Question
A courier firm has exactly four scheduled deliveries each day. Its manager reports the following probabilities for the number \(L\) that finish late: 80% of days have at least one late delivery, 45% have at least two, 20% have at least three, and 5% have all four late. A separate report says no deliveries are late on 20% of days.
The reports refer to the same randomly selected day. Recover \(\Pr(L=k)\) for every \(k=0,1,2,3,4\) and decide whether the five published figures can all hold simultaneously. In particular, explain why the four at-least percentages cannot simply be treated as separate probabilities of exactly that many late deliveries.
Get hint
Hint 1
The event \(L\ge2\) is included in \(L\ge1\); the tail events overlap. Subtract neighbouring tails to isolate exact counts.
Check your answer
Solution
Write \(S_k=\Pr(L\ge k)\) and note that \(L\) can take only the values 0 through 4, so \(S_5=0\). The reported tails give \(S_1=.80\), \(S_2=.45\), \(S_3=.20\) and \(S_4=.05\).
Successive differences yield
\(\Pr(L=1)=S_1-S_2=.35\), \(\Pr(L=2)=S_2-S_3=.25\), \(\Pr(L=3)=S_3-S_4=.15\), and \(\Pr(L=4)=S_4-S_5=.05\).
Also \(\Pr(L=0)=1-S_1=.20\), agreeing with the separate zero-late report. The five probabilities are nonnegative and sum to \(.20+.35+.25+.15+.05=1\), so they form a valid distribution. To verify the original cumulative reports, sum the reconstructed masses: \(\Pr(L\ge1)=.80\), \(\Pr(L\ge2)=.25+.15+.05=.45\), \(\Pr(L\ge3)=.15+.05=.20\), and \(\Pr(L\ge4)=.05\).
The at-least events are nested. For example, a day with four late deliveries contributes to every at-least report, so adding those headline percentages as though they were disjoint would double-count days.
Assumptions
Exactly four deliveries per day, so \(L\) is an integer in \(\{0,1,2,3,4\}\).
All listed percentages are unconditional probabilities for the same day-selection process.
Final Answer
\(\Pr(L=0,1,2,3,4)=\mathbf{(0.20,0.35,0.25,0.15,0.05)}\). All reports are consistent.