Rasmussen has to consider his own business, as well. He’s been producing the most Republican numbers of national polls. Most other polls are using D+6 or greater. So, with a D+3, Rasmussen will be closest as long as turnout is more Republican than D+4.5... which would be the second biggest D wave since Reagan.
Second, we have yet to hold an election where Ras showed more Rs than Ds the month before the election. November 2010 was the very first time Ras ever showed Rs outnumbering Ds. So, he doesn’t really have any data to guide him on what will happen on Election Day under that situation. He DOES know, though, that since Reagan got elected, turnout has been between even and D+4 every single election except 2008, with an average of D+2.5. That gives him a nice, safe model that is easy to defend. I don’t blame Ras one bit for that.
I think Ras is playing it smart. No fault can be found with that. He’ll still end up closer than anyone else, IMO.