Decision Tree Expected Value Interview Questions
Decision tree expected value interview prep for staged choices, chance nodes, backward calculation, and branch probability mistakes.
Candidates who struggle with staged decisions and conditional outcomes.
Separate decisions and chance
Decision trees distinguish choices you control from random outcomes you do not control.
Work backward
For a staged problem, compute expected values at later chance nodes first, then roll those values back to earlier decisions.
Concrete example
If you can stop for 4 dollars or flip a fair coin for 10 on heads and 0 on tails, the continue value is 5, so continuing beats stopping in the toy EV model.
Conditional branches
Each branch probability should be conditional on reaching that branch. Do not reuse original probabilities after information changes.
Decision rule
At decision nodes, compare available action values. At chance nodes, average outcomes by probability.
Common mistakes
Candidates often average decision branches as if they were random. Choices and chance events need different treatment.
Practice the pattern
Use the LeetQuidity curriculum and calibration to turn this topic into a focused practice plan.