LYON vs G2 Esports MSI Draft: Lee Sin Bet Flips Game 2

LYON (2024 American Team) opens this draft by trusting Inspired on Lee Sin into Pantheon, and that choice sets the tone for the whole game. The pick is not automatically wrong — Lee Sin has a 60.0% global WR into Pantheon over 35G — but at MSI the champion has only a 42.9% WR over 14G, so LYON is betting on execution rather than patch-level comfort. If that early tempo lands, the draft can snowball fast; if not, G2 Esports has enough mid-game control to punish every missed window.

Compositions

LYON (2024 American Team) drafted a classic front-to-back teamfight shell with layered engage and steady scaling. Dhokla on Gnar and Isles on Leona give reliable initiation, Saint on Viktor supplies zone control and late damage, and Berserker on Varus adds long-range DPS and poke. Inspired’s Lee Sin is the only truly volatile piece, which means LYON wants proactive early skirmishes and dragon setups before Viktor and Varus are forced to play from behind.

G2 Esports answered with a more awkward but dangerous skirmish-and-pick composition. BrokenBlade’s Yasuo, SkewMond’s Pantheon and Labrov’s Nautilus create explosive access to fights, while Caps on Cassiopeia punishes short-range dives and Hans Sama on Ziggs gives waveclear and siege. This comp is less clean in full 5v5 front-to-back fights, but it can win through roam, side pressure and fast collapses if Pantheon unlocks mid first.

Compared with last night’s pre-draft view, G2 still showed the broader identity pool that made them attractive before champion select. Pantheon was one of the forecast ways for G2 to win, so that read was confirmed. What was not confirmed was LYON’s expected comfort package around Rumble, Lulu and Yunara: none of those appeared, which makes the Lee Sin choice feel even more deliberate.

Key Picks and Stats

Top lane is better for LYON on paper. Dhokla’s Gnar comes in with a 54.1% global WR over 788G, a 46.7% MSI WR over 15G, and he is 50.0% on the pick at MSI in 2G with a 5.6 KDA. Into him is BrokenBlade’s Yasuo at 39.5% global over 43G, 25.0% at MSI over 4G, and 0.0% at MSI in 2G for BrokenBlade with a 0.8 KDA. The matchup data also leans hard blue side: Gnar vs Yasuo is 61.5% globally over 13G and 100.0% at MSI over 2G.

Jungle is the swing role. Inspired’s Lee Sin has a 55.5% global WR over 384G, but only 42.9% at MSI over 14G; still, Inspired himself is 50.0% in 2G with a 10.5 KDA. SkewMond’s Pantheon sits at 45.4% global over 866G, yet 80.0% at MSI over 5G, and he is 100.0% in 1G with a 10.0 KDA. The head-to-head global stat favors Lee Sin at 60.0% over 35G, which is why this remains playable for LYON despite the tournament risk.

Mid lane tilts red. Saint’s Viktor is 49.8% global over 436G, 66.7% at MSI over 9G, but only 33.3% for Saint in 3G with a 4.2 KDA. Caps’ Cassiopeia is 54.7% global over 360G, and the Viktor matchup is brutal: Cassiopeia holds 68.8% globally over 32G into Viktor, while Viktor is only 31.2%.

Bot lane is closer. Berserker’s Varus has a 50.6% global WR over 696G, but 36.4% at MSI over 11G and 0.0% for Berserker in 1G despite a 8.0 KDA. Hans Sama’s Ziggs is 51.6% global over 157G and 33.3% at MSI over 12G. Support slightly favors LYON in direct matchup data because Leona vs Nautilus is 59.0% over 61G, even if Isles’ Leona MSI sample is only 33.3% over 3G.

Draft Edge

The raw draft edge is narrow, and that matches the model better than public sentiment does. LYON has the cleaner teamfight, the more stable top-side matchup, and a strong engage structure if Gnar and Leona reach objectives first. G2 Esports has the sharper mid-jungle punishment pattern because Cassiopeia plus Pantheon can trap Lee Sin’s early entries and turn every overforce into a lost map state.

So the winning conditions are clear. LYON (2024 American Team) needs Inspired to accelerate lanes before Caps takes over mid control, then let Viktor and Varus scale behind Gnar-Leona engage. G2 Esports needs SkewMond and Caps to break the game open through skirmish timing, with Ziggs converting priority into tower pressure.

Polymarket Market

Polymarket is the strongest outside signal here, and it is far more skeptical of LYON (2024 American Team) than the draft model is. The Game 2 market prices LYON at 36% and G2 Esports at 64%, while the live series market sits at LYON 72% and G2 Esports 28%. Those numbers are not close, so this is not a reused deciding-map market snapshot.

There is no pre-match series price provided here, so we cannot measure the exact move from before the series. What we can say is that the market is much more optimistic about LYON winning the series than this specific game, likely because LYON already won G1 despite the 8-17 kill score in 35:40 and now only needs to convert one more map. For this draft alone, however, Polymarket is clearly trusting G2’s season-level strength, their 0.547 elo signal, and the Cassiopeia-Pantheon access to early punishment more than it trusts LYON’s cleaner front-to-back shape.

Prediction

The model starts at LYON (2024 American Team) 51% to G2 Esports 49%, and I would trim that slightly to LYON 50% and G2 Esports 50%. LYON’s team form signal at 0.700 and the momentum of leading the series after G1 matter, but G2’s stronger season WR at 0.660 versus 0.611 and the mid-jungle matchup quality pull this back to even.

In other words, the draft is playable for LYON, not comfortable. If Inspired’s Lee Sin gets the first clean tempo lead, blue side can justify the gamble; if not, G2 has the more punishing map.