Sounds like the DM theory is great for postdictions but not for predictions, like you can't know in advance whether it is applicable or not