Curtis ended up submitting a help ticket that we resolved privately.
We did tell him to test them via ice water (32) and boiling water (212) and to plug the probes in before the power.
Our probes are tested against NSF equipment and have an accuracy within +/-2 degrees Fahrenheit.
Here is what he did to resolve his issue.
seemed to be more accurate when I plugged the probes in before powering on. that's strange that it makes a difference but lesson learned.
How can this be marked as solved when there isn't any public solution?