Half-wrong, half-right, or just half-wit?

I made 10 predictions at the beginning of the year:

1. The Democrats will lose 7 or 8 seats in the Senate in the November 2010 election. FAIL
2. There will be a kerfuffle when images supposedly from airport backskatter X-ray machines are leaked to the Internet. Score
3. A Schleck will win the Tour De France. FAIL
4. The president will sign a Cash for Caulkers bill. FAIL
5. More than 100 banks will fail in the US. Score
6. General Motors will sell fewer cars than they sold in 2009. FAIL
7. Ford and Toyota will sell more cars than they sold in 2009. Score
8. We'll have a white Christmas. FAIL
9. Spring Town Meeting will last 5 or fewer nights. Score
10. Amherst will pass an override. Score

I said I thought each of them had a greater than 50% chance of happening.  I got 50% correct.  Oops.

What lessons I should draw from my failures, besides the general fact that I'm overconfident?

I'm a programmer, so I should expect to make off-by-one errors.  The Democrats only lost 6 seats. 

I should avoid making predictions about sports.  So even though I think the Patriots are going to win the Super Bowl, I shouldn't predict that they're going to win.  But they are.  But I'm not predicting that.  Just so you know.

I don't know lesson to learn from my 'cash for caulkers' FAIL -- maybe that predicting what legislation will catch the fancy of our congress-critters is hopeless?  Two weeks ago I would have predicted that there was almost no chance Don't Ask Don't Tell would be repealed.

I should trust large organizations more; GM and Chrysler were not horribly mismanaged after being bought by Uncle Sam.

And I've lived here long enough to know that trying to predict New England weather is just plain dumb.

But I've learned my lesson, and I'm 100% positive that my 2011 predictions will be brilliant, insightful, and accurate-- guaranteed, or double your money back!