Reading confidence intervals
Understand simulation uncertainty without exaggerating small differences.
A point estimate and a range
A simulated success rate is the share of successful trials and can move slightly with another seed. The 95% Wilson confidence interval uses the observed rate and sample size to show a plausible range for the underlying probability.
It does not mean there is a 95% probability that the fixed value lies in this completed interval. It describes a procedure whose intervals contain the true value about 95% of the time across repeated samples.
Compare two decks cautiously
When intervals overlap substantially, do not declare a tiny decimal difference decisive. Compare with the same seed and more trials. Separate intervals strengthen evidence about the configured opening condition, not the overall strength of either deck.
Choose a trial count
Small runs are enough to validate settings. For a subtle one-card change, increase trials until the interval width is smaller than the difference that would affect your decision. Very rare or nearly certain outcomes also deserve care because one side has few observations.
Share enough context
Keep the success condition, target turn, seed, trial count, and interval with the rate. A percentage alone cannot be reproduced or audited. DrawState result cards place this context in one image.