G2 Esports vs Top Esports MSI: Mel gamble shapes Game 5

Caps and G2 Esports have turned Game 5 into a real conviction test by putting Mel into Yone, a matchup that looks uncomfortable on paper but signals a clear read on how they want fights to unfold. Mel sits at 39.8% WR over 244G globally and only 20.0% vs Yone over 5G, so this is not a comfort percentage pick; it is a bet that Caps can survive lane, control mid waves, and give G2 Esports first access to objective setups.

Compositions

G2 Esports drafted a composition that mixes front-to-back teamfight with strong anti-engage tools and late-game insurance. BrokenBlade on Anivia, SkewMond on Poppy, Caps on Mel, Hans Sama on Sivir and Labrov on Rell gives them layered zone control, reset potential and decent scaling, but it is a comp that needs clean spacing. Early game, Poppy and Rell can force skirmishes while Mel and Anivia try to slow the map; mid game, G2 Esports want choke-point fights where Sivir can hit safely behind peel; late game, their win condition is surviving TES engage and turning the fight through terrain and disengage.

Top Esports answered with a more direct engage and pick structure. ZUIAN on Ornn, Tian on Maokai, Creme on Yone, JackeyLove on Ziggs and fengyue on Pantheon gives TES easier initiation and simpler execution around side lanes and river fights. Their early game wants Maokai and Pantheon to create roam pressure, their mid game is about Yone engage layered with Ornn and Maokai, and their late game still has reliable scaling through Ornn upgrades plus Ziggs siege.

Key Picks and Stats

Top lane is volatile. BrokenBlade's Anivia has 48.4% WR over 448G, but only 33.3% vs Ornn over 9G. ZUIAN's Ornn sits at 48.5% over 241G and 55.6% vs Anivia over 9G. That makes Anivia a real counter-idea rather than a safe meta blind, and it asks BrokenBlade to create side pressure or superior setup value instead of expecting a stable lane win.

In jungle, SkewMond's Poppy is 50.3% over 195G, yet 0.0% vs Maokai over 5G. Tian's Maokai is 49.5% over 93G and 80.0% vs Poppy over 5G. That lane-adjacent matchup matters because TES have the cleaner point-and-click access if Poppy cannot interrupt the first engage angle.

Mid is the draft hinge. Caps on Mel brings 75.0% MSI WR over 4G, but the broader sample is harsh at 39.8% over 244G and 20.0% vs Yone over 5G. Creme's Yone is 49.8% over 203G and 80.0% vs Mel over 5G. This is the surprise of the draft: G2 Esports are trusting current-event form over the larger historical matchup.

Bot lane also leans TES in raw matchup data. Hans Sama's Sivir is 51.4% over 477G, but only 20.0% vs Ziggs over 5G. JackeyLove's Ziggs is 51.1% over 139G, 33.3% MSI over 3G, and 80.0% vs Sivir over 5G. Labrov's Rell is 43.0% over 358G, 20.0% MSI over 5G, and 43.2% vs Pantheon over 37G, while fengyue's Pantheon is 44.2% over 858G, 0.0% MSI over 2G, and 56.8% vs Rell over 37G.

Draft Edge

The pre-draft read that G2 Esports would arrive with a cleaner draft map is mostly confirmed, but not in the expected way. G2 did not land the forecasted high-WR anchors like Lulu, Xin Zhao or Bard; instead they chose a more fragile, higher-variance structure around Mel and Anivia. Meanwhile, TES avoided the forecasted trap picks such as Azir and Corki, which is meaningful. Because the actual bans and B1 sequence are not included here, the safest comparison is that the strategic expectation was right on style, but the execution diverged sharply through Mel.

Even so, G2 Esports still have a playable edge because Poppy, Anivia and Mel can all distort how TES want to engage. If G2 control first setup, force TES through narrow corridors and keep Hans Sama untouched, their comp becomes easier to teamfight with than the individual matchup numbers suggest. TES win more simply: Tian or fengyue start the fight, Creme reaches Sivir or Mel, and JackeyLove breaks towers before G2's scaling shell is online.

Polymarket Market

Polymarket is the most important external signal here, and it says this decider is dead even. The Game 5 market is G2 Esports 50% — Top Esports 50%, and the Series market now is also G2 Esports 50% — Top Esports 50%. Because this is Game 5, that is not a contradiction: Polymarket does not create a separate per-map market for the deciding game and effectively reuses the series moneyline.

The bigger move is from pre-match. Around 90 min before the series, Polymarket had G2 Esports 34% — Top Esports 66%. G2 have gained +16.5 puntos porcentuales since then, which fits the series flow: after losing G1 8-17 in 34:05 and G2 11-27 in 32:04, G2 answered with G3 29-10 in 27:04 and G4 14-12 in 41:16. The market is pricing momentum, resilience and a full reset into a deciding map. This draft does not fully justify a market swing toward TES either, because while TES have several winning lane matchups, G2's composition has clearer anti-engage structure than the individual WR lines alone imply.

Prediction

The model opened at G2 Esports 56% — Top Esports 44%, and after the draft I would trim that only slightly to G2 Esports 54% — Top Esports 46%. TES have the better lane-by-lane matchup profile on paper, especially through Yone, Maokai and Ziggs, but G2 Esports have the stronger recent form signal in the model at 0.600 versus 0.500, plus the psychological lift of forcing reverse-sweep pressure after the last 2 maps. If Caps survives the first 15 minutes on Mel and SkewMond blocks the first engage windows, G2's draft should scale into the cleaner Game 5 teamfight.