Turnout in Knox County was 92,541, a new record. Looking at the presidential primaries, there were 90,323 votes cast. The vote breaks down 42.6% Democrat (38,545) to 57.3% Republican (51,778), a strong showing for Knox County Democrats.

Previous predictions here were all over the place. First predictions from last November were:

• Greg Mackay: 70,000 (R: 40,000/57%, D: 30,000/43%)

• Bill Young: 95,000 (R: 65,000/68%, D: 30,000/32%)

• R. Neal: 95,000 (without Thompson, R: 53,000/56%, D: 42,000/44%)

• Don Daugherty: 50,000 total

Bill Young revised his prediction upward at one point to 33,000/38% for Democrats, and later revised GOP turnout down to 55,000/63%, for a total of 88,000.

I revised downward sharply based on the first few days of early voting to 47,503 total (lesson: early voting in primaries is not only unpredictable, it's probably not a good idea -- just ask Thompson and Edwards voters, plus be sure to include extra EV days in your calculations). After a few more days of early voting, I revised back upward to 61,000 with a possibility of 82,000. The Mrs. predicted 69,591.

So Bill Young and I were closest to the pin in November, with Bill Young having the closest overall projection as early voting progressed. (But he was off on the breakdown. Does this suggest big GOP crossover to the Dem primaries, and if so, are those reliable Dem votes in November?) Greg Mackay nailed the final percentages exactly with his first November projection, and I wasn't far off.

So, who's the winner, besides Knox County voters?

UPDATE: Turnout was approx. 41% based on registered voters as of Dec. 2007, as compared to 36% statewide.

UPDATE: For future reference, early voting turount of 36,719 was 39.6% (call it 40%) of the total.