Arizona Harris +2, (7/24-28 Bloomberg/Morning Consult)
Georgia tie (ditto)
Michigan Harris +11 (ditto)
Nevada Harris Harris +2 (ditto)
North Carolina Trump +2 (ditto)
Pennsylvania Trump +4 (dits)
Wisconsin Harris +2 (ditts)
Although I have no specific reason to doubt Bloomberg, and these are the most recent soundings on unstable terrain, I am nonetheless skeptical and need replication to be convinced. Michigan particularly is an extreme outlier. A Fox poll 7/22-24 found the race tied. Either Fox or Bloomberg is very wrong. It is not near plausible that there has been a shift of 11 points in six days of polling with an overlap on 7/24.
The change in Arizona is almost as great and represents a flip, from Trump +5 (7/22-23 The Hill/Emerson) to Harris +2 four days later.
Those four state polls were conducted post-Biden withdrawal.
In Nevada the last state poll before Bloomberg was conducted 7/15-16 (post-Trump ear piercing but pre-Biden withdrawal) and showed Trump +10. It's a tremendous shift, and a flip, but more understandable than Michigan and AZ.
NC: The last poll (pre-Bloomberg) was in February, so who knows.
Pennsylvania appears to be a wrong the other way. Fox has good polling and they showed the race a dead heat 7/22-24. Trump +4 7/24-28? What, did he have a good half week that we all missed. I would believe Fox.
Georgia (Trump +2, The Hill; tie, Bloomberg) is plausible.
Wisconsin, from Trump +1 (Fox) to Harris +2 (Bloomberg) is plausible.
