Powerball · Backtesting lab
Would a strategy have beaten random selection?
Seven selection rules were run over the full 1,984-draw Powerball history with a strict rule: when a strategy picks numbers for a draw, it only sees the draws that came before it — never the future. Results are compared against a seeded random baseline, so “beat chance” has a concrete meaning.
| Strategy | Avg matches | vs random | ≥3 match | ≥4 match | ≥5 match | Full match |
|---|---|---|---|---|---|---|
| Random baseline Picks k numbers uniformly at random each draw. | 0.383 | 0.000 | 0.31% | 0.05% | 0.00% | 0.00% |
| Most frequent Picks the k numbers drawn most often in the past. | 0.380 | -0.003 | 0.41% | 0.00% | 0.00% | 0.00% |
| Least frequent Picks the k numbers drawn least often in the past. | 0.264 | -0.120 | 0.10% | 0.00% | 0.00% | 0.00% |
| Longest gap Picks the k numbers absent for the longest. | 0.265 | -0.119 | 0.15% | 0.05% | 0.00% | 0.00% |
| Most recent Picks the k numbers drawn most recently. | 0.382 | -0.002 | 0.15% | 0.00% | 0.00% | 0.00% |
| Pair co-occurrence Greedily builds a set using the most co-occurring pairs. | 0.372 | -0.012 | 0.20% | 0.00% | 0.00% | 0.00% |
| Balanced mix Most frequent while balancing odd/even and high/low. | 0.381 | -0.003 | 0.36% | 0.00% | 0.00% | 0.00% |
How the test works
- Start after a training period of at least 20 draws.
- For each draw, every strategy picks 5 numbers using only the draws before it — the future is never visible (no lookahead).
- Count how many picked numbers matched the actual draw.
- Repeat for all 1964 evaluated draws; average the results.
Why the random baseline matters
Any strategy picks 5 of 69 numbers, so even blind luck matches about 0.38 numbers per draw. A strategy only shows anything at all if it clears that bar consistently — and on a fair lottery, none of them do. Backtest results describe the past; they change nothing about the next draw.
Dataset version 1357a99f · calculation version 1 · deterministic seeds.
What this doesn't mean
- Historical frequency and patterns do not change the probability of any future draw — every combination is equally likely.
- This tool does not predict, guarantee, or improve winning odds. Anyone claiming otherwise is selling something.
- Simulated and backtested results describe the past and model a fair random process; they are experiments, not advice.
- Lottery tickets are a form of entertainment, not an investment. Play within your means.