How to build a consistent Pokémon TCG deck

Most decks don't lose to their opponents. They lose to themselves — a dead opening hand, a mulligan streak, an attacker that shows up a turn late. Competitive deck building is mostly the craft of making that happen less often, and it is far more measurable than most players treat it.

This article is the overview of a series that teaches deck building from the math up. Every claim in it can be computed exactly from a decklist — no simulations, no vibes — and each section links to a deeper article on that topic.

A deck is a probability machine

A deck is 60 cards. Your opening hand is 7 of them, dealt blind, and every draw after that comes from a shrinking pile. That makes every question about consistency a question about draw odds — and those odds can be computed exactly. Run 4 copies of a card and the chance at least one is in your opening hand is 39.9%. Run 1 copy and it's 11.7%. Each extra copy buys less than the last: going from 1 to 2 copies adds about 10 points, going from 3 to 4 adds about 8. The 4th copy of anything has to earn its slot.

Basic count sets your mulligan rate

You mulligan when your opening 7 has no Basic Pokémon — not a missing Supporter, not a missing Energy. That makes the number of Basics you run the single biggest consistency dial in the list. A 60-card deck with 10 Basics starts about 25.9% of its games with a reveal, shuffle, and redraw; at 14 Basics that drops to 13.9%. Neither number is wrong — but a deck that mulligans a quarter of the time should be doing it on purpose, for power it gets back elsewhere. The mulligan math article walks the full table.

The seat you're in changes the numbers

The player going first skips their first draw, so by any given turn the player going second has seen one card more. That single card is worth more than it sounds. With 4 copies of a key Basic, the chance you've seen at least one by turn 2 is 44.5% going first and 48.8% going second. Every turn-based figure about your deck is really two figures, one per seat — the going-first tax explains why that matters for both building and playing.

Order of plays is free percentage

The same hand played in a different order wins a different number of games. The rule of thumb: play your search cards before your shuffle-draws, thin the deck of everything you don't need to draw, and only then draw. The edge per decision is small — often under a point — but it costs nothing and compounds across a tournament. The sequencing article works a concrete example with real numbers.

Test changes like a scientist, not a fan

Building the list is half the job. The other half is finding out whether it works, and this is where most players fool themselves — in both directions.

• Know what the math predicts. Before you trust your results, know your deck's theory floor: what raw draws alone should deliver, seat by seat. Theory vs practice shows how to compare honestly.
• Judge cards by what they did, not what they promise. A card that sits dead in your hand a quarter of your games is telling you something. Card utilization covers how to read it.
• Distrust small samples. Five games tell you almost nothing; a single matchup needs on the order of 50 games before its win rate means much. Why small samples lie puts numbers on it.
• Separate the list from the pilot. Some losses are variance, some are the deck, and some are play. The six disciplines give you a vocabulary for the last category.

The loop

Put together, competitive deck building is a loop: build with the odds in view, play with correct sequencing, log your games, and read the results against what the math predicted. Where the math and your results disagree is usually where the real improvement is.