Choose one question before drawing
A solo table becomes most useful when the session has a narrow question. Examples include whether the deck reaches a particular civilization by turn three, whether an opening hand can recover after a slow start, or whether a defensive line leaves enough resources for the following turn. Avoid trying to prove that the entire deck is good in a single game.
Write the question in one sentence and decide what evidence would answer it. This protects the session from drifting into a highlight reel where both sides conveniently draw the cards needed for an impressive sequence.
Set up two honest roles
Load the deck under test on one side and a stable reference deck on the other. The reference deck does not need to be the strongest possible opponent; it needs to represent the pressure, removal, speed, or resource pattern you want to examine. Keep its plan consistent across several runs.
When both hands are visible to one tester, make each decision using only the information that player would normally know. If hidden information changes a choice, note the fork instead of silently choosing the outcome that helps the deck under test.
Use a repeatable turn loop
At the start of each turn, confirm the active player, draw, and any pending state. Before moving a card, say the purpose of the move: develop mana, preserve a response, create pressure, or prepare a later sequence. Use tap, reveal, reorder, shield, and stack controls to keep the visible table aligned with that decision.
- Confirm turn, hand size, mana, shields, and any pending manual task.
- State the active player's best plan using only available information.
- Make the minimum table operations needed for that plan.
- Resolve ambiguous card behavior manually and record the assumption.
- At turn end, note the resource or decision that mattered most.
Test branches instead of replaying whole games
When a decision has two credible options, capture the table state in your notes before choosing. Play the first branch for one or two turns, then restore the position manually or begin a fresh run for the second branch. A short branch comparison often teaches more than finishing two long games with many unrelated differences.
Use the action log and undo only as aids, not as a substitute for understanding the state. If a complex sequence contains hidden choices or random movement, a clean restart may be more trustworthy than repeatedly rewinding individual actions.
Close with a three-line result
End the session when the original question has enough evidence, even if the game is unfinished. Record what happened, why it happened, and what you will change or test next. Separate a deck-list change from a play-pattern change; sometimes the correct result is to keep the list and practice a different mana decision.
- Observation: the repeatable fact seen in the session.
- Interpretation: the resource, timing, or matchup reason behind it.
- Next action: one card change, one play rule, or one follow-up test.