Like most other mainstream polls Rasmussen usually errors to the negative on GOP and to the positive on Dems, meaning the approximate 45% they have Trump at is probably well over 50%.
I looked to see if this was true. So I looked at the final Rasmussen poll compared to the actual results for every race they submitted a poll in the last 2 weeks of the campaign since 2010. (All from RCP)
Errors in favor of GOP: 48
2010: (GCB, California-Sen, Nevada-Sen, West Virginia-Sen, Washington-Sen, Pennsylvania-Sen, Illinois-Sen, Colorado-Sen, Connecticut-Sen, Wisconsin-Sen, Kentucky-Sen, Alaska-Sen, California-Gov, New Hampshire-Gov, Colorado-Gov, Connecticut-Gov, Florida-Gov, Hawaii-Gov, Illinois-Gov, Massachusetts-Gov, Ohio-Gov, Oregon-Gov, Arizona-Gov, Georgia-Gov, New Mexico-Gov, South Carolina-Gov, Wisconsin-Gov.)
2012: (National-Pres, Nevada-Pres, Colorado-Pres, Iowa-Pres, Pennsylvania-Pres, New Hampshire-Pres, Michigan-Pres, Wisconsin-Pres, Ohio-Pres, Virginia-Pres, Minnesota-Pres, Wisconsin-Sen, Virginia-Sen, North Dakota-Sen, Nevada-Sen, Montana-Sen, Indiana-Sen, Washington-Gov, New Hampshire-Gov)
2016: (National-Pres)
2018: (GCB)
Errors in favor of Dems: 6
2010: (Missouri-Sen, Minnesota-Gov, Texas-Gov)
2012: (Arizona-Pres)
2013: (Virginia-Gov)
2014: (GCB)
If I rounded the polls off then there would be a couple ties such as 2016's national poll and 2012's Arizona Presidential poll.