Predict It currently has Republican chances of winning the Senate up to 52 cents

[Polybius]

WTF is up with that? After all that's come out against them lately (especially with Walker in GA)?

https://www.predictit.org/markets/search?query=Midterms2022

[NewHendoLib]

1. 66 to 34 chance Dems at 538

[Polybius]

3. Better, but down slightly

It was 71%.

[regnaD kciN]

2. I suspect the phrase you're looking for is "gas prices"...

[Mr.Bill]

4. I'd like to ask a republican how them controlling the Senate

will bring gas prices down.

[TheFarseer]

7. Because

Republicans will finally let us drill -even though Biden has done practically nothing to stop drilling in any way and the US is the biggest energy producer in the world. So I have no idea what is meant by finally let us drill.

[ColinC]

5. Polls.

Which are going to be proven historically inaccurate after election day. It might be a good idea to start making ridiculously large bets in favor of Dems.

[Elessar Zappa]

6. I don't pay attention to polls, good or bad.

They’ve been entirely inaccurate these past six years. They didn’t predict Trump winning, they didn’t predict the Kansas abortion result nor Palin losing in Alaska. I suspect they’ll be off this time in our favor but I’m not sure.

[orleans]

8. i don't answer/respond to polls. so... my voting intention hasn't been included

[Genki Hikari]

9. You don't have to be polled

For them to be accurate.

Really.

The numbers surveyed matters only to the margin of error. Once they hit 1500-2000, the results will be a good snapshot for the entire country. 1500-2000 won't usually be that far off from what hundreds of millions will do.

Really.

From there, the way you figure out if the poll is any good is to check he composition of the polled sample (how many of X group versus Y group vs Z group), and then the questions that were asked, to check for bias. Even the best sampled poll can be screwy if the demographic splits and questions are biased.

For instance, running a poll that asks, "What did you do this weekend?" will yield very different answers from one that asks, "Did you go to church this weekend?" The latter is a loaded question which implies a values/morality judgment, while the former is fact-finding, not values-loaded.

So a poll with a proper sample that asks unbiased questions will come out accurate if only 2000 people are surveyed. You're not that different from a sizeable percentage of that sample of 2000.

[Deminpenn]

10. It's the turnout models, i.e. "likely voters"

that have been and are problematic. It appears that a number of pollsters got freaked out by missing Trump's win 2016 because a lot of "unlikely" voters showed up at the polls and have adjusted their turnout models/weighting to account for them showing up in droves. The polling "misses" on the KS abortion vote, the NY special election and others where Dems have "overperformed" indicate that the turnout/LV models pollsters are using might not be right.

Trafalger's model has always been weighted to capture "shy" voters, i.e. rw/conservatives. It was right in 2016 and 2020 only because of Trump himself being on the ballot and his cultlike hold on his voters, but that doesn't mean the clock wasn't broken.

[Amishman]

11. PredictIt seems slanted right significantly right now

Just stuck another $100 in my PI account this morning.

70 more in Fetterman (zero Oz enthusiasm and I'm in the red aprts of PA Oz needs to win.)

30 on Val in FL. PI has Rubio No at 8 cents. Given Rubio's hard anti abortion stance and the relatively favorable generic ballot polling recently with Hispanic voters, she's and underdog but definitely has better chances than that. Longshot, but I like the wager price vs actual chances.

[PCIntern]

12. Betting on propositions is not like a poll...

A bookie wants half the money on one side and half the money on the other for the best return to them. They read public perception and undoubtedly the oddsmakers feel that there will be more people laying down money on one side or the other given that proposition and those odds. The worst thing that can happen to a bookie is that all the money comes down on one side and they have to pay out prodigiously on one side of the equation.