The 2.2% is not a "flipping a coin" style probability. It is a "we lack complete knowledge" style probability. So the odds that it drops to 0 with better observ

The 2.2% is not a "flipping a coin" style probability. It is a "we lack complete knowledge" style probability. So the odds that it drops to 0 with better observations are 100%-2.2% = 97.8%, which is pretty likely. These odds are the odds systematically incorporating all knowledge. I think the line about changing probabilities is there because us humans are bad at understanding probability.