Based on the "Last Updated" time on CNN's election results site, my best estimate for when Missouri was called by CNN is 21:09 UTC yesterday. So consider this to me an update covering the time immediately prior to 21:15 UTC yesterday.
Missouri is called for McCain. This is the last outstanding state.
This makes the results (assuming no faithless electors) Obama 365, McCain 173... which has been what it would look like it would be for quite a bit now, but now it is official. (Well, not really, not until the electoral college votes... but all states have been called by CNN now.)
So, how did I do?
Well, every single state that I did not say was a too close to call swing state, I got right. But lets be a little more critical than that. If you look at my last regular daily update on November 3rd, the "Current Status" where everybody got their leans had Obama 338, McCain 200. What did I miss? I thought McCain would win North Carolina, Indiana and all of Nebraska. Obama of course ended up winning North Carolina, Indiana and Nebraska's 2nd District.
Of these, I had considered both North Carolina and Indiana as too close to call and that they could easily go either way. So not too upset about that. On Nebraska I had decided early on not to look at the possibility of Nebraska splitting unless the state itself looked somewhat close. That seems like an error now. However, I'm not sure what else I would have done, as I did not see regular polling on the separate Nebraska districts anyway.
BUT... there was one more error.... after my last daily update on the 3rd, I continued to log polls as they came in on election day itself, before the real polls started closing. I logged one change early in the day that flipped Missouri to just barely on the Obama side of the line instead of just barely on the McCain side of the line.
So my actual final "everybody gets their leans" prediction was Obama 349, McCain 189 which was actually closer to the actual final result than my final prediction on the 3rd, but only because a couple of the states I got wrong canceled each other out. (Missouri and Indiana both have 11 electoral votes, so when I got them both wrong the total remained the same.)
So, in the end, out of 50 states and DC, I ended up predicting correctly on all except North Carolina, Indiana, Missouri and the 2nd district of Nebraska. And my electoral college total was only off by 16. And with the exception of that pesky single electoral vote from Nebraska, all of the states which I "missed" were states that I had actually described as being too close to call.
All and all, I think those are pretty good results. So I am happy with that.
I'll do one more update when I combine the wiki pages for the historical predictions through the race with the pages reflecting actual results and get everything all cleaned up to sit there forever for random people googling about the 2008 race. And then I'll finally be done with this! :-)
Edit 21 Nov 04:06 UTC: Fixed Typo in Obama's actual electoral college vote total (365 instead of 265). Thanks to the reader who pointed this out.