US Election post:
Win probability today: Biden 83%; Trump 17%
Nat'l vote projection today: B 54%; T 45% (B+9)

Win prob 2016 (Polls+, 18 days before): Clinton 84%; T 16%
Nat'l votes t-18 2016: C 49%; T 44% (C+5)

All nos. from @NateSilver538/ @538politics

Typo yesterday - the win probabilities should have been Biden 87%; Trump 13%

Today:
Win prob: Biden 87%; Trump 12% (tie 1%)
Nat'l vote projection: B 54%; T 45% (B+9)
Nat'l poll: B 52%; T 42% (B+10)

Win prob 2016 (Polls+, t-17): Clinton 83%; T 17%
Nat'l vote proj 2016 (t-17): C 49%; T 44% (C+5)
Nat'l poll (t-17): C 46%; T 39% (C+7)

Today (18/10/20):
Win prob: Biden 87%; Trump 12% (tie 1%)
Nat'l vote projection: B 54%; T 45% (B+9)
Nat'l poll: B 52%; T 42% (B+10)

Win prob 2016 (Polls+, 23/10, t-16): Clinton 84%; T 16%
Nat'l vote proj 2016 (t-16): C 49%; T 44% (C+5)
Nat'l poll (t-16): C 46%; T 39% (C+7)

19/10/20:
Win prob: Biden 88%; Trump 12%
Nat'l vote projection: B 54%; T 45% (B +8: rounding)
Polling Average: B 53%; T 42% (B+11)

Win prob 2016 (Polls +, 24/10, t-15): Clinton 84%; T 16%
Nat'l vote proj: C 49%; T 44% (C+5)
Polling Average: C 46%; T 39% (C +6: rounding)

One of the reasons for this thread is to show how 2016 is not 2020. The @538politics 2020 model has far more uncertainty, increasing Trump's chances to turn things around. And in 2016, it underestimated how many undecideds ended up voting for Trump. A much smaller factor in 2020.