhin*Ahri*Caitlyn*Taliyah*Alistar*Sylas*Mordekaiser*ShenJayceJarvan IVVayneSyndraPoppyNocturneCassiopeia
Each champ is a dot. x is how often pros pick or ban it this patch. y is our model's predicted prominence, on a log-odds spread rather than linear, so the near-certain picks don't all pile up at the top. We show only the actual pro drivers and our highest-confidence calls, and * marks an emerging call. Top-left is our edge: climbing in high-elo soloq before the draft catches up. Bottom-right is played by pros but not backed by our data.
Model is our scorer's predicted likelihood the champ shows up in pro play. Presence is how often it actually did (pick and ban share). They're related but not the same unit, so read the comparison as a direction. Agreement on the current patch is partly a mirror, since the model is trained to predict presence. The real test of an emerging call (high model, low presence now) is whether it climbs next patch, which is what the Track record measures.