I can't really speak to VA and CO, but I have talked to people in the know in both states. As is the case in Iowa, Romney is one the move and polls might not pick up on it right away.
I've been privy to a lot of polling data on a number of races in Iowa and the Romney movement is remarkable. If its that way in other states he's going to be the President next January.
If there is one problem with polls its that they are always late on identifying geographical movements. For example, the divide to watch in Iowa is the urban/rural split. Romney is going to crush Obama in rural Iowa, but a statewide poll might have difficulty picking it up because the sample needs to be geographically balanced, meaning more respondents are from urban areas, or its impossible to identify pockets of intense support.
In Iowa look at Santorum's caucus victory. Polls had him moving up, but they never captured his actual strength. Why? Because his support was almost entirely rural. Same was true in the 2010 Republican gubernatorial primary. One candidate had intense support in about a 5th of the counties. The Des Moines Register poll was off by 19 points. The guy still lost, but still.
I love polls. I commission polls. But I also know that they are always flawed in one way or another.