Articles

Two labels do not always mean equal probability

Two labels need not have equal probability. In this hypothetical example, three distinguishable tokens are equally likely to be selected: A1, A2 and B1.

UpdatedEditorial review

Two labels need not have equal probability. In this hypothetical example, three distinguishable tokens are equally likely to be selected: A1, A2 and B1. Each token has probability 1/3. The printed letters group the tokens; they do not describe equally sized groups.

List the outcomes before grouping the labels

The complete sample space is:

TokenPrinted labelProbability
A1A1/3
A2A1/3
B1B1/3

There are three equally likely elementary outcomes, not two. The event “label A” contains the set {A1, A2}, while the event “label B” contains {B1}. Therefore:

P(A) = 2/3
P(B) = 1/3

Events group outcomes; counting requires equal likelihood. Reference: OpenStax terminology for sample spaces and events.

Why counting label names gives the wrong ratio

If someone counts only the visible names, they see A and B and may write 1/2 for each. That calculation silently changes the sample space from three tokens to two labels. It also discards multiplicity: label A appears on two distinct equally likely tokens, whereas label B appears on one.

You can check the grouping by assigning three counters, one per token: two go to A and one to B. Moving both A counters into one labelled box does not turn them into a single counter. The two groups still exhaust all three possibilities.

What changes when the tokens change?

Consider a second explicitly hypothetical model with only two distinguishable tokens: A1 and B1. Keep the same uniform-token assumption: each token is equally likely. Now the sample space is {A1, B1}. Label A contains one outcome and label B contains one outcome, so:

P(A) = 1/2 and P(B) = 1/2.

The ratio changed because the list now contains one token per label. Before assigning probabilities to a screen's categories, identify the underlying outcomes and whether they are equally likely. Merely seeing two names cannot establish that information. This model is an illustration, not evidence about a real game's mechanism.

For a different question about interpreting past observations, see our history-screen reading note. The hypothetical counts above are not observed results.

Reading note:This article is for information, rule reading, and risk awareness only. It does not provide account, payment, or betting services, and it does not promise any result. Adults should follow applicable local rules.
Telegram