How accurate has Gallup been predicting midterm elections in the past?

[Valienteman]

You be the judge, and comment please. These are the final pre-election Gallup generic polls among likely voters prior to every midterm (all decimals are rounded)

1950
Gallup predicts: Dems +2%
Actual result: Tie
Gallup error: 2 points

1954
Gallup predicts: Dems +6
Actual Result: Dems +4
Gallup error: 3 points

1958
Gallup predicts: Dems +14
Actual result: Dems +13
Gallup error: 1 point

1962
Gallup predicts: Dems +10
Actual result: Dems +6
Gallup error: 5 points

1966
Gallup predicts: Dems +5
Actual result: Dems +3
Gallup error: 2 points

1970
Gallup predicts: Dems +6
Actual result: Dems +8
Gallup error: 2 points

1974
Gallup predicts: Dems +20
Actual result: Dems +17
Gallup error: 3 points

1978
Gallup predicts: Dems +10
Actual result: Dems +8
Gallup error: 1 point

1982
Gallup predicts: Dems +10
Actual result: Dems +12
Gallup error: 2 points

1986 (no likely voter data)

1990
Gallup predicts: Dems +8
Actual result: Dems +8
Gallup error: 0 points

1994
Gallup predicts: Republicans +7
Actual result: Republicans +6
Gallup error: 1 point

1998
Gallup predicts: Dems +4
Actual result: tie
Gallup error: 4 points

2002
Gallup predicts: Republicans +6
Actual result: Republicans +5
Gallup error: 1 point

2006
Gallup predicts: Democrats +7
Actual result: Democrats +8
Gallup error: 1 point

link: http://www.gallup.com/poll/24493/election-polls-accuracy-record-midterm-congressional-elections.aspx

[Ozymanithrax]

1. So in most cases they got it right, but were off on the point spread. n/t

[MilesColtrane]

2. Don't most polls have the margin of error as +/-2 points?

If that's the case, they're batting .715.