That's the question that started Salsamigo. Turns out, it's not an easy question to answer.

Salsamigo ranks and scores salsas using only head-to-head, blind tastings. Each salsa's win-loss record determines its overall score, and the score predicts whether one salsa will be liked more than another.

Head-to-head comparisons are the simplest way to learn about preferences. Blind tastings help avoid bias for or against familiar brands or packaging.

But don't get too hung up on the rankings. Many of these salsas taste great when you try them on their own. Head-to-head comparisons can bring out differences that you would not otherwise notice.

Salsamigo uses a statistical model, the Bradley-Terry model, to calculate the scores from the head-to-head results. The Bradley-Terry model has been used to rank and compare everything from college football teams to ice cream.

The model factors in the scores of the opposing salsas in the head-to-head comparisons. Wins over highly ranked salsas count more than wins over lower ranked salsas.

You can predict how one salsa will do against another by comparing their scores. Divide a salsa's score by the sum of the scores of the two salsas being compared.

The reliability of a new salsa's score will continue to improve as it is tasted against other salsas, especially those with similar scores.