Models/theories that are good at predicting are not necessarily true.

When the theory is not falsifiable, observations that support it don’t increase the likelihood it’s true, it just props up assumptions.[^G1]