LYON vs FURIA MSI Game 3: Lulu Gamble Shapes Draft Edge

LYON (2024 American Team) opened this draft by leaning into a bot lane that asks for protection rather than pure lane control, and that makes Isles on Lulu the defining choice of Game 3. The number behind it is uncomfortable on paper—43.7% global over 700G for Lulu and only 37.1% into Milio over 62G—but the idea is clear: enable Berserker’s Kog'Maw, stretch fights, and trust the series lead to buy time for scaling.

Compositions

LYON (2024 American Team) drafted Dhokla on Varus top, Inspired on Lee Sin, Saint on Viktor, Berserker on Kog'Maw, and Isles on Lulu. It is an unusual mix of poke, peel, and late-game scaling. Lee Sin is the only truly natural early skirmish engine, so LYON need Inspired to bridge the first 15 minutes while Viktor and Kog'Maw come online. If they reach structured front-to-back fights, Lulu plus Kog'Maw can carry through sustained DPS, while Varus and Viktor add ranged control.

FURIA answered with Guigo on Olaf, Tatu on Xin Zhao, Tutsz on Sylas, Ayu on Lucian, and JoJo on Milio. This composition is much cleaner in its timing: stronger early skirmish from Olaf-Xin Zhao, a more direct engage pattern, and Lucian-Milio for lane pressure. Sylas also gives Tutsz access to high-value ultimates in teamfights. FURIA’s best window is early to mid game, before Kog'Maw and Viktor fully stabilize the map.

Key Picks and Stats

The draft surprise is still Isles on Lulu. Lulu sits at 43.7% global over 700G, 66.7% at MSI over 3G, and only 37.1% into Milio over 62G. JoJo’s Milio is steadier at 47.5% global over 375G, 100.0% at MSI over 3G, and 53.2% into Lulu over 62G. On raw lane data, FURIA’s support matchup is favored, so LYON are drafting for comp function rather than lane numbers.

Bot lane becomes volatile because Berserker’s Kog'Maw has just 37.1% global over 35G, but the direct matchup into Ayu’s Lucian is 57.1% over 7G. Ayu’s Lucian is 47.3% global over 509G and 75.0% at MSI over 4G, yet only 14.3% into Kog'Maw over 7G. That is a small sample, but it explains why LYON can justify the gamble.

Top lane is also strange. Dhokla’s Varus top shows 49.2% global over 689G, but 0.0% at MSI over 6G and 38.9% into Olaf over 18G. Guigo’s Olaf has 51.6% global over 95G, 0.0% at MSI over 3G, and 44.4% into Varus over 18G. Neither side owns a convincing tournament sample, but Olaf’s side-lane threat is easier to execute.

In jungle, Inspired’s Lee Sin is one of LYON’s best conventional picks at 53.8% global over 372G, 40.0% at MSI over 5G, and 53.2% into Xin Zhao over 47G. Tatu’s Xin Zhao is 47.5% global over 1108G, 40.0% at MSI over 5G, and only 38.3% into Lee Sin over 47G. Mid is tighter: Saint’s Viktor is 47.2% global over 424G and 50.0% at MSI over 2G, while Tutsz’s Sylas is 41.4% global over 203G, 0.0% at MSI over 1G, but 55.2% into Viktor over 29G.

Draft Edge

Last night’s pre-draft read expected jungle and support pressure to define the map, and that part absolutely holds: Inspired on Lee Sin and Isles on Lulu are the two picks that decide whether LYON’s draft functions. The predicted blue-side B1 path toward a priority jungle or tempo core was closer to reality than the earlier Bard idea, because LYON ended up building around tempo from Lee Sin and protection from Lulu rather than standard blind priority. The expected ban story cannot be confirmed because the ban phase is not listed in the provided sheet.

The edge is narrow. FURIA have the cleaner early curve and more natural engage. LYON have the better late-game insurance if Berserker survives fights. The real hinge is whether Inspired can stop Olaf-Xin Zhao from snowballing side control before Viktor-Kog'Maw hit scaling.

Polymarket Market

Polymarket makes LYON (2024 American Team) 66% for Game 3 and 94% for the series, with FURIA at 34% and 6% respectively. No series pre-match market number is provided here, so the exact move cannot be quantified, but after LYON won G1 23-26 in 44:00 and G2 23-22 in 37:45, a major shift toward LYON is the logical reading.

The split between 66% in the game market and 94% in the series market is also coherent. Polymarket is saying this draft is playable for FURIA on a single-map basis, but LYON’s 2-0 lead makes the overall series far more secure. That fits the draft: FURIA’s comp has real early-game paths, yet LYON only need one clean scaling game to close.

Prediction

The model starts at 53% for LYON (2024 American Team) and 47% for FURIA. After the draft, I would shade it slightly toward the market but not all the way there: 55% LYON, 45% FURIA.

The reasons are external as much as tactical. LYON bring the stronger team-form signal at 0.900 versus 0.800, the better h2h at 0.556 versus 0.444, and obvious momentum from the first 2 wins. FURIA’s route is still clear—win early lanes, force skirmishes through Guigo on Olaf and Tatu on Xin Zhao, and deny Berserker the protected Kog'Maw setup—but if Game 3 slows down, LYON’s gamble starts to look deliberate rather than reckless.