- expectation is linear, no independence needed
E(a1X1+⋯+anXn)=a1E(X1)+⋯+anE(Xn),E(aX+b)=aE(X)+b
- if the $X_i$ are identically distributed, $E(X_1 + \dots + X_n) = nE(X_1)$
- variance is not this nice, see [[Variance of Sums]]
- the trick: write a count as a sum of indicators Xj∈{0,1} for events Aj, then E(Xj)=P(Aj)
- example: four students order noodles independently, each picks Japanese pan noodles 40% of the time
- Xj indicates student j orders them, E(Xj)=0.4
- E(X)=E(X1+⋯+X4)=4(0.4)=1.6 students
- example: draw cards without replacement until the ace of hearts, expected number of draws
- directly you’d sum over the whole distribution, indicators are faster
- Aj = at least j draws are needed = no ace of hearts in the first j−1 cards, so P(Aj)=1−52j−1
- the number of draws X=X1+⋯+X52, since X≥j exactly when Aj happens
E(X)=j=1∑52(1−52j−1)=52−521⋅251⋅52=52−251=26.5