It was not post hockey, it was part of the SAP submitted to regulators, pursuant to regulations, prior to unblinding. The result happened to be very positive, it also could have been otherwise.
The crossover was in fact part of the total trial and yes, not having a true placebo arm when the drug works so well, causes confoundment. That is a key reason why the FDA and other regulators allowed for a statistical concurrent external and very large placebo arm to be used as the placebo arm instead. However, placebo arms are generally required where the differences are very slight in oncology. In this instance you have substantial signs of extended survival with virtually no side effects or toxicity at all.
Most experts seem to be optimistic as to what this means.
You can keep playing up a false technical argument that you clearly do not even understand and a counterfactual argument about data dredging that falsely requires that they only updated their SAP after unblinding, which is false by the historical record. But you’d just be spreading misinformation with your false arguments.
Market Data 
Markets 
AI
