To calculate the probability of matching exactly three numbers in a simple six-from-49 draw, count the ways to choose three of your six selected numbers and three of the other 43 numbers. Divide that count by all possible six-number results. The answer is about 1.765%, or roughly 1 in 56.66.

That answer applies to the stated mathematical model: six distinct numbers drawn uniformly from 49, with order ignored and no additional prize conditions. It is not a universal “three numbers” probability for every lottery.

The method is useful because it forces you to define the event precisely. A five-number game, a bonus-ball condition, or a question about at least three matches requires a different calculation.

Set up the model before using the formula

Suppose you hold one fixed selection of six distinct labels from a pool of 49. A fair drawing selects six distinct labels from that same pool. Every unordered six-number result is assumed equally likely.

Your selection divides the pool into two groups: six labels you chose and 43 you did not. Exactly three matches means the result contains three labels from each group.

There are C(6,3) ways to choose the matching labels and C(43,3) ways to choose the nonmatching labels. Multiplying those counts gives the number of results meeting both conditions.

The total number of possible results is C(49,6). The exact-match probability is therefore C(6,3) × C(43,3) / C(49,6). This is a counting argument, not a forecast based on previous results.

Work through the numbers

C(6,3) equals 20. C(43,3) equals 12,341. Their product is 246,820 qualifying results. C(49,6) equals 13,983,816 possible results.

Dividing 246,820 by 13,983,816 gives approximately 0.0176504. Multiplying by 100 gives 1.76504%. Taking the reciprocal gives approximately 56.66.

Each form expresses the same probability. The decimal is convenient for further calculations, the percentage is often easier to read, and “1 in 56.66” resembles the format used in many prize tables.

Do not interpret 56.66 as a promise that a match will occur within 57 plays. It is a ratio for a specified event on one entry. Independent sequences can have long gaps, clusters or no such result in a particular sample.

Why both groups must be counted

A common mistake is to count only the three matching numbers. That leaves the other three drawn numbers unspecified, which can accidentally include results with four, five or six matches.

Exactly three requires three successes and three selections outside your ticket. The second combinations term enforces the word “exactly.” Without it, you have not fully described the event.

Another mistake is to choose three matching labels and then allow the remaining positions to use any of the 49 labels. That can repeat labels or introduce additional matches, neither of which belongs to the stated model.

The division into selected and unselected groups avoids those problems. Every qualifying result is counted once: its three labels from your set and its three labels from the complement identify it uniquely.

The general exact-match formula

If a pool contains n labels, your entry selects k distinct labels, and the draw also selects k labels, then the probability of exactly r matches is C(k,r) × C(n−k,k−r) / C(n,k).

The first term chooses the r matching labels. The second chooses the remaining k−r labels from outside your selection. The denominator counts all possible results.

This formula assumes an unordered draw without replacement. It should not be applied unchanged to a digit game where repeated digits or exact order matter. It also does not automatically include supplementary balls or separate special-ball fields.

When a binomial coefficient requests an impossible selection, its count is zero. For example, you cannot choose more matching labels than your entry contains. That convention keeps the formula consistent at the boundaries.

Exactly three is not at least three

“At least three” includes exactly three, exactly four, exactly five and exactly six. These outcomes are mutually exclusive for one entry, so their probabilities can be added.

For the six-from-49 model, exactly three is about 1.76504%. Exactly four is about 0.096862%, exactly five about 0.001845%, and exactly six about 0.000007151%.

Adding those gives an at-least-three probability of approximately 1.86375%. It is larger than the exactly-three probability because it includes additional qualifying outcomes.

A question about “three or more” and a prize row labeled “match three” may therefore have different answers. Read whether the prize row describes exactly that match count or a broader event.

A smaller example you can enumerate

Consider a game that draws three labels from six. Your entry is 1, 2, 3. The other labels are 4, 5, 6. There are C(6,3) = 20 possible results.

Exactly two matches can happen in C(3,2) × C(3,1) = 9 ways. The probability is 9/20, or 45%. Exactly three matches has just one result, 1–2–3, giving 5%.

At least two matches therefore has probability 50%. The difference between 45% and 50% is the complete-match outcome. You can list all 20 results to verify the count.

This small model is useful for checking a spreadsheet formula. If your calculation does not return nine exactly-two outcomes, something is wrong with the arguments or with your interpretation of order.

What changes with a separate special ball?

Suppose a game has a main-number field and an independent special ball drawn from b equally likely labels. If the event requires exactly r main matches and a special-ball match, multiply the main-field probability by 1/b.

If it requires exactly r main matches and a special-ball miss, multiply by (b−1)/b. If the question is about main matches regardless of the special ball, do not add either restriction.

For example, in a hypothetical game with the six-from-49 main field and ten special labels, exactly three main matches plus the special ball would have one tenth the main-only probability.

That hypothetical game is a teaching device, not a description of a named operator. A same-pool supplementary ball has a different relationship to the main draw and must be modeled using its own rules.

A spreadsheet calculation

For the six-from-49 example, a spreadsheet with COMBIN can use COMBIN(6,3)*COMBIN(43,3)/COMBIN(49,6). Format the result as a percentage rather than multiplying by 100 and also applying percentage formatting.

To calculate the complete match distribution, put r values from zero through six in a column and apply the general formula to each. The probabilities should add to one, apart from tiny numerical rounding.

That sum is an important validation check. The seven exact-match categories cover every possible result, and no result belongs to two different exact-match counts.

A second check is the expected number of matches. Multiply each match count by its probability and add. For this model, the answer is k squared divided by n, or 36/49, approximately 0.7347 matches per entry.

Why a common match category does not identify better numbers

Every fixed six-number selection has six chosen labels and 43 unchosen labels. The exact-match count therefore does not depend on whether the selection looks balanced, contains a sequence, or uses dates.

Historical frequency rankings do not appear anywhere in this calculation. The formula describes the fair uniform model. Replacing a “cold” label with a “hot” one does not change the number of labels in either group.

A set of several entries can have a more complicated joint distribution because the entries may overlap. You should not assume that knowing the single-entry distribution makes all tickets independent within one drawing.

Our odds calculator handles jackpot coverage for distinct complete combinations and repeated independent drawings. It does not claim to calculate the full partial-match distribution of an arbitrary ticket portfolio.

A practical reading checklist

When you see a probability for “matching three,” ask how many main labels are drawn, how large the pool is, whether order matters, whether labels can repeat, and whether a special-ball condition is included.

Then ask whether the event is exactly three or at least three. Finally, confirm whether the calculation concerns one entry, several entries in one draw, or repeated drawings.

Those details are the model. Leaving one out can produce a mathematically correct answer to the wrong question. A precise description is more valuable than a long decimal attached to an ambiguous event.

For named games, start with the official format and prize definitions. Our Powerball guide and Mega Millions guide explain why a special-ball distinction matters.

Try the method yourself

In the three-from-six teaching game, calculate exactly one match. Choose one of your three labels and two of the other three: C(3,1) × C(3,2) = 9 qualifying results. Divide by 20 to obtain 45%.

Exactly zero matches has one result, 4–5–6, and exactly three also has one. The distribution is therefore 5%, 45%, 45%, 5% for zero through three matches. It sums to 100%.

Now compare exactly one with at least one. At least one is 95%, because only the zero-match result is excluded. This exercise shows how a small change in wording can produce a large change in probability.

Once the event is clearly defined, the arithmetic follows a simple pattern: choose what matches, choose what does not, and divide by every possible result. That is the foundation for reading partial-match odds accurately.

Check that the exact-match categories exhaust the possibilities

A full exact-match distribution provides a stronger check than one isolated formula. In the six-from-49 teaching model, calculate the probabilities for zero through six matches using the same method.

Every possible draw belongs to one and only one of those seven categories. Their total must therefore be one. If it is larger, some outcomes were counted more than once. If it is smaller, some outcomes were omitted or a term was calculated incorrectly.

You can also inspect the extreme categories. Exactly six matches has one qualifying result: your own selected set. Exactly zero requires all six drawn labels to come from the 43 outside your set. Those endpoints provide intuitive checks on the general formula.

For a small three-from-six model, write every result on a card and sort the cards into boxes labeled zero, one, two and three matches. The box counts should be one, nine, nine and one. No card should appear in two boxes.

This physical exercise makes the denominator visible. Probability is the fraction of all cards in the desired box, not the number of ways to select a few matching labels in isolation. Once the complete partition is correct, adding boxes gives broader events such as at least two matches without double counting.

Browse all ticket-checking and probability guides in the Lottery Learning Library.

Sources and further reading

• Powerball official format for comparison with the teaching model
• Mega Millions official match categories

Reviewed October 18, 2026. Examples are illustrative unless identified otherwise. Winning Numbers Today is an independent information website; the selling lottery validates tickets and pays prizes.