And vice-versa; Clinton won in 1992 partly due to the (R) vote being split between Bush and Perot. Or, more accurately, more people who would have voted for Bush (or not voted) than people who would have voted for Clinton (or not voted) voted for Perot. Maybe. You see how complicated this is? Without Perot in the 1992 election it's impossible to say what would have happened -- would the Perot voters have stayed home, or voted for Clinton, or Bush? Even a survey at the polling locations couldn't tell for sure.
There have been other elections with "independents" where the vote was split in odd ways, like the 2006 gubernatorial election in Texas, where Rick Perry (the incumbent, on the (R) ticket) was up against Carole Keeton Strayhorn, a (R) who skipped the primaries since she couldn't win them, the (D) candidate Chris Bell, a libertarian candidate, the truly independent and famous (in Texas) Kinky Friedman, and a write-in campaign for someone forgettable. The vote broke down as:
39% Perry (R)
29.8% Bell (D)
18% Strayhon
12.6% Friedman
0.6% Libertarian
Now, whose votes did Kinky Friedman "steal"? And whose did Strayhorn? And what would have happened with an IRV system? And how many elections in the U.S. would be different (in ways good and bad) with an IRV election?