Build Guide
Chapter 11

Testing and Iteration

How to refine your deck through deliberate testing — goldfishing, live play, what to track, when to cut cards, and learning from established lists.

Testing and Iteration

Building a deck is the first half. Refining it through testing is the second — and it's where good decks become great decks. A list that looks perfect on paper will always have surprises when you actually play it.

Goldfishing

Goldfishing means playing your deck alone, with no opponent. You draw your opening hand, play through your first 2-3 turns, and see how your setup develops. It tests one thing: does my engine work?

What Goldfishing Reveals

  • Brick hands — How often do you open with an unplayable hand? (Acceptable: less than 10% of games)
  • Setup speed — How many turns until your main attacker is ready? (Track the average)
  • Search chains — Can your search cards find what you need, or do you dead-end?
  • Energy attachments — Are you hitting your energy attachment every turn?

How to Goldfish

  1. Shuffle, draw 7 cards for your opening hand
  2. Play out 3 turns as if you went first (no attacking, just setup)
  3. Record: Did you get your attacker ready? Did you brick? What was missing?
  4. Repeat 10-20 times
  5. Track results — you want 80%+ of games with a functional setup by turn 2-3
If you're bricking more than 20% of the time, your engine has a problem before you ever face an opponent.

Live Testing

Goldfishing catches engine failures but misses interactive problems. Live testing against real opponents reveals:
  • Whether your deck can take six prizes before your opponent does
  • Which matchups feel impossible (might need techs)
  • Which matchups feel free (confirms your strategy works)
  • How disruption affects your plan (opponent's hand disruption, ability lock, gusting)

Testing Games vs Tournament Games

Testing games are for learning, not winning. In testing:
  • Play both sides of a matchup (or ask your testing partner to swap decks)
  • Discuss decision points after the game — what mattered?
  • Try different lines of play with the same hands
  • Track results but don't obsess over win rate in small samples
You need 10-15 games against a matchup before drawing conclusions. Fewer than that is noise.

What to Track

Keep notes during testing. The things that matter:

Per-Game Notes

  • Opening hand quality (good/acceptable/brick)
  • Turn your main attacker was ready
  • What card you wished you had (the card you kept looking for but didn't find)
  • What card sat dead in your hand (never played it, never mattered)
  • Win or loss + why (one sentence: "lost because couldn't find energy turn 2")

After 10+ Games

  • Most-wished-for card — If you keep wanting the same card, add more copies or more ways to find it
  • Most-dead card — If something sits in hand game after game, it might not deserve its slot
  • Average setup speed — Is it fast enough to compete?
  • Matchup spread — Which common opponents are favorable/unfavorable?

When to Cut Cards

A card should be cut when:
  1. You never play it. If a card sat in your hand 5 games in a row unused, it's dead weight.
  2. You always use something else instead. If you have two search options and always pick the same one, the other is cuttable.
  3. It only matters in one matchup you rarely face. A tech for a matchup that's 5% of the field doesn't justify its slot.
  4. Adding another copy of something else would help more. The real question isn't "is this card good?" — it's "is this card better than the next copy of something I already run?"
A card should NOT be cut when:
  • It won you the game once and felt amazing. (That's results bias — one moment doesn't justify a permanent slot.)
  • It would be good "if I could just find it." (That means you need more search, not more copies of a situational card.)

Learning from Established Lists

There is no shame in studying established decklists. Tournament-winning decks represent hundreds of hours of testing compressed into 60 cards. They are research material.

How to Use Established Lists

  1. Compare your counts. If every top list runs 4 of a card and you run 2, ask why.
  2. Look for patterns. When 8 out of 10 lists agree on a card choice, that choice has been validated by experience.
  3. Notice what's absent. Cards that never appear in winning lists have usually been tested and rejected.
  4. Understand their meta context. A tournament list was built for a specific expected metagame. Your local meta may differ.

Net Decking vs Understanding

Copying a list exactly is a valid starting point — especially when learning. But a copied list without understanding is fragile:
  • You won't know which cards are flex slots and which are core
  • You won't know how to adapt when the meta shifts
  • You won't know what to side out in specific matchups
The goal is to reach a point where you could reconstruct a competitive list from first principles. That's when you truly understand deck building.

The Iteration Loop

Deck building is never done. The loop:
  1. Build (stencil → filled list)
  2. Test (goldfish → live play)
  3. Identify weakness (track what's missing/dead)
  4. Adjust (cut dead cards, add what's missing, try techs)
  5. Test again (does the change help or hurt overall?)
  6. Repeat until the deck handles your expected metagame
Each pass through this loop makes your deck slightly better. Most competitive players cycle through this loop dozens of times before a tournament. The deck you register is never your first draft.

What Comes Next

You now have the complete mental model for building a competitive deck: from philosophy to structure to testing. The Build guide has given you what to build — the Learn guide teaches you what the metagame demands, and the Play guide teaches you how to pilot what you've built.
Building the deck is the foundation. Playing it well is where you win.