LOL. No DA koolaid drinking here. We left the DA over a year ago and wouldn't go back.
Numbers don't lie, but models have limitations.
YSR's predictions are great, but as I said before, the numbers are not as good for teams further away in the graph. They favor teams that play more against lessor teams. If two teams have the same rating and one has played more games against lessor competition, the other team is more likely to win. This is why non SoCal teams have inflated numbers (and fare so poorly in tournaments vs. SoCal teams). It's not ysr's fault. It's how the model works. There's no possible way it can compensate for the lack of games between DA and non DA because the longer the cycle (using data from A played B played C played D to predict A vs. D) the less accurate the prediction. I'm sure the creator of ysr, whoever he may be, is aware of this. It's just how the math works.
If ysr is considering a former DA team as the same team, that can help tighten the cycles, but only so much because when a team ages out of DA, they usually loose one or two of their top players, which can make a big difference.