To goldfish a Commander deck, shuffle normally, draw an opening hand, make a real mulligan decision, and play turns as though opponents do nothing. Record land drops, early actions, commander timing, and the first turn the deck can establish or close its plan. Repeat under the same rules so the results can be compared.
Choose one test question
“Is this deck good?” is too broad for solo testing. Ask whether 36 lands and ten ramp pieces reliably cast a five-mana commander, whether the deck has an action before turn three, or whether a revised draw package prevents empty hands. One test can produce one actionable answer.
A repeatable goldfish protocol
- Shuffle thoroughly and draw seven without selecting a favorable hand.
- Apply your normal mulligan standard and record the final hand size.
- Play one land per turn and obey all costs, timing, and color requirements.
- Choose lines using only information the game would provide.
- Stop at a defined turn or milestone, such as turn eight or a deterministic win.
- Log the result, reshuffle, and repeat at least 20 times.
What to record
| Signal | What it diagnoses |
|---|---|
| Mulligans | Land count, color access, and early-play density |
| Missed land drops | Whether mana development supports the curve |
| First meaningful action | How often the deck spends early turns idle |
| Commander cast turn | Whether ramp and fixing support the central engine |
| Cards stranded in hand | Color, cost, or prerequisite problems |
| Setup / win turn | The unopposed ceiling, not expected pod performance |
Add realistic friction
A second test pass can simulate common disruption. Remove the commander once, destroy the first engine, or clear the battlefield on turn five. Apply the same event to every run in that batch. This does not reproduce an intelligent opponent, but it reveals whether the deck has alternate lines and enough recovery resources.
Do not invent perfect opposition one turn at a time. Constantly choosing the worst possible answer creates a different bias and makes runs incomparable.
Know what goldfishing cannot tell you
Solo play cannot measure threat assessment, politics, target selection, interaction timing, combat against blockers, or whether a plan survives three opponents. It systematically favors linear strategies and overstates fragile combos. Real pod games remain the evidence for matchup and win-rate claims.
Turn results into one focused revision
If eight of 20 runs miss the fourth land drop, test a small land or card-access change. If the deck always has mana but no early action, review the curve. Save the revision as a new version, rerun the same protocol, then take it into real games. WUBRG's version history keeps the hypothesis and pod results together.
Commander goldfishing FAQ
How many times should I goldfish a Commander deck?
Twenty opening hands provide a useful first screen; 30–50 make repeated mana and curve patterns easier to see. More runs do not replace real games.
Should I assume opponents use removal?
Run a separate, standardized disruption test—such as removing your commander once—rather than choosing arbitrary answers in every game.
Can goldfishing determine a deck's power level?
No. It can reveal speed and consistency without resistance, but not resilience, interaction, politics, or performance against comparable pods.