LGD’s Taliyah Gamble vs Invictus Gaming in LPL

TL;DR: LGD GAMING have gambled on LGDTangyuan’s Taliyah into IGRookie’s Orianna in LPL Game 3 against Invictus Gaming: Taliyah holds a 37.0% LPL WR over 54G, while the direct matchup is only 20.0% global over 10G. The draft model favors Invictus Gaming 60% to 40%.

LGDTangyuan’s Taliyah is a deliberate attempt to unlock Maokai-controlled skirmishes and Camille flanks, but it asks LGD GAMING to create the first move before Orianna controls grouped fights. A 37.0% LPL WR over 54G makes this a high-risk response rather than a standard meta comfort pick.

Compositions: Invictus Gaming vs LGD GAMING

Invictus Gaming have the cleaner five-versus-five composition. IGTheShy’s Rumble supplies poke and zone control, IGWei’s Skarner offers reliable engage, and IGRookie’s Orianna turns that setup into a decisive wombo combo. IGJiaQi’s Aphelios gives the blue side scaling carry damage, while IGMeiko’s Thresh can peel, pick, or lantern a vulnerable carry out of danger.

LGD GAMING want faster, more fragmented fights. LGDBurdol’s Camille needs a flank or split-push angle, LGDHeng’s Maokai provides engage, and LGDTangyuan’s Taliyah can cut off escapes. LGDShaoye’s Kalista and LGDCrisp’s Renata Glasc must snowball bot-side priority; otherwise, Aphelios and Orianna become harder to reach in late-game teamfights.

Key Picks and Stats

Top lane is volatile: IGTheShy on Rumble owns a 53.2% global WR over 1297G and 54.4% LPL WR over 169G, but Rumble is 25.0% over 4G versus Camille. LGDBurdol’s Camille is 51.1% global over 362G, 48.8% LPL over 41G, and 75.0% over 4G versus Rumble. Camille is the clearest counter-pick lane, though the samples are small.

In jungle, IGWei’s Skarner is 53.2% global over 438G but 49.1% LPL over 55G; his Maokai matchup falls to 33.3% LPL over 3G. LGDHeng’s Maokai is stronger at 61.3% LPL over 31G versus 50.0% global over 144G, including 66.7% LPL over 3G against Skarner.

The mid lane favors IGRookie. His Orianna posts 51.3% global over 949G and 58.9% LPL over 90G, with an 80.0% global WR over 10G into Taliyah. LGDTangyuan’s Taliyah is 47.9% global over 622G and 37.0% LPL over 54G, including that 20.0% global WR over 10G versus Orianna.

Bot is less decisive: IGJiaQi’s Aphelios is 47.0% global over 283G and 43.2% LPL over 44G, while LGDShaoye’s Kalista is 49.3% global over 203G and 48.0% LPL over 25G. IGMeiko’s Thresh has a 53.3% LPL WR over 30G and a 75.0% LPL WR over 4G versus LGDCrisp’s Renata Glasc, whose figure is 25.0% over 4G.

Draft Edge

Rumble confirms last night’s Invictus Gaming forecast, but Jarvan IV did not arrive. Orianna was forecast as a ban, yet Invictus Gaming secured her instead; the supplied draft does not include a ban slate to verify the other expected bans. LGD GAMING’s Taliyah is the sharp deviation from that forecast.

Invictus Gaming win through Skarner picks, Rumble Equalizer zones, and Orianna-Aphelios teamfights. LGD GAMING need Maokai and Taliyah to force early river control, then let Camille access Aphelios before scaling takes over.

Polymarket Market

Polymarket prices Game 3 at 64% for Invictus Gaming and 36% for LGD GAMING. Its series market is even more emphatic: Invictus Gaming are 94% now versus 6% for LGD GAMING, up from 62% to 38% pre-match—a 31.0 percentage points move toward Invictus Gaming.

The Game market is less optimistic than the series market for Invictus Gaming, suggesting the market sees LGD GAMING’s best remaining chance in this map despite the series damage. Two one-sided wins—20-4 kills in 26:50 and 27-11 kills in 25:30—and the Orianna-Taliyah lane data explain the series surge, while Camille and Maokai preserve LGD’s upset route.

Prediction

The model gives Invictus Gaming 60% and LGD GAMING 40%. I would move slightly to 62% for Invictus Gaming: their 0.700 team-form signal, 0.623 head-to-head signal, and 2-0 momentum outweigh LGD GAMING’s 0.504 tempo edge. Still, a successful early Maokai-Taliyah roam could change the mental state of a Game 3 quickly.

FAQ

*Data: draftlol live draft model + Polymarket, draft close.*