So, how'd I do? Badly! But, being wrong is far more interesting than being right. Let's start with a strict count of yes/no, whether I picked winners.
W:4 L:3 I'd be better off asking a coin, it seems! Within my MOE, Y:6 N:2, but most of that comes from my lack of confidence in a few predictions that allowed for things to just squeak in. Using the Wilcoxon Signed Rank test, I get a W score of W=4 and W=18, neither is statistically significant. Alas, I can't even pick lose consistently! :}
I'm in good company on a lot of things, though. The pollsters, as I figured, butchered the numbers on 4 and 2 - Robopolling's a dumb idea for rural issues. For the pollsters, it doesn't matter that they got the net result of 4 right - it was outside their confidence interval, which to a pollster is the same as getting it wrong.
Begich - I knew he was big, but I should have gone bigger.
Stevens - I anticipated even fewer people would vote `3rd candidate,` AKA, Mr. Viky vik vickster from Vikanistan.
Berkowitz - Benson put in about as good of a showing as I estimated. Is this because even a broken clock is right twice a day?
Prop 1: I generalized from Fairbanks too much. Repeat to self: Fairbanks is more libertarian leaning! I was right, but not how I'd hoped I'd be right (e.g., only within my MOE)
Prop 2: I was within my MOE, but I a) took too much from previous electoral results and b) over estimated Anchorage's influence.
Prop 3: Very wrong, here. I'm a tad surprised. Generally, people like spending other people's money.
Prop 4: I think I had this one nailed, within my `stated confidence.` We'll see if it played out like I thought it did, when we look at how various regions broke down. I hope the state will be releasing that data (Edited to add: They did. Woo!).
Woo! And that's that. I'm a gambling man, and I'm willing to bet a whole nickel I do better in the general election. There's more data, and it'll be better vetted - always is - which gives me more to play with than my `gut feeling` how various electorates will break down. :)