Deck Building 101: How Our Deck Score Actually Works

Every deck in the Deck Builder gets scored out of 100 the moment it hits 50 cards. Before you trust a number like that, you should know exactly what it's measuring — so here's the honest breakdown, including where it's a real read on your list and where it's a rough heuristic.

This is not a trained model and it's not correlated against real win rates, because no tournament result dataset existed to train it against when we built it. Treat it as what an experienced player would eyeball from your decklist, not a promise of how it'll perform. (Our Meta page is where real win-rate data will eventually live, pulled from actual logged tournament results as stores start reporting them — see our tournament formats post for how those brackets work.)

The seven axes

Each sub-score is 0–100, and they're weighted differently into the overall grade:

• Curve (15%) — rewards a low, smooth cost curve. Decks heavy on 0–2 cost cards score well here; decks stacked with 5+ cost bombs get penalized.
• Consistency (20%) — fewer colors and a higher average copy-count per card (closer to running 4 of everything) both push this up. Splashing a third color for one flashy card costs you here.
• Removal (10%) — how much of your deck can actually answer an enemy Unit, read off combat-damage and destroy-effect keywords in card text.
• Aggression (15%) — average AP across your Units, weighted with curve speed and burst/tempo keywords.
• Defense (10%) — average HP plus Blocker/Repair keyword density.
• Synergy (20%) — the same engine behind the Deck Builder's Synergy Reader: pilot/unit link matches, trait clustering, color focus.
• Resource Efficiency (10%) — draw and search density, discounted if your curve is top-heavy.

Synergy and Consistency carry the most weight (20% each) — deliberately, since a deck that's internally coherent and reliably draws its plan tends to matter more than any single axis in isolation.

Reading the tier label

The overall number maps to a label so you don't have to memorize thresholds:

• 85+ — Competitive
• 70–84 — Solid
• 50–69 — Developing
• Below 50 — Needs Work

Where the keyword-matching gets rough

Removal, Aggression, and Defense partly lean on regex matches against card effect text — looking for phrases like "deals damage," "destroys," "Blocker," or AP/HP modifiers. That's a real signal, but it's still text-pattern matching, not game-state understanding. A card with unusual phrasing for a removal effect might get under-counted; a card that mentions damage in a completely different context could get over-counted. We calibrated the multipliers against real constructed decks so the numbers spread out sensibly instead of everything pegging at 100 (an early version had ~27% of the entire card pool matching the removal keyword set, which meant almost every deck maxed that score out — not useful), but "calibrated" isn't the same as "verified against actual play."

What it's good for

Use Deck Score as a second opinion, not a verdict. If Consistency is dragging your overall number down, that's worth looking at — it usually means you're running too many colors or too many one-ofs. If Synergy is low, check whether your Pilots actually match the Units you're running. It's a fast way to catch a list that looks aggressive on paper but doesn't actually hang together, before you find that out the hard way across the table.