As of early afternoon of The Day After Election, Obama has 63.4 million votes (52.4%) to McCain's 56 million (46.3%). As expected, Obama has exceeded Bush's 2004 record of 62 million votes. McCain got more votes than Reagan in 1984, but needs 3 million more to surpass John Kerry's 2004 count.
I am not sure what is going on with the turnout. AP is reporting estimates ranging from 134 to 137 million, but so far only 121 million votes have been counted. Is it possible that 10% of the votes haven't been counted yet? All states except Washington, Oregon and Maine have reported more than 90% of precincts, and the only others below 98% are California, Colorado, Arkansas, South Carolina, and New Hampshire. If the final vote tally increases by 10% in proportion to the votes already counted, Obama could end up with close to 70 million, and McCain could equal Bush's 2004 numbers. Alternatively, there could be a lot of invalid ballots and the estimates could fail to take that into account.
If the current popular vote percentages hold, it will mean that the pollsters were very accurate. The pollster.com average was 52.0-44.4; RealClearPolitics average was 52.1-44.5; and FiveThirtyEight.com projections were within 0.2% of the actual results. The best individual pollster was Rasmussen, whose latest poll was 52-46 for Obama, which means no error other than rounding.
Most state polls were excellent, too, at least for us who relied on FiveThirtyEight's aggregation methods. The three states that were supposed to be extremely close - MO, IN, NC - are still uncertain. Polls also did very well in most states that were expected to be close (OH, FL, VA, CO, MT, GA) or that McCain hoped would be close (NH). Obama outperformed the polls in Pennsylvania (by 4%), Nevada (by 5%) and New Mexico (by 6%), while McCain did better in Iowa (by 4%), Arizona (by 4%) and North Dakota (by 7%). It also seems that, in the states won by large margins, the winner typically outperformed the polls, but of course there were fewer polls in those states and they may not have been done as carefully.
As many have pointed out, Obama's share of the popular vote is greater than any Democrat's since 1964 and greater than any non-incumbent's since 1952 (if we count the sitting Vice President, i.e., Bush in 1988, as an incumbent). More impressive fact, in my opinion, is that, since Andrew Jackson, only two Democrats - FDR and LBJ - have won a greater share of the popular vote.
(Total nerds will also find noteworthy that in the last 10 presidential elections (since 1968), only Reagan in 1984 and Bush in 1988 got a higher percentage of the vote. Obama's popular vote margin is greater than Clinton's in 1992 and, if he ends up winning all three uncertain state, his electoral vote advantage will be greater as well.)
Obama outperformed Kerry in 44 states and DC. The only states where he did worse were Arkansas, Louisiana, Oklahoma, West Virginia, Tennessee, and (probably) Alaska - and all but the first two were within 1 percentage point of Kerry's result. By contrast, he outperformed Kerry by 39 percentage points in Hawaii, by 22 in Indiana, 19 in North Dakota, 18 in Montana, 17 in Nebraska, 16 in New Mexico, Utah and Vermont, and 15 in Delaware and Nevada.
Obama won the majority of states, only the third Democrat since Truman to do so. (The other two were LBJ and Clinton.)
Obama did better in Delaware, DC, Illinois and Vermont than any Democrat since 1856 (when Democrats and Republicans emerged as the two major parties). His margin of victory in Delaware was second-highest for either party since 1856. (Hoover in 1928 won by 30%.)
It was the best Democratic result in California since 1936 and the second best ever (unless the final results lower the margin from the current 24% to below 21%, which was FDR's 1932 margin).
No Democrat has done better since 1964 in CO, CT, HI, IN, MD, MI, NE, NV, NH, NM, PA, VA, WI, and probably OR (where Obama seems almost certain to exceed Clinton's 10% margin from 1992). That is 14 states, in addition to the 4 (+DC) already mentioned.
It was also the second-best Democratic result "ever" (meaning since the later of statehood or 1856) in CT, HI and NH, third-best "ever" in ME, MI and NY, fourth-best "ever" in NM, PA and WI, and fifth-best "ever" in MD, NJ and OH.
It was the second-best result for a Democrat since FDR's time in CO, IN, MD, MI, NV, NM, PA, WA and WI.
In Nebraska, it was the second-best Democratic result since 1948 and third-best since 1940.
In Florida, it was the third-best Democratic result since 1948 (and better than in 1964).
In Utah, although Clinton lost by smaller margins, the Democratic share of the vote was the highest since 1968; in Texas, it just about equals Clinton's in 1996 (43.83%), which was the highest since 1976. And the result in North Carolina is the best for a Democrat since 1976.
On the other hand, Arkansas is the third-worst ever for a Democrat, Oklahoma is looking about the same as 2004, which was the third-worst ever (in both cases, only 1972 and 1984 were worse), Louisiana is the fourth-worst ever (after 1968, 1972 and 1984), Tennessee is the second-worst since 1868 (only 1972 was worse), and Kentucky and Wyoming were the worst of any year in which the Democrat won.
As the Republican streaks ended for Virginia and (probably) Indiana, the most reliable Republican states remain AK, ID, KS, NE, ND, OK, SD, UT and WY. The nine (the Nazgul?) have voted Republican in every election since 1968, 11 times in a row. By contrast, the most reliably Democratic state has been Minnesota (9 elections, since 1976) and the most reliable jurisdiction DC, which has voted Democratic in every election since the 23rd Amendment (12 elections, since 1964).