I'm assuming 1/10 probability of Trump winning. Most poll aggregators are going 1/6; I actually suspect we'll see a bigger Clinton blowout than expected, but that's just MHO.
The Princeton Election Consortium currently has Clinton's win probability as: random drift 95%, Bayesian 98%. [1]
From the numbers so far coming out from early voting you should add +2 to the poll margin. The Clinton ground game is looking like being more formidable than the famed Obama one. And with RNC bailing on Trump and moving money to down ballot races expect to see there being a massive discrepancy in late minute advertising and GOTV operations.