Tournament complete!
Thanks for participating. Read about the results, or continue betting on Manifold.
Over the last few election cycles, polls have tended to show a bias towards Democratic candidates in both presidential and Senate elections in Pennsylvania and parts of the Midwest, namely Iowa, Michigan, Ohio, and Wisconsin.
This market will settle as YES if the 538 popular vote forecast has on average at least a three point bias towards Democrats in the 2022 Senate races in Ohio, Pennsylvania, and Wisconsin.
The bias will be calculated in each state based on the difference between the two candidates in the 538 popular vote forecast on Election Day compared to the actual results. Popular vote forecasts on Election Day will be taken from the pages below. The relevant numbers on each page come from the chart that says "popular vote."
https://projects.fivethirtyeight.com/2022-election-forecast/senate/ohio/
https://projects.fivethirtyeight.com/2022-election-forecast/senate/pennsylvania/
https://projects.fivethirtyeight.com/2022-election-forecast/senate/wisconsin/
💬 Proven correct
Fox Resolves No on All 3:https://www.foxnews.com/elections
CNN Resolves NO on all 3 : https://www.cnn.com/election/2022/results/senate?iid=politics_election_hp_ticker&admin1=39&election-data-id=2022-SG&selected-election-data-id=2022-SG-OH&election-painting-mode=projection&filter-key-races=false&filter-flipped=false
ABC news resolves NO on all 3: https://abcnews.go.com/Elections/2022-us-senate-election-results-live-map
At this point it is impossible for any of the 3 to resolve Yes on any of the sites I had access to
ussgordoncaptain made S$116!
@SalemCenter All major Televisions stations resolve no, as does the NYT. IDK what's left to count
All 3 states have >99% of votes counted which means that it is mathematically impossible for there to be an error
Fox Resolves No on All 3:https://www.foxnews.com/elections
CNN Resolves NO on all 3 : https://www.cnn.com/election/2022/results/senate?iid=politics_election_hp_ticker&admin1=39&election-data-id=2022-SG&selected-election-data-id=2022-SG-OH&election-painting-mode=projection&filter-key-races=false&filter-flipped=false
ABC news resolves NO on all 3: https://abcnews.go.com/Elections/2022-us-senate-election-results-live-map
At this point it is impossible for any of the 3 to resolve Yes on any of the sites I had access to
@SalemCenter The New York times now reports 94% of ballots counted in Pennsylvania and >95% of ballots counted in the other two states. Here’s the numbers for Republicans:
State, 538 Forecast, Result from NYT
Ohio, 53.2%, 53.3%
Pennsylvania, 49.5%, 47.2%
Wisconsin, 52.4%, 50.5%
So that would give us an average bias of ((53.2% - 53.3%) + (49.5% - 47.2%) + (52.4% - 50.5%))/3 = 1.36% in favor of Republicans, less than 3% against. And if my math is correct, it’s now impossible for that to change even if it’s just Republican ballots from now on.
@1941159478 I think you may have not chosen the most appropriate way to represent the situation: the market is for at least 3% bias towards *Democrats*, but the results so far indicate a -1.32 bias (i.e. actually a bias towards *Republicans*).
Dollars on sale for 70 cents apiece here.
I put most of my S$ into this market weeks ago, when it was trading closer to 50-50. IMO it was always way more likely to resolve NO (because YES was a bet against 538 being good at modeling...), but enough results are in that this is now free money for anyone with spare S$ not locked up in other markets.
@Max yeah the issue was that while it will be resolving NO, the market about red wave will also resolve NO and that had a lot better payout
@Max ran out of capital to throw against it we'll see
silly market cause 538 attempts to remove bias in weighting , so its misleading and pointless. I
Polling bias was nil for the 2018 midterms (when Trump was not on the ballot, as in 2022): https://fivethirtyeight.com/features/the-state-of-the-polls-2019/
@TimH
If we look at the high profile races in the midwest in 2018
Indiana senate: Final RCP average - Dem 1.3 Result Rep 5.9
Ohio senate: No RCP average, but if I use the last 3 polls - i.e. October or later, no others close to ED - Dem 11 Result Dem 6.4
Ohio gov: Again no RCP average, but ave of last 3 polls - Dem 4.7 Result Rep 4.3
Wisconsin gov: Last 3 polls - Dem 2.3, Result Dem 1.3
Wisconsin senate: RCP average - Dem 10.6, Result Dem 10.8
Michigan senate: RCP average - Dem 8.3, Result Dem 6.4
Michigan gov: RCP average - Dem 10.0, Result Dem 8.5
Pennsylvania gov: RCP average Dem 19.6, Result Dem 17.1
Pennsylvania senate: RCP average Dem 14.3, Result Dem 12.9
And if Missouri counts, Missouri senate: RCP average Rep 0.6, Result Rep 6.0
So, 9 out of 10 went more Republican than the RCP polling average (if available) or the last 3 polls listed on RCP's website. The average error looks to be about 3.4 points towards the Dems. (And perhaps cherrypicking, but if you look at the ones expected to be close - Ohio gov, Indiana senate, Missouri senate, Wisconsin gov - those misses were larger than average).
@MatthewLilley Taking your numbers as given, that's more informative than the national averages I pointed towards. I guess that means the rest of the country had a polling bias in the opposite direction, but that's less relevant for this market.
Mis-worded, to say the least; 538 forecasts are not polls - they are the output of a statistical model which includes polls (among other things)
Play-money betting
Mana (S$) is the play-money used by our platform to keep track of your bets. It's completely free for you and your friends to get started!