> I would expect a similar approach (progressively build a graph of the most promising breaking points) to be effective. Why wouldn't it be the case here? That

> I would expect a similar approach (progressively build a graph of the most promising breaking points) to be effective. Why wouldn't it be the case here? That's…how he does it.