G2 Esports 3-2 Top Esports — MSI 2026 Results & Stats

TL;DR: G2 Esports fell behind 0-2, looked one game from elimination, then ripped the series back with a reverse sweep over Top Esports at MSI 2026. It matters because the comeback was not luck: G2 solved the draft, won the jungle tempo battle, and turned a lost series into a 3-2 statement.

G2 Esports did not just survive against Top Esports on 2026-07-03; they changed the sound of the series. After getting pushed around for 34:05 and 32:04, G2 found the answers, won 3 straight, and transformed a near-collapse into one of the defining reverse sweeps of MSI 2026.

Key Takeaways

• Caps was the series MVP because his control of the middle of the map kept evolving: Syndra in Game 3 for +1190 GoldDiff@15, Twisted Fate in Game 4 at 3/1/7 with a 10.00 KDA, and Mel in Game 5 at 13/4/10. When G2 needed a carry voice, Caps gave them 3 different versions of it.
• The decisive turn came after Game 2, when Polymarket dropped G2 Esports to just 8% in the series. From there G2 answered with a 29-10 stomp, then a 41:16 macro grind, then a 30-14 closer, which is the clearest possible proof that the reverse sweep was built, not gifted.
• The final 3-2 score hides how sharply the momentum swung: Top Esports controlled the first 2 games by a combined 44-19 in kills, but G2 owned the last 3 by a combined 73-36. That split tells the real story of the series better than the scoreboard alone.

Before the Series

The pre-match read said draft would matter from game 1, and that part was dead right. The spotlight fell on Vi, Cassiopeia, Varus, Nautilus, and Rumble, with Vi especially framed as premium because of its 84.2% presence, 68.4% ban rate, and 100% WR in 3 games at MSI 2026. The other key idea was that Top Esports, with their lane-first identity, would be comfortable around Jayce and Cassiopeia, while G2 might need smarter adaptation than raw lane power.

That early script fit the first half of the series. Cassiopeia delivered exactly as predicted when Creme used it to control Game 1, finishing 4/0/6 and helping TES turn a slim draft edge into total map command. Varus and Nautilus also mattered across the series as the kind of stable bot-lane tools that let fights start cleanly. Rumble, though, was the warning sign that prediction 1 was not automatic: in Game 1, BrokenBlade's Rumble was punished hard, and the pick never became the series-defining weapon the pre-draft note hinted it could be.

Game 1 — Setting the Tone

Game 1 sounded close only in champion select. The live draft model favored Top Esports at 52%, and this was one of the spots where prediction 2 held cleanly: the draft edge translated to the result. Once ZUIAN blew open top lane on Ambessa at 8/1/3, TES had room everywhere. Creme's Cassiopeia then supplied the control promised in prediction 1, and the 17-8 kill score by 34:05 felt like the map had narrowed around G2 one corridor at a time.

Game 2 — The Pivot

This was the game that nearly ended the series, and it was also the clearest draft-model miss. G2 were favored at 51%, yet Top Esports won 27-11 in 32:04. That is prediction 2 failing in the most obvious way across the series arc: a small paper edge for G2 vanished under Tian's Jarvan IV at 8/2/16 and Creme's Aurora at 11/0/9.

At 0-2, G2 were not simply behind in score. They were behind in pace, confidence, and the terms of engagement. Top Esports had made the series feel like it would be decided by their acceleration.

Game 3+ — The Climax

Then the series broke open the other way. Game 3 was G2's answer, and here prediction 2 worked again: the 53% draft lean toward G2 became a 29-10 demolition in 27:04. Caps' Syndra took mid with +1190 GoldDiff@15, SkewMond's Skarner finished 5/0/18, and G2 finally made TES react.

Game 4 was where prediction 1 came full circle. Vi appeared in the hands of SkewMond, and the champion delivered exactly as advertised: +1106 GoldDiff@15, early tempo, and the engage structure that let G2 close a tense 41:16 game. The model liked G2 at 52%, and this time the Rift agreed. If Game 3 was the punch, Game 4 was the proof that G2 had actually solved the series.

Game 5 completed the reversal. The model favored Top Esports at 52%, so prediction 2 failed again, but G2's execution was overwhelming. BrokenBlade's Anivia ripped top lane open for +924 gold@15, Caps' Mel exploded for 13/4/10, and SkewMond's Poppy shut down clean entries. The deciding map ended 30-14 in 36:11, not like a coin flip, but like a team that had learned faster than its opponent.

Aftermath

So how should this 3-2 be remembered? As the moment G2 turned adaptation into identity. Caps, SkewMond, BrokenBlade, and the rest stopped playing the series TES wanted and forced a new one built on cleaner engage timing, stronger solo-lane leverage, and better objective structure. Prediction 1 ended mixed but revealing: Cassiopeia and Vi absolutely paid off, Varus and Nautilus stayed relevant as control pieces, and Rumble never truly became a winner. Prediction 2 finished 3/5 overall, which is a reminder that draft edges matter, but only when a team can make them audible on the map.

Polymarket Trajectory

The market began slightly on the side of Top Esports, with G2 entering at 46% in the series, and after 2 games that skepticism looked brilliant when G2 crashed to 8%. Up to that point, the market was reading both the scoreboard and TES's superior pace correctly. Where it lost the thread was in how quickly G2's internal adjustments changed the matchup. It failed hardest in Game 3, when G2 won as a 44% underdog, and it never fully priced in how much control Caps and SkewMond had seized until the series had already tilted back to even. In retrospect, the earlier signal was not just draft percentages, but who was winning the map's first move after draft ended.

Series Stats

FAQ

*Odds via Polymarket, 2026-07-03 14:54 UTC.*