Let the four candidates be A, B, C, and D.
| #voters | their vote |
|---|---|
| 28 | A>B>C>D |
| 25 | B>C>D>A |
| 24 | C>D>B>A |
| 23 | D>C>B>A |
In this situation, A would lose to any opponent in a head-to-head election by a huge 72-to-28 margin, far larger than the hugest "landslide" in US presidential election history. And A is ranked dead last by 72% of the voters.
But thanks to the insanely idiotic plurality, aka "first past the post" voting system, A is elected with 28% of the vote, beating out everybody else (each with ≤25%).
In the 3-candidate random elections model this happens in approximately one in six plurality 3-candidate elections. It is plausible that it happened in the Nicaraguan 2006 presidential election, and it definitely happened in the 2009 Burlington (VT) mayor election among the top 3 candidates.
In this example, B beats each opponent pairwise by at least a 53-47 margin and hence presumably "should" have been the easy winner; and indeed B would have won with both Condorcet and Borda voting. On the other hand C would have won IRV voting, using these ballots. (B or C also would usually have won with range or approval voting, but we cannot which tell with certainty because these ballots have not stated the voters' scores or approval-decisions on the candidates.)