Use one explicit price format

Select one combined CSV or up to eight files with the exact header date,asset,price. Rows contain an exact date, a case-sensitive asset label and a positive price, for example 2025-01-01,A,100. Use one quote currency across all assets and avoid duplicate asset/date pairs. Supply one long-only weight for each label, totaling 100%. The combined maximum is 1 MiB and 16,000 observations.

Intersect prices first

Simple return = current shared-date price / previous shared-date price − 1

The analyser finds dates shared by every asset before it calculates returns between consecutive shared observations. This gives the same holding interval to every return in a pair. It does not correlate raw price levels or fill missing observations. At least three shared price dates are needed for two return intervals. Rows outside the intersection are counted as discarded.

Read the matrix and portfolio together

Pearson correlation uses sample covariance divided by the two sample standard deviations. If an asset has zero return variance, its correlations, including its diagonal, are undefined. The portfolio adds weighted returns using your fixed weights at every interval. Opposite +10%, −10%, +10% returns can cancel at equal weights; zero portfolio volatility leaves the diversification ratio without a finite denominator.

Keep the observation interval visible

Volatility is reported per shared interval, without annualization. Different date gaps and fewer than 30 returns produce explicit warnings. Fixed weights assume rebalancing each observed interval; transaction costs and buy-and-hold drift are excluded. Correlation from a short historical sample does not establish future protection. File selection is processed in the browser. Saved scenarios retain weights, ranges, row counts and a digest; you must select matching files again. Raw prices and filenames are excluded from saving and result export. Without JavaScript, pasted text is sent transiently to this site.