The headline number is an estimate
Suppose 40 of 61 historical examples moved in one direction. The observed rate is about 65.6%. It is tempting to treat that as an exact property of the pattern. In reality, a different set of 61 comparable examples would probably produce a somewhat different rate.
That movement is not a mistake. It is part of learning from a limited sample. A useful presentation shows the estimate while making clear that the true repeatable tendency, if one exists, could be lower or higher.
Precision has to be earned
More informative, representative observations usually narrow uncertainty. Small samples leave a wide range of plausible answers. Dependence, selection effects and changing market conditions can make standard calculations look more precise than the real situation deserves.
This is why extra decimal places are not the same as extra knowledge. Writing 65.5738% does not make the evidence more exact. Good rounding and a visible interval can be more honest than a screen full of digits.
Uncertainty is not the same as ignorance
Saying that a result is uncertain does not mean saying that we know nothing. It means being specific about what the data can support. A wide range may still rule out some possibilities. A narrow range may still sit around an effect too small to matter after costs.
Statistical significance is another separate question. A detectable difference can be unstable, economically unimportant or produced by a biased design. Tradour therefore exposes constituent evidence rather than compressing everything into a mysterious confidence score.