What each base is actually worth — ~21,000 simulated games across 3 meta decks

Up front: this is simulator data from a tool I built (gundambay.com), not tournament results. It's a rules engine that plays both decks out with a deterministic in-house policy (no LLMs, no machine learning, just a play algorithm I wrote), so treat it as directional, not gospel. 100 games per matchup is enough to see big gaps but not small ones, so I've marked which differences are real and which are inside the noise.

How the sim works (and why I trust it for this): it's not guessing or curve-fitting — it actually plays the games. Both decks get dealt out and played turn by turn under the real rules: resources, timing, keywords, blocking, all of it. It makes a reasonable decision at each step and plays a matchup out to a win or loss, then does that hundreds of times. The comparison is fair because every variant is played the exact same way against the exact same opponents on the same seeds, so when a number moves, it's the card that moved it. And running it hundreds of times averages out the good and bad draws, so a real gap survives and a fake one washes out. It won't catch every line a top player would find, so read it as "what happens when both decks are piloted competently and consistently."

What I wanted to know: are bases worth their slots, or are you better off cutting them for more units and events? I took three meta decks and cut each base type on its own, holding the other 49 cards as fixed as I could, so you can see what each individual base is worth instead of just "more bases vs fewer." 100 games per matchup, same seed, and the same 5 meta opponents for every variant: two Amuro Ray / Gundam lists (one of them running Zaku Ⅱ), two different Gundam Lfrith builds, and a Marida Cruz / Kshatriya / Banshee (Destroy Mode) list.

Two things to keep in mind while reading:

• Every cut also swaps in replacement cards, so each number is the base plus its replacement, not the base in a vacuum. For Zeon and Minerva I list one refill per row; for White Control I tested every valid refill and averaged them (more on that below).
• Copy count matters as much as the card itself — 3 copies of a base you draw reliably, 2 copies you often don't.

R/G Zeon Ping

Decklist — runs 4 Falmel (ST03-016), so I swept the count, refilling each freed slot in a fixed order: Zeong (GD04-017) up to 2 copies first, then Turning Point of History (GD02-104), then Sinanju (ST03-001).

The gap I trust here is top to bottom: 4 Falmel beats 0 Falmel by 7 points, well clear of the noise. The steps in between are each only about a point, too small to call individually, but all four point the same direction, which is exactly what you'd see if every copy is quietly earning its slot. No sweet spot below four.

White Control (V-Gundam / Unicorn)

Decklist — 5 bases: Corsica x2, Reineforce Jr. x3.

Rather than pick one arbitrary refill per cut, I filled the freed slots with every valid combination from a fixed pool — +2 Graceful Demeanor (GD04-117), +2 A Show of Resolve (GD01-100), +2 Striker Pack (ST04-012), +1 Victory Gundam (GD04-011) — and averaged across all of them. So the average column is the real signal: "what cutting this base gets you, regardless of refill." The best/worst columns are just the high and low of that spread.

One caveat on those best/worst numbers: when you measure 13 refills at 100 games each, the highest one is partly just the luckiest one. At this sample size that top-of-spread figure is biased upward, so don't read the 61% as "this exact list is a 61% deck" — read it as "cutting the 3 Reineforce lands around baseline on average, and the refill choice swings it several points either way."

Across the whole spread, not just the lucky end: cutting the 2 Corsica is a real ~3-point loss no matter how you fill it, while cutting the 3 Reineforce averages at or slightly above baseline. And the refill ranking was consistent — in every cut, the best combination ran Graceful Demeanor at the full +2, and the worst leaned on A Show of Resolve. For the Reineforce cut specifically, the three refills that maxed Graceful Demeanor landed at 58.5%, 58.7% and 61.1%, all clearly above the 55.0% baseline. Going fully baseless still loses on average (~49%), but a Reineforce-for-Graceful-Demeanor swap is a genuine sideways-to-up move, not a sacrifice.

Minerva / Barbatos

Decklist — 4 bases.

About that 44.9% baseline: this gauntlet is a rough neighborhood for the deck (the Amuro Ray / Zaku Ⅱ list farms it at ~33%). That's fine for this experiment — every variant plays the exact same field, so the number to read is the change from baseline.

Here cutting either base gained ~4, and going fully baseless still came out ahead of baseline. It's a fast Barbatos deck that wants to be attacking, and the bases ask it to spend tempo it would rather put into damage. The ~4-point gains are near the edge of what I'd trust at 100 games, but they point the same way across every matchup. And if you're wondering why both cuts together (+1.8) scored below either cut alone (+4): those three variants are too close for this sample size to rank against each other — the solid finding is that all three beat baseline, not which cut is best.

Takeaway

"Bases are filler, shave them for action" isn't true across the board, but it isn't simple either. Zeon clearly wants its bases: every Falmel earns its slot. Aggro Barbatos clearly doesn't. White Control sits in between — cutting a base is fine, or even an upgrade, if you refill into the right card, and a downgrade if you don't. The lever that moved the number most there wasn't the base at all, it was the replacement. So before you shave a base "for action," check what the action actually is.

One aside on Corsica (anecdotal — the list is illegal)

The run above already hints Corsica is the base pulling weight here — cutting its 2 copies costs more per copy than cutting the 3 Reineforce. Corsica is also the only restricted card in the game (max 2, per the official banlist), so I ran a what-if with Corsica at 4 and Reineforce at 1 (still 50 cards, still 5 bases) just to see what the cap was hiding:

Up in all five matchups. So Corsica is the stronger base per copy — you just can't run enough of it to lean on it, which lines up with Bandai's stated reason for the restriction (it was showing up in a majority of blue decks, and those decks were winning). Don't go build the illegal deck; it's only here to show why the base-count numbers undersell that card.

The sim's at gundambay.com if you want to run your own list. Happy to share the method or rerun anything if you think a variant's set up wrong.