I would warn that RCP's aggregate 'betting odds' is in no way reliable or robust and offers nothing authoritative.
Just for starters, polling not only has ground rules, but statistics are inviolate when properly performed, and that includes self-correcting mechanisms.
Betting markets are non-uniform and a mixed of fixed odds versus pari-mutuel, lumping all those together does not create an appetizing "apples and oranges" fruit compote.
Comparing betting markets to correct polling data is like comparing analog to digital -- a stopped stem-wound Rolex Oyster is still only right twice a day.
Here you have every market but one -- PredictIt -- including the Mooch as a wagering entry! This creates garbage that pollutes any possible 'data' that could be gleaned from this nonsense.
Meanwhile, let's take the one bet market that didn't include the girlrilla -- PredictIt. It's not as if RCP actually drilled down to find the actual head-to-head matchup or make any effort to reduce the intangibles...
Furthermore, as indicated by the green arrow, this was actually Trump 40c - Biden 38c just this morning, and volatility in small populations samples is not your trend friend.
If that isn't enough, PredictIt has this odd marketing ploy of adverting their wagers as "shares" and advising bettors to hold onto their "shares" and "trade" them as if on an exchange, whereas other bet markets subverted into RCP's smoldering heap of aggregation are simply either head-to-head props where your smart bet on Trump is matched against a Leftard for Biden, or all the way down to fixed odds determined by some bloke in a surburban London shithole.