My theory is that the supposedly "predictive" models serve more to understand the past than to predict the future.

What usually happens is the following: the past is observed, in this case of Bitcoin, some regularities are observed (which to some extent are arbitrary), and from there a model is assembled assuming that the future will be similar.

I prefer to rely on events that happen in the real world rather than on charts drawn on the basis of a continuation of patterns. Like the miners' ban in China last year, which was not foreseen in any technical analysis or predictive model, also the reduction of liquidity by central banks, or the cascading liquidation of leveraged positions as we have seen recently. All this is much more important than standing in front of a blackboard and drawing future lines based on the supposed regularities you observe in the past.

In the case of PlanB, he put together a very elaborate and logical model based on supply, and I understand that he does not want to give it up for lost after so much time and effort invested in it, but I think he would do better moving on and stop making a fool of himself.