| Name | Prize ▼ | Deadline | EV/hr | Domain | Diff | Format |
|---|
EV = (prize for realistic placement) x (probability of that placement) / (hours needed). Per-unit payout models (HackAPrompt, audit contests) rank highest because they pay proportionally to output, not rank.
| Hackathon | Prize Pool | Competitors | P(win) | E[Prize] | Hours | EV/hr Low | EV/hr High | Risk |
|---|
Hackathon outcomes are mostly uncorrelated. Diversifying across 3-5 competitions turns low-probability bets into a portfolio where P(earning something) is high.