Random seed
The starting number for the random-number generator that draws a simulation's paths.
Explanation
Computers do not produce truly random numbers for simulations. They use a random-number generator: a formula that turns one starting number, the seed, into a long sequence of numbers that look random. The same seed always produces the same sequence.
In a Monte Carlo run, that sequence decides every path: which returns each month brings, and which historical years are picked in a bootstrap. So:
- Same seed + same inputs = exactly the same paths and results. Run it again tomorrow and you get identical numbers.
- Change only the seed and you get a fresh set of futures drawn from the same assumptions.
Why this helps: when you compare two versions of a plan, you want the difference to come from what you changed. If each run drew different random futures, part of the gap would just be luck. Keeping the seed fixed means both versions face the same set of good and bad markets. This is a fair, side-by-side test.
Checking for luck: try a few different seeds. If your success rate or percentiles move a lot between seeds, your results depend heavily on which futures were drawn, and you should increase the number of paths. If they barely move, the result is stable.
Changing the seed is not a way to improve a plan. A seed that happens to show a higher success rate drew friendlier markets; it does not make your future any safer.
The seed is saved in your scenario file, so you or anyone you share the file with can reproduce the exact same results.
Example
You run a plan with seed 42 and 1,000 paths and get an 84% success rate. You raise savings by $500 a month, keep seed 42, and get 89%. Both runs used the same 1,000 futures, so the 5-point gain comes from the extra savings.
Now you check stability by rerunning the original plan with seeds 7 and 1,234 and get 83% and 85%. The results move about 1 point either way, which matches the roughly ±1 point sampling error expected at 1,000 paths for a rate near 85% (√(0.85 × 0.15 ÷ 1,000) ≈ 1.1%). The 5-point gain from saving more is well beyond that noise.