This estimator tells you how reliably a combo deck assembles its pieces. Enter how many copies of each piece you run and the tool returns the probability that you are holding the full combo by a given turn, using the same hypergeometric math the pros use to tune consistency.
How it works
For a single combo piece with k copies in an N-card deck, the chance of
drawing at least one copy within d cards seen is the complement of drawing
none:
P(at least one) = 1 − C(N − k, d) / C(N, d)where C(n, r) is the binomial coefficient “n choose r”. To hold every piece of
a multi-card combo, the tool multiplies the per-piece probabilities together:
P(full combo) = Π over pieces of P(at least one of that piece)Cards seen d depends on play or draw: on the play d = 7 + (turn − 1), on the
draw d = 7 + turn.
Example and tips
A two-card combo with 4 copies of piece A and 1 copy of piece B in a 60-card deck, on the play by turn 4, has roughly a 39 percent chance per the lone B copy dominating the result. Adding a second copy of the rare piece, or a tutor that fetches it, dramatically improves the assembly odds. Because the multiply step ignores that draws compete for slots, treat the figure as a slightly conservative estimate of true consistency.
Reading the probability output — what the numbers mean
A probability of 39% does not mean you will assemble the combo by turn 4 four times in every ten games. It means that in a single shuffled game, there is a 39% chance you hold all pieces at the start of your turn 4 draw step. In practice, over many games you should see the combo assembled roughly that often — but variance in any small sample is high.
Some useful benchmarks for thinking about combo consistency:
| Assembly probability by target turn | Assessment |
|---|---|
| Below 20% | Very inconsistent; you will often miss |
| 20–35% | Inconsistent; support with tutors or redundancy |
| 35–55% | Moderate; reasonable for a combo subtheme |
| 55–70% | Solid; this is a reliable combo deck strategy |
| Above 70% | Very consistent; primary win condition territory |
Copy count matters more than anything else
The single most powerful lever for combo consistency is the number of copies of the least-represented piece. In a 60-card deck:
| Copies of a key piece | Probability of seeing at least one by turn 7 (on play) |
|---|---|
| 1 copy | ~55% |
| 2 copies | ~80% |
| 3 copies | ~92% |
| 4 copies | ~98% |
This is why tutor effects are so powerful in combo decks — a tutor that finds any combo piece functions like adding copies of all pieces simultaneously. If your combo has three pieces and you run 4 copies of two and 2 copies of one, the rare piece caps your overall consistency even if the common pieces are maxed out.
Two-card versus three-card combos
Adding a third combo piece to the equation multiplies the per-piece probabilities, which quickly reduces the combined assembly chance:
Example: three-piece combo, each at 4 copies in 60 cards, by turn 7 (on play, cards seen = 13)
- Piece 1 (4 copies): ~98% of seeing at least one
- Piece 2 (4 copies): ~98%
- Piece 3 (4 copies): ~98%
- Combined estimate: 0.98 × 0.98 × 0.98 ≈ 94%
But if one piece is only a 2-of:
- Piece with 2 copies: ~80%
- Combined: 0.98 × 0.98 × 0.80 ≈ 77%
Three-piece combos with restricted-copy pieces are significantly harder to assemble consistently. This is why competitive three-piece combos usually include tutors, cantrips, or card selection that effectively increases the functional copy count.
Accounting for tutors and card filtering (manually)
This tool does not model tutors or scry effects, but you can approximate them:
- A tutor for any combo piece is roughly equivalent to adding 1 copy of every piece you could find with it. If Demonic Tutor can find piece B (1 copy), running the tutor is similar to having 2 copies of piece B for probability purposes.
- Cantrips that let you draw and select (like Brainstorm or Ponder) can be modeled approximately by treating your effective hand size as larger. If Ponder lets you look at 3 cards and keep the best 1, it is somewhat like drawing 2–3 extra cards for probability purposes.