How does drawing with replacement differ from drawing without replacement?
Quick answer
With replacement, you return the selected item before drawing again, so the next draw has the same probabilities as the first. Without replacement, the item stays out, so the mix—and the next-draw probabilities—change. For two draws, keep the order in your tree: red then blue is different from blue then red. Multiply the probabilities along each path to find its probability.
The question, the details, the next step.
What does “with replacement” mean, and why does it matter?
It describes what happens to an item after you pick it. With replacement, put it back before the next draw. Without replacement, leave it out. So with replacement the next draw starts from the original mix; without replacement, the first pick changes what remains and can change the next probabilities. [S1]
To keep the setups straight, write down what is in the container before draw one, then ask what is in it before draw two. The same container and number of draws can produce different probabilities depending on whether the first item goes back.
Can you show both setups with the same example?
Assume a bag contains 3 red and 8 blue marbles, and two marbles are drawn at random. This is the example used in the cited tree-diagram material. [S2] The tree records colors in draw order. Each second-draw branch shows its probability given the first color.
With replacement — the bag returns to 3 red and 8 blue after draw one:
Start
├─ Red (3/11)
│ ├─ Red (3/11) → RR: 9/121
│ └─ Blue (8/11) → RB: 24/121
└─ Blue (8/11)
├─ Red (3/11) → BR: 24/121
└─ Blue (8/11) → BB: 64/121Without replacement — the first marble stays out:
Start
├─ Red (3/11)
│ ├─ Red (2/10) → RR: 3/55
│ └─ Blue (8/10) → RB: 12/55
└─ Blue (8/11)
├─ Red (3/10) → BR: 12/55
└─ Blue (7/10) → BB: 28/55For example, without replacement, after a red marble there are 2 red and 8 blue marbles left, so the next red probability is 2/10. Multiply along a path to get its outcome probability: (3/11) × (2/10) = 3/55 for RR. [S2]
What exactly are the sample spaces here? Are RR and RB the actual marbles?
In those trees, outcomes record colors in order, not individual marble identities. The color sample space for two draws is {RR, RB, BR, BB} in both setups. RR means red first, red second; RB means red first, blue second. Order matters, so RB and BR are separate outcomes. [S1][S2]
If you label each of the 11 marbles as a distinct object, the ordered sample space has 11 × 11 = 121 possible pairs with replacement and 11 × 10 = 110 without replacement. A repeated marble identity is allowed in the first setup but not the second. These counts follow from the number of available marbles and whether the first one is returned. The color tree groups identity pairs together; it is simpler when the question asks only about colors.
How do I use a tree for a different problem without mixing up the probabilities?
Write the first-draw branches from the original mix. For each branch, update the contents before adding the second-draw branches: restore the original counts for replacement, or subtract the first item for no replacement. Label every branch with its probability at that stage. To find a complete path’s probability, multiply its branch probabilities. Then check that the probabilities of all final outcomes add to 1.
For a category-only tree, use the same labels throughout (such as red and blue), and make clear that each pair is ordered. If items are not equally likely to be selected, or the selection process is not random, the simple fractions in this example may not apply; define the selection rules first.
A practical checklist
Write the starting counts for each category.
State whether a selected item is returned before the next draw.
Label branches with conditional probabilities for that stage.
Keep the draw order in each final outcome.
Multiply along each path and check that all final probabilities sum to 1.
A few more useful answers
Does “without replacement” always make the next draw less likely to be red?
No. It depends on what was drawn first. Removing a red marble lowers the red share in this example; removing a blue marble raises it. [S2]
Why are RB and BR separate outcomes?
They describe different orders: red then blue versus blue then red. Combine them only if the question asks for “one of each,” regardless of order.
Can I use the same tree if there are more than two draws?
Yes. Add another level for each draw, updating the counts after each selection when there is no replacement. With replacement, keep the original probabilities at each level.
Sources
References used for this guide.
- [S1] Independent and Mutually Exclusive Events — Introductory Statistics openstax.org · Checked
- [S2] Tree and Venn Diagrams — Introductory Statistics openstax.org · Checked