On the perils of 'Instant Runoff' voting - SFGate

Ah, the Chronicle, going to press with inaccuracies.  For the record, Lynette Sweet had the most first place votes (2137, 12.07%), not Kelly. He came in second (2095, 11.83%), Cohen third (2083, 11.77%), a mere 12 votes behind Kelly.  What John Diaz of the Chronicle seems to be saying is that Cohen's victory with "less than a quarter of all votes cast" (24.4%) is somehow better than a Sweet victory with 12.07% of first-choice votes cast (with no runoff).  Either that, or, during a runoff between Sweet and either Cohen or Kelly (there surely would have been a recount for the #2 spot), we somehow would have gotten the 17705 people back to the polls to vote again.  In fact, as Richard DeLeon of BeyondChron points out, 

In December 2001, San Francisco paid approximately $3 million to hold a runoff election in which 70,000 voters, only 17 percent of those registered, turned out to elect the city attorney. Turnout plunged all over the city, especially in minority precincts.

In this 2010 election, 60% of registered voters voted.  17% of registered voters in a December runoff this year would have been 5000, and the winner would have taken the seat with presumably about 50% of that, more or less, which is 2500.  2500 is only 14% of the total number of people who voted in the November general election, and not much better than having no runoff at all.  As it is, Cohen took the seat with 4321 votes in the end, which is almost double that which the winner would have received during a December runoff.  

Read more at SFGate and BeyondChron