BLG vs LYON MSI: Naafiri Risk Shapes a Tight Game 2 Draft

Xun is staking Game 2 on Naafiri, a jungle pick sitting at 33.3% in 9G at MSI, and that immediately makes this draft more volatile than the market suggests. The logic is clear, though: Bilibili Gaming are not drafting comfort for the tournament sample, they are drafting a skirmish tool that has posted 61.5% vs Skarner over 13G globally, so if Xun can unlock early fights, the pick stops looking reckless and starts looking targeted.

Compositions

Bilibili Gaming draft a front-to-back teamfight setup with layered engage and strong mid-game skirmish. Bin on Gnar, Xun on Naafiri, Knight on Orianna, Viper on Ashe and ON on Seraphine give BLG initiation through Ashe arrow and Gnar flank, plus Orianna follow-up and Seraphine sustain. Their ideal game is active from level 6 onward: contest river, force grouped fights, and snowball before Viktor and Ziggs fully stabilize waveclear.

LYON (2024 American Team) answer with a more disruptive red-side composition. Dhokla on Ambessa and Inspired on Skarner give engage angles, Saint on Viktor provides scaling control, and Berserker on Ziggs with Isles on Pyke creates a volatile bot lane built around poke, pick and map tempo. This comp can play slow through Viktor and Ziggs siege, but Pyke means LYON also have a roam lever that can break side lanes before 20 minutes.

Key Picks and Stats

Top lane is close. Bin’s Gnar owns 54.3% global WR in 781G and 50.0% MSI WR in 10G, while Dhokla’s Ambessa sits at 48.3% global WR in 812G but a much stronger 62.5% MSI WR in 8G. The direct matchup slightly favors Gnar historically: Gnar has 47.6% vs Ambessa in 84G globally, while the draft model’s matchup sample gives 51.16% to Gnar. This looks playable for both, but Bin has the cleaner teamfight role.

Jungle is the pivot. Xun’s Naafiri shows 52.2% global WR in 335G, yet only 33.3% MSI WR in 9G, and Xun himself is 0.0% in 1G on it at MSI despite a 14.0 KDA. Inspired’s Skarner has 51.9% global WR in 270G and 66.7% MSI WR in 3G. The oddity is that Naafiri has 61.5% vs Skarner in 13G globally, while the model’s smaller matchup bucket only gives BLG 47.89%. That is exactly why this pick feels deliberate rather than desperate.

Mid lane leans LYON. Knight’s Orianna is 50.2% globally in 633G and 60.0% at MSI in 5G, but only 41.0% vs Viktor in 39G. Saint’s Viktor is 49.9% globally in 427G, 80.0% at MSI in 5G, and Saint is 100.0% in 1G on the pick with an 8.5 KDA. If Saint gets first move into objectives, LYON’s setup becomes much easier to execute.

Bot lane is where LYON’s draft gets most interesting. Viper’s Ashe has 55.0% global WR in 654G but 0.0% MSI WR in 4G, and ON’s Seraphine has 53.0% globally in 651G but 0.0% MSI WR in 5G. Across from them, Berserker’s Ziggs shows 52.8% global WR in 144G, 50.0% MSI WR in 8G, and 100.0% in 1G at MSI with a 17.0 KDA. Isles’ Pyke owns 54.1% global WR in 111G and has a favorable 55.6% vs Seraphine in 27G. Ashe into Ziggs is fine at 50.0% in 22G, but Seraphine into Pyke at 40.7% in 27G is a real pressure point.

Draft Edge

I give a narrow draft edge to LYON (2024 American Team). The Viktor, Ziggs and Pyke trio attacks BLG’s slower backline pieces, and Saint plus Berserker have the cleaner scaling damage profile. BLG’s win condition is still very real: Naafiri must accelerate side skirmishes, Ashe arrow has to start fights on BLG’s terms, and Bin’s Gnar needs decisive flank timing. If BLG do not build an early lead, LYON’s siege and pick tools become harder and harder to answer.

Polymarket Market

Polymarket is aggressively pro-Bilibili Gaming: 86% for Game 2 and 96% for the Series right now, leaving LYON at 14% and 4%. Series pre-match odds were not provided, so exact movement from pre-match cannot be measured here. What we can measure is the gap between the current Game and Series prices: the market is far more optimistic about BLG winning the series than about BLG winning this specific map. That makes sense after BLG already took G1 12-11 in 32:03. Traders can believe BLG remain overwhelming favorites to close the set while still respecting that LYON’s red-side draft, especially Viktor-Ziggs-Pyke and the Skarner response, gives them a better shot in this single game than the overall series line implies.

Prediction

The model opens at Bilibili Gaming 47% — LYON (2024 American Team) 53%, and I only trim that slightly to 48% — 52%. LYON’s lanes and scaling profile look better on paper, but BLG’s 0.691 season WR, 0.800 team form signal, and the momentum of winning G1 matter in a short series. If Xun’s Naafiri reaches targets first, BLG can flip the whole read; if not, LYON’s draft should age better.