Borda Count

The Borda count is an aggregation and voting method that turns several rank lists (several experts' opinions, several jury members, or the ranks several sub-methods produce) into a single combined rank. Every rank list awards a score based on the position an alternative holds in it: coming last is worth 0 points, each step up adds 1 point; coming first earns the highest score. An alternative's final Borda score is the sum of the scores it receives across all lists. Its output is a score, and a rank based on that score. Proposed by the French mathematician Jean-Charles de Borda in 1781 for election boards, it is one of the best-known positional voting rules.

The idea behind the Borda count is to look not only at "who came first" but at every alternative's position across all the lists. If a voting rule counts only first preferences (as majority rule does), how often an alternative is preferred second or third remains invisible. Borda folds this information into the score too: an alternative widely seen as "second best" can overtake one with narrow but strong first-preference support. The philosophical consequence is that the Borda count is compensatory and consensus-seeking; it rewards broad acceptance across the lists rather than sharp first preferences.

The method proceeds through three steps.

First, gather the rank lists. K rank lists (experts, jury members or sub-methods) are gathered, each assigning a rank number (1 = best) to m alternatives.

Second, score. Every alternative's score on every list is found by subtracting its rank on that list from the total number of alternatives; coming last earns 0 points, coming first earns m minus 1 points. An alternative's final Borda score is the sum of these scores across all lists.

Third, rank. Alternatives are arranged from the highest Borda score to the lowest; the alternative with the highest score comes first.

The formula behind each step and the intermediate tables are given on the DecisionMind method page; this card carries no formulas.

The Borda score shows how broadly an alternative is accepted across the lists; on its own it does not answer "how many people ranked it first." A high Borda score means the alternative was never ranked very poorly on any list and sat near the top on most lists; it does not mean the alternative was the most-preferred first choice. An important feature of the Borda count is this: adding or removing an alternative from the set can change the relative order of the remaining alternatives, because the scores depend on the number of alternatives (m).

Thus instead of writing:

"The alternative with the highest Borda score is the most-preferred first choice"

the report should read:

"The alternative with the highest Borda score is the one most broadly accepted across the lists; this may not be the same as the most-preferred first choice"

The Borda count works with crisp data, and its input is a rank matrix: alternatives in the rows, rank lists (experts, jury members or sub-methods) in the columns, and every cell holding an integer rank from 1 to m, not a raw score. DecisionMind currently holds no fuzzy or other-data-type extension of this method; it works only in its base form. You must ensure every list ranks all alternatives completely; an incomplete ranking distorts the scores. At least two alternatives and two rank lists are required; three to twelve rank lists work comfortably. The method neither produces weights nor requires them; its input is rank numbers directly.

The Borda count is suitable when you have several rank lists and your goal is a combined rank that reflects broad acceptance across the lists, not just first preferences; it is used often in voting, jury assessment, and decision processes with several sub-methods. In a setting where the set of alternatives could change during the analysis (a new alternative might be added later), the Borda count becomes risky, because that addition can change the order of the remaining alternatives. If you only want to find "the most-preferred first choice," the Borda count is unsuitable; simple majority counting is sufficient in that case.

Several rank lists exist, a combined rank reflecting broad acceptance is needed → Borda Count

Same rank lists, but combining by mean position alone is wanted → Mean Rank

Only the most-preferred first-choice alternative is sought → simple majority count

The set of alternatives could change during the analysis → interpret the result with caution, fix the set before the analysis

The Borda count's greatest advantage is that it uses the full rank information, not just first preferences; this makes it a richer aggregation than majority rule. Its computation is simple, requires no extra parameter, and is easy to explain to a decision-maker. Because it can bring forward an alternative that is broadly accepted but never ranked first on any list, it rewards balanced alternatives rather than extreme preferences.

Its limitations follow from this same structure. The Borda count violates the independence of irrelevant alternatives condition; adding or removing an alternative from the set can change the order of the remaining alternatives (Young, 1974). In what is known as the Borda paradox, the most-preferred first-choice alternative may not come first in the Borda ranking; this is a consequence of the method operating on a different logic from majority rule. The scoring treats the distance between rank steps as equal; the gap between second and third on a list is weighted the same as the gap between first and second, even though the actual strength of preference may differ.

The most common mistake is adding a new alternative once the analysis is finished and being surprised that the order of the remaining alternatives changes; the set of alternatives must be fixed before the analysis. A second mistake is reading the Borda score as "how many people made this alternative their first choice"; the score sums position across the lists, it does not simply count first preferences. A third mistake is proceeding to scoring without first correcting a rank list that contains an incomplete ranking; an incomplete ranking distorts the value of m and, with it, the scores.

The governing principle is this:

The Borda score measures an alternative's broad acceptance across the lists; unless the set of alternatives is held fixed and the score is not read as simply "most preferred," the result can be misunderstood.