The opening round of the Grand Round Robin produced a fast, volatile, and revealing 14-game slate. Round 1 featured early combo kills, combat blowouts, control failures, lock pieces, and one major deck-discovery moment that may affect how the field is evaluated going forward.

The final scoreline was Gavvin 10, ChatGPT 4, but the round did not begin that way. ChatGPT opened hot, winning three of the first four games. From there, Gavvin took control of the round, winning nine of the final ten.

That shape matters. Round 1 was not a wire-to-wire rout. It was a round where ChatGPT started ahead, then Gavvin repeatedly converted the more complicated middle and late games.

Round 1 Results

GameWinning DeckPilotLosing DeckTurnWin Type
1Almost Done with My TurnChatGPTJor Kadeen and His Amazing Friends5Scoop
2Bad Hare DayChatGPTShirei Stax6Combat
3Terra Test OneGavvinBrudiclad’s Counterfeit Catastrophe6Scoop
4Angry OmnathChatGPTBrutal Brago8Combat
5Azami WizardsGavvinDaretti Artifact Fun9Combo
6Toph, Broken MetalbenderGavvinExperiment Kraj7Combo
7Grey GooGavvinLive and Let Die5Combat
8Grimgrin’s Undead HordeGavvinXiahou Dun Combo/Control8Combo
9Death & TaxesGavvinHello Again 3.012Combat
10Silver Bullet MagazineGavvinShelob’s Legendary Taxidermy6Combo
11Where Are We Going and Why Are We in a Handbasket?GavvinSliver Overlord9Scoop
12Sultai ElfballChatGPTMarchesa5Combat
13Bring Out Your DeadGavvinWort Goblin Shenanigans10Lock/Combat
14Yedora’s Compost PodGavvinAcademic Dishonesty7Combat

Final pilot record:

PilotRecordWin Rate
Gavvin10-471.4%
ChatGPT4-1028.6%

Games 1-12 and 14 used GPT-5.6 Sol with the High reasoning setting. Game 13 used ChatGPT 5.5; its reasoning setting was not recorded.

Fast Starts Defined the Round

Several games ended before the opponent could fully stabilize. Almost Done with My Turn opened the tournament with a turn-five scoop after generating more than sixteen Alaundo activations in one turn. Bad Hare Day followed with a turn-six combat win over Shirei Stax, and Angry Omnath gave ChatGPT a 3-1 lead by beating Brutal Brago with visible lethal combat damage.

Those early results made ChatGPT look dangerous immediately. The wins were direct, fast, and decisive.

But the round shifted after that.

Grey Goo produced one of the loudest finishes of the day, beating Live and Let Die on turn five with two 14/14 trampling Skittering Cicadas and multiple Blightsteel Colossi making lethal infect unavoidable. Sultai Elfball also showed how quickly the field can punish incomplete defenses, beating Marchesa on turn five after Marchesa survived an earlier Craterhoof swing but could not survive the follow-up.

The opening round made clear that this tournament field is not slow. Several decks can put the opponent into must-answer territory almost immediately.

Combo and Engine Decks Performed Strongly

Combo finishes were a major part of Round 1’s identity.

Azami Wizards defeated Daretti Artifact Fun after Trickbind stopped Nevinyrral’s Disk, allowing Azami to convert through Palinchron, Psychosis Crawler, Magus of the Future, and Sensei’s Divining Top.

Toph, Broken Metalbender defeated Experiment Kraj by generating infinite colorless mana with Storm Cauldron and an earthbent Sol Ring, then using Staff of Domination to draw through the deck.

Grimgrin’s Undead Horde assembled Geralf’s Messenger, Ashnod’s Altar, Nim Deathmantle, and Diregraf Captain to defeat Xiahou Dun Combo/Control.

Silver Bullet Magazine also showed high-end combo efficiency, using Ashaya, Magus of the Candelabra, and Gaea’s Cradle to generate infinite green mana, then converting through Wooded Bastion, Emiel, Sisay, and Kamahl.

These wins did not all come from the same strategic family. Round 1 showed artifact loops, sacrifice loops, land-mana engines, commander engines, and toolbox inevitability all functioning at tournament speed.

Control and Stax Had Mixed Results

The slower interactive decks had a more uneven round.

Death & Taxes produced one of the most notable wins, defeating Hello Again 3.0 on turn twelve despite severe mana issues. Mangara’s permanent-exile pressure and a final all-creature attack showed that the deck can still grind and close through awkward development.

Bring Out Your Dead defeated Wort Goblin Shenanigans through a Night/Maha lock that suppressed Wort’s ability to function, then converted by using Beast Within on its own Night to create a lethal attacker. That game also included noted Wort pilot errors, so it should be treated as both a deck result and a pilot-execution data point.

On the other side, several control-oriented decks failed to find the necessary second piece of interaction. Academic Dishonesty, in particular, interacted once against Yedora but could not keep pace once Yedora’s recursion engine came online.

The Biggest Deck Discovery: Yedora

The most strategically important discovery of Round 1 may have come in the final match: Yedora’s Compost Pod versus Academic Dishonesty.

Yedora’s draw was strong, but the game revealed more than a lucky opening. The deck assembled overlapping pieces that all rewarded the same behavior: sacrifice creatures, return them as Forests, trigger landfall, generate mana, and convert those triggers into more board presence.

The key interaction was:

Yedora + sacrifice outlet + Tireless Provisioner + Springheart Nantuko bestowed on Avenger of Zendikar

Once Springheart Nantuko was bestowed on Avenger, Yedora’s sacrifice loops became dramatically more dangerous. Sacrificed nontoken creatures returned as face-down Forests, triggering both Provisioner and Avenger. Provisioner generated Treasure, the new Forest could tap for green, and those resources could pay Nantuko’s {1}{G} cost to create additional Avenger copies.

By the end of the sequence, Yedora had produced:

  • 5 Avengers of Zendikar
  • 45 Plants
  • Multiple Plant groups ranging up to 13/14

Yedora attacked with six 13/14 Plants. Saruman blocked two, but the remaining four dealt lethal damage.

This was not merely a strong performance. It uncovered an engine that had not been fully mapped before the match. That matters for future evaluation. Yedora appears less dependent on specific tutor lines than it first seemed, because many of its cards function as overlapping roles rather than isolated combo pieces.

Player-Level Analysis

The player results were lopsided, but the underlying pattern is more useful than the raw 10-4 score.

ChatGPT’s wins were faster on average:

PilotAverage Win TurnMedian Win Turn
ChatGPT6.05.5
Gavvin7.97.5

This suggests ChatGPT was most successful when piloting decks that presented direct, visible pressure early. The ChatGPT wins came from Almost Done with My Turn, Bad Hare Day, Angry Omnath, and Sultai Elfball: decks that established clear momentum and converted without requiring especially intricate late-game navigation.

The loss pattern tells a different story:

PilotAverage Loss TurnMedian Loss Turn
Gavvin6.05.5
ChatGPT7.97.5

Gavvin’s losses tended to be faster games where the opposing deck simply got ahead quickly. ChatGPT’s losses lasted longer, which suggests those games often reached more complex midgame states before slipping away.

The win-type distribution reinforces that distinction.

Gavvin’s wins included:

  • 4 combo wins
  • 3 combat wins
  • 2 scoop wins
  • 1 lock/combat win

ChatGPT’s wins included:

  • 3 combat wins
  • 1 scoop win
  • 0 recorded combo wins

From a neutral standpoint, Round 1 suggests that Gavvin was better at converting complex board states and multi-piece engines, while ChatGPT performed better when the path to victory was already clear and board-visible. ChatGPT’s strategic evaluations were often reasonable, but there were more state-management and sequencing issues, especially in games involving hidden zones, stacked triggers, copied objects, life-total corrections, or complicated engines.

Statistical Interpretation

The 10-4 result is meaningful but not decisive. A 14-game sample is still small, and Commander games include substantial variance from opening hands, matchup pairings, deck power, play/draw dynamics, and pilot assignment.

If every game were treated as an independent 50/50 pilot contest, Gavvin would be expected to finish around 7-7. Instead, Gavvin finished 10-4.

Under that simplified 50/50 assumption:

  • Probability of Gavvin going 10-4 or better: about 9.0%
  • Probability of a result this extreme in either direction: about 18.0%

That is not enough to prove a persistent pilot gap. It is enough to justify watching the trend closely.

The strongest data point is not only the win rate. It is the combination of win rate, win diversity, and conversion quality. Gavvin won through combo, combat, lock pressure, and complicated engine sequencing. ChatGPT won mostly through faster and more direct board pressure.

Round 1 Takeaways

Round 1 established that the GRR field is fast, powerful, and strategically uneven in interesting ways. Several decks can end games before turn seven. Others can survive longer and then convert through layered engines or lock pieces.

The tournament is unlikely to be decided only by raw deck strength. Pilot execution is already visibly affecting outcomes.

The most important deck-level development was Yedora’s emergence as a more dangerous engine deck than previously documented. The most important player-level development was Gavvin’s stronger conversion rate in complex games after ChatGPT’s early 3-1 start.

At the same time, the statistical sample remains early. A 10-4 opening round is a strong signal, not a final verdict. Future rounds will clarify whether this was variance, matchup distribution, deck assignment, or a real pilot-performance trend.

Round 1 did not settle the tournament.

It gave it shape.