Bob Somersby of Daily Howler noted that the media was already predicting a year ago (in February) that Obama would quickly catch Hillary when he announced.

There was no inevitability. Hillary had name recognition but also tons of bad press, so much of her candidacy has been trying to resurrect her image nationwide, to get past the dire predictions that she would be so divisive and polarizing that she would be a poster child for Republican fund-raising.

Of course one of the absurdities of all this was the polls comparing different candidates of each party to see who would win which matchup. Why absurdity? Well consider McCain. By conventional wisdom he was one of the strongest candidates, but didn't properly abase himself with the party "base" (code word for "bigots and flat earthers"). By December McCain was running about 4th place in the pack and was quite broke. Who would have guessed that when it actually came to voters voting, he would quickly rise to the top and never yield?

And then there's Giuliani, who came close to being a Republican "inevitable" until he actually started opening his mouth and the people of New York started remembering a few short years ago. So he began the slow descent into the depths of pollster hell, even challenging Ron Paul for the cellar.

So what do all those Clinton-McCain, Clinton-Giuliani, Obama-McCain, Obama-Giuliani polls mean? Pretty much nothing. McCain January would beat McCain December by about 30 points. The idea that Giuliani by January could beat anybody seemed delusional. But even these evaluations are simplistic. The public gets to know candidates fairly late, doesn't bother to make up its mind until shortly before the election, and then votes. Meanwhile there are news cycles, the horse-race effect building up from each successive contest, TV ads and canvassing, and of course lots more chatter in the blogosphere, plus an occasional YouTube or other video. Where candidates are relatively acceptable, we've proven there can be rapid shift of opinions just before the vote. But mostly people just don't pay attention until late, unlike pundits and wonks, and there's not much reason to pay attention until the candidates really get going.

It should also be obvious that until the voting starts, no one really knows how their "strengths" and "weaknesses" will work out. This can be demographics, specific issues or qualities, etc. One badly spoken phrase can break a campaign or lose credibility on an issue. It can often be hard to scramble back into position because like with McCain, once you start falling back you're frequently dismissed - and all eyes turn to the new "inevitable" frontrunner. Though "frontrunner" may not always be an advantage, as Gore struggled with. In some types of bike racing and running, it's better to tail the frontrunner by a small bit and give a last spurt to get ahead at the end when it matters. And in some sports such as car racing it's very difficult to switch positions at full speed rounding a turn. So candidates do find themselves "desperate" to change perception, and it often looks like two fighters clenched, not full out whacking each other, but doing little jabs and headbutts and side-steps and other moves to try to find an opening, get a movement, make enough space to throw in a flurry of punches. And sometimes it leads to a surprise KO, like with McCain, sometimes it leads to a slow comeback, and sometimes it's a lost cause. But you really never know until the final decision. At one point Mohammed Ali seemed unbeatable, and then the thoroughly whipped Frazier came out of nowhere and floored Ali. Ali beat Frazier, but Frazier won the bout, and spent a few days in the hospital cherishing his victory.

Every so often someone tells me what the weatherman predicted for the day, and I flip a coin and am usually as accurate. Same goes for politics, sometimes one of the least predictable sports.