Example election [from electology.org]:
| #voters | Their Vote |
|---|---|
| 35 | Rightist > Centrist > Leftist |
| 33 | Leftist > Centrist > Rightist |
| 22 | Centrist > Leftist > Rightist |
| 10 | Centrist > Rightist > Leftist |
[E.g. the first row says that 35% of the voters prefer the rightist, over the centrist, over the leftist.]
In this 100-voter example election, under Instant Runoff Voting (IRV), "Leftist" wins. ("Centrist," with the fewest top-rank votes at 32%, is eliminated in the first round then Leftist wins 55 to 45 over Rightist.) This is a fairly realistic scenario of a kind that often arises with voters and candidates positioned along a one-dimensional line.
That's a common flaw exhibited by IRV that tends to hurt centrist parties and instead push countries toward extreme views. This example election's result is probably unfair and bad for your country. Why? Consider Centrist running head-to-head against anybody else.
It's hard to believe these two landslide head-to-head margins were wrong. (Each has greater margins than any US presidential election contest, at least up to year 2015.)
So usually, in this kind of situation, IRV (1) elects the wrong choice, and (2) elects an extreme, not a centrist, choice.