Nate Silver (and others) Dance with the Green Knight
Nate has been dismissive of the election markets this cycle. He’s noted his own model is much more confident in Biden. Fair enough. Let’s let them battle. However, this isn’t simple Full Kelly wagering. That’s easy mode. Nate’s claim is quite bold, so let’s give him a chance to prove it.

Nate shall get a half-Kelly bet into the markets and the markets shall in turn get a half-Kelly wager into Nate’s numbers. Each side shall have identical vig on each wager. All of this was recorded on the final Saturday before the election. Let’s start with the main event:
| Biden-President | 133.44% | 29.62% | 0.00% |
To explain how this works, let me break that down into the two individual bets:
| Nate into the market | Biden-President | 34.70% | 16.84% | 0.00% |
| The market into Nate | Trump-President | 12.77% | 98.74% | 0.00% |
The market was offering Biden -206, Trump +181. Nate was offering Biden -1070, Trump +773. Identical vigs, two half-Kelly wagers.
That’s a lot of confidence in the Biden. Let’s break it down state-by-state:
| Biden-AK | 2.56% | 16.96% | 0.00% |
| Biden-AZ | 29.49% | 16.96% | 0.00% |
| Biden-FL | 49.21% | 42.10% | 0.00% |
| Biden-GA | 29.71% | 29.50% | 0.00% |
| Biden-IA | 9.94% | 14.03% | 0.00% |
| Biden-ME | 3.17% | 0.39% | 0.00% |
| Biden-MI | 205.84% | 23.45% | 0.00% |
| Biden-MN | 200.16% | 20.03% | 0.00% |
| Biden-MT | 7.66% | 65.51% | 0.00% |
| Biden-NC | 31.58% | 20.90% | 0.00% |
| Biden-NH | 68.99% | 12.32% | 0.00% |
| Biden-NM | 128.52% | 10.28% | 0.00% |
| Biden-NV | 83.15% | 15.63% | 0.00% |
| Biden-OH | 20.71% | 34.64% | 0.00% |
| Biden-PA | 107.33% | 31.84% | 0.00% |
| Biden-TX | 15.14% | 33.84% | 0.00% |
| Biden-WI | 168.20% | 22.70% | 0.00% |
Michigan and Minnesota are of special interest here as Nate has gone HAM (as have the markets, betting into Nate). Florida is worth noting as well given how strongly Nate has Biden favored. Wisconsin, Pennsylvania, and Georgia are all key states where Nate has shown significant confidence in Biden.
There will be a lot of eyes on Senate control as well:
| Dem-Senate Control | 54.15% | 22.89% | 0.00% |
There are a number of competitive races to watch:
| AK-Dem Senate | 5.34% | 20.59% | 0.00% |
| AL-Dem Senate | 6.03% | 36.75% | 0.00% |
| AZ-Dem Senate | 4.84% | 1.21% | 0.00% |
| CO-GOP Senate | 0.93% | 5.79% | 0.00% |
| GA-Dem Special | 40.85% | 35.97% | 0.00% |
| IA-Dem Senate | 3.86% | 3.58% | 0.00% |
| KS-Dem Senate | 2.60% | 9.54% | 0.00% |
| ME-GOP Senate | 8.76% | 15.45% | 0.00% |
| MI-Dem Senate | 42.78% | 13.43% | 0.00% |
| MN-Dem Senate | 82.02% | 11.01% | 0.00% |
| MS-Dem Senate | 2.49% | 17.67% | 0.00% |
| MT-Dem Senate | 1.73% | 3.44% | 0.00% |
| NC-Dem Senate | 9.30% | 5.50% | 0.00% |
| TX-GOP Senate | 0.44% | 0.05% | 0.00% |
Michigan and Minnesota are fairly strong opinions, but I’m most fascinated by the GA Dem Special. Nate is predicting a Dem pickup there. The markets are skeptical.
As for the House I must confess there is significant less liquidity in these markets (save for House control), but for science I’ve recorded them:
| Dem-House Control | 141.97% | 9.63% | 0.00% |
That seems like an easy pickup. As for the individual races:
| Dem-CA 26 | 32.67% | 10.92% | 0.00% |
| Dem-CO 3 | 3.07% | 4.85% | 0.00% |
| Dem-IA 1 | 26.19% | 4.91% | 0.00% |
| Dem-IA 2 | 55.86% | 11.67% | 0.00% |
| Dem-IA 3 | 35.03% | 8.81% | 0.00% |
| Dem-IL 13 | 3.35% | 5.88% | 0.00% |
| Dem-MI 3 | 21.68% | 36.12% | 0.00% |
| Dem-MI 6 | 1.65% | 4.27% | 0.00% |
| Dem-NC 11 | 5.91% | 19.44% | 0.00% |
| Dem-NC 8 | 2.33% | 4.50% | 0.00% |
| Dem-NJ 3 | 72.32% | 10.56% | 0.00% |
| Dem-NY 2 | 5.46% | 8.96% | 0.00% |
| Dem-OH 1 | 7.07% | 12.64% | 0.00% |
| Dem-PA 10 | 12.96% | 18.90% | 0.00% |
| Dem-VA 2 | 23.22% | 4.27% | 0.00% |
| Dem-VA 5 | 29.26% | 33.09% | 0.00% |
| GOP-AZ 6 | 7.45% | 6.09% | 0.00% |
| GOP-CA 21 | 8.83% | 13.88% | 0.00% |
| GOP-CA 25 | 4.74% | 7.33% | 0.00% |
| GOP-CA 48 | 1.02% | 2.62% | 0.00% |
| GOP-GA 7 | 33.02% | 44.37% | 0.00% |
| GOP-KS 2 | 11.58% | 1.62% | 0.00% |
| GOP-MO 2 | 18.75% | 12.46% | 0.00% |
| GOP-NJ 2 | 1.90% | 1.97% | 0.00% |
| GOP-NM 2 | 3.35% | 5.88% | 0.00% |
| GOP-NY 11 | 9.95% | 16.00% | 0.00% |
| GOP-NY 24 | 7.84% | 5.90% | 0.00% |
| GOP-PA 18 | 48.30% | 16.44% | 0.00% |
| GOP-TX 21 | 9.61% | 6.38% | 0.00% |
| GOP-TX 22 | 40.88% | 39.56% | 0.00% |
| GOP-TX 24 | 21.87% | 24.93% | 0.00% |
| GOP-UT 4 | 10.36% | 17.72% | 0.00% |
| GOP-WI 3 | 1.38% | 9.78% | 0.00% |
G. Elliot Morris of the Economist has been equally dismissive of markets, even while sparring with Nate over modeling choices. In fact, Morris has been much more confident in Biden than Nate for the bulk of this cycle:
| Biden-President | 306.25% | 36.44% | 0.00% |
This shows up on the state level as well:
| Trump-AK | 39.76% | 3.72% | 0.00% |
| Trump-MT | 101.88% | 4.09% | 0.00% |
| Biden-AZ | 61.40% | 28.25% | 0.00% |
| Biden-FL | 99.57% | 61.18% | 0.00% |
| Biden-GA | 27.31% | 27.69% | 0.00% |
| Biden-IA | 11.76% | 16.27% | 0.00% |
| Biden-ME | 138.75% | 9.51% | 0.00% |
| Biden-MI | 317.59% | 25.59% | 0.00% |
| Biden-MN | 858.99% | 24.26% | 0.00% |
| Biden-NC | 39.61% | 24.69% | 0.00% |
| Biden-NH | 438.80% | 22.12% | 0.00% |
| Biden-NM | 167.43% | 11.85% | 0.00% |
| Biden-NV | 176.92% | 21.20% | 0.00% |
| Biden-OH | 9.62% | 18.34% | 0.00% |
| Biden-PA | 260.58% | 41.33% | 0.00% |
| Biden-TX | 0.48% | 1.36% | 0.00% |
| Biden-WI | 302.27% | 25.95% | 0.00% |
Minnesota or bust, eh? Wisconsin, Pennsylvania, New Hampshire, and Michigan are all fairly huge as well. As with Nate, we see Florida as a fairly strong play.
There’s confidence in the Dems in the Senate as well.
| Senate-Dem Control | 90.59% | 30.60% | 0.00% |
We see that show up in the individual races:
| AK-Dem Senate | 10.66% | 36.86% | 0.00% |
| AL-GOP Senate | 63.37% | 4.59% | 0.00% |
| AZ-Dem Senate | 50.29% | 9.57% | 0.00% |
| CO-GOP Senate | 75.04% | 6.83% | 0.00% |
| GA-Dem Senate | 31.66% | 28.48% | 0.00% |
| GA-Dem Special | 35.45% | 32.59% | 0.00% |
| IA-Dem Senate | 10.43% | 9.10% | 0.00% |
| ME-Dem Senate | 7.40% | 3.20% | 0.00% |
| MI-Dem Senate | 188.80% | 28.55% | 0.00% |
| MN-Dem Senate | 139.53% | 14.74% | 0.00% |
| MT-Dem Senate | 1.73% | 3.44% | 0.00% |
| NC-Dem Senate | 35.44% | 17.24% | 0.00% |
| SC-Dem Senate | 16.22% | 40.07% | 0.00% |
Plenty of headliners here, but it’s Georgia that is holding my attention. Both races will be fascinating.
There’s a similar tale in the House:
| House-Dem Control | 146.47% | 10.19% | 0.00% |
As for the individual races:
| Dem-CA 21 | 11.53% | 5.27% | 0.00% |
| Dem-CA 39 | 8.71% | 2.25% | 0.00% |
| Dem-FL 26 | 198.22% | 30.68% | 0.00% |
| Dem-GA 7 | 20.57% | 7.57% | 0.00% |
| Dem-GA 7 | 124.86% | 16.33% | 0.00% |
| Dem-IA 2 | 111.28% | 17.29% | 0.00% |
| Dem-IL 13 | 18.91% | 27.46% | 0.00% |
| Dem-MI 3 | 13.01% | 23.97% | 0.00% |
| Dem-MI 6 | 20.11% | 39.58% | 0.00% |
| Dem-MT at Large | 11.02% | 24.92% | 0.00% |
| Dem-NC 8 | 8.70% | 15.21% | 0.00% |
| Dem-NJ 2 | 3.12% | 2.43% | 0.00% |
| Dem-NJ 3 | 72.32% | 10.56% | 0.00% |
| Dem-NJ 7 | 70.57% | 11.30% | 0.00% |
| Dem-NV 4 | 190.12% | 22.31% | 0.00% |
| Dem-NY 2 | 14.11% | 20.86% | 0.00% |
| Dem-NY 22 | 84.59% | 21.19% | 0.00% |
| Dem-NY 24 | 16.97% | 15.05% | 0.00% |
| Dem-OK 5 | 5.18% | 4.31% | 0.00% |
| Dem-OR 4 | 139.53% | 14.74% | 0.00% |
| Dem-PA 10 | 33.16% | 39.63% | 0.00% |
| Dem-SC 1 | 10.33% | 4.50% | 0.00% |
| Dem-TX 21 | 1.00% | 1.21% | 0.00% |
| Dem-TX 24 | 21.22% | 11.03% | 0.00% |
| Dem-TX 7 | 68.06% | 11.93% | 0.00% |
| Dem-UT 4 | 24.78% | 8.88% | 0.00% |
| Dem-VA 2 | 68.95% | 9.92% | 0.00% |
| GOP-CA 25 | 23.33% | 29.45% | 0.00% |
| GOP-CO 3 | 21.48% | 9.98% | 0.00% |
| GOP-FL 15 | 30.67% | 5.43% | 0.00% |
| GOP-KS 2 | 40.22% | 4.97% | 0.00% |
| GOP-MO 2 | 56.06% | 28.55% | 0.00% |
| GOP-NE 2 | 5.71% | 5.40% | 0.00% |
| GOP-NY 11 | 6.59% | 11.06% | 0.00% |
| GOP-PA 1 | 29.40% | 11.29% | 0.00% |
| GOP-TX 22 | 6.23% | 8.35% | 0.00% |
Finally, there’s the Epstein-Bitecofer model. You might remember them from the Biden 99.5% projection:
That was… something. They’ve come a bit closer to reality and were at 93.4% at the time of my tracking. That’s leads to a Presidential wager that almost seems reasonable in comparison:
| Biden-President | 201.43% | 33.52% | 0.00% |
Different model. Same confidence in Biden in Minnesota:
| Biden-AK | 24.85% | 105.92% |
| Biden-AZ | 98.43% | 36.15% |
| Biden-FL | 93.92% | 59.65% |
| Biden-GA | 24.98% | 25.85% |
| Biden-ME | 105.38% | 7.98% |
| Biden-MI | 66.48% | 14.53% |
| Biden-MN | 200.16% | 20.03% |
| Biden-NC | 80.72% | 38.19% |
| Biden-NH | 154.29% | 17.86% |
| Biden-NM | 167.43% | 11.85% |
| Biden-NV | 214.25% | 22.28% |
| Biden-OH | 2.65% | 5.58% |
| Biden-PA | 142.71% | 35.47% |
| Biden-TX | 40.92% | 70.06% |
| Biden-WI | 21.15% | 6.30% |
Pennsylvania is also worthy of note. As for the Senate:
| Senate-Dem Control | 65.65% | 25.77% | 0.00% |
There’s more excitement in some of the individual races:
| AK-Dem Senate | 14.86% | 47.89% | 0.00% |
| AL-GOP Senate | 27.12% | 1.90% | 0.00% |
| AZ-Dem Senate | 21.52% | 4.88% | 0.00% |
| CO-Dem Senate | 75.04% | 6.83% | 0.00% |
| GA-GOP Senate | 4.25% | 3.08% | 0.00% |
| IA-Dem Senate | 1.76% | 1.67% | 0.00% |
| ME-Dem Senate | 94.67% | 23.70% | 0.00% |
| MI-Dem Senate | 115.28% | 24.09% | 0.00% |
| MN-Dem Senate | 68.95% | 9.92% | 0.00% |
| MT-GOP Senate | 35.53% | 12.63% | 0.00% |
| NC-Dem Senate | 74.28% | 27.94% | 0.00% |
| SC-Dem Senate | 7.36% | 20.79% | 0.00% |
| TX-Dem Senate | 5.34% | 20.59% | 0.00% |
We’ll see how all of this goes. They have House ratings, but not quantifiable projections, so my job is done. Again, all of this was recorded on the final Saturday before the election.
Good luck to everyone involved and may all your votes count!
I’m having trouble understanding what your columns mean?
“The market into Nate Biden-President 12.77% 98.74% 0.00%”
12.77, 98.74 where are these coming from?
The first column seems like the half Kelly wager %, I get ~13%, but then I have no idea what
“Biden-President 133.44% 29.62% 0.00% ”
would mean?
The market is risking 12.77% of its bankroll to win 98.74% on Trump. It’s my bad for not correcting the second column.
Pingback: 2020 Election: Prediction Markets versus Polling/Modeling Assessment and Postmortem | Don't Worry About the Vase