More of Nate Silver's Predictions
I looked though the FiveThirtyEight website for predictions on all of the upcoming caucus and primary states and, except for Nevada and South Carolina which I have already posted here, this is what I found (all projections in terms of chances of winning in that state):
Virginia
http://projects.fivethirtyeight.com/election-2016/primary-forecast/virginia-democratic/
Clinton: 98% - Sanders: 2%
Oklahoma
http://projects.fivethirtyeight.com/election-2016/primary-forecast/oklahoma-democratic/
Clinton: 81% - Sanders: 19%
Arkansas
http://projects.fivethirtyeight.com/election-2016/primary-forecast/arkansas-democratic/
Clinton: Greater than 99% - Sanders: Less than 1%
Michigan
http://projects.fivethirtyeight.com/election-2016/primary-forecast/michigan-democratic/
Clinton: 95% - Sanders: 5%
Most of the other states pages showed "insufficient polling data"
NOTE: If you have questions about which polls were used for a particular prediction, go to the website using the link provided for that prediction Scroll down and you will see a list of polls. Keep in mind that not all of the polls may have been used. Look in the column titled WEIGHT. Silver weights polls in his calculations (gives more emphasis to some polls than others) based on the past accuracy of the polling organization and the bias of their polls.
For example, for the Virginia prediction, the "Public Policy Polling" poll (weight -1.18) received a lot more weight than the "Christopher Newport University" (wieght - 0.11)and the other polls below them weren't used at all in Silver's prediction calculations - notice that they all have 0.00 for the weight. That is almost certainly that is because the are too old to be valid.
= new reply since forum marked as read
But I am anxious to see how they react over the next few weeks. 
