问题如下：
The model that Martin is tasked with designing will likely be most effective:
选项：
A.for testing new markets.
in a well-understood market environment
during periods of higher than normal market volatility
解释：
B is correct. Many trading problems are ideally suited for machine learning analyses because the problems repeat regularly and often. For such problems, machine-based learning systems can be extraordinarily powerful. However, these systems are often useless—or worse—when trading becomes extraordinary, as when volatilities shoot up. Machine learning systems frequently do not produce useful information during volatility episodes because they have few precedents from which the machines can learn. Thus, traders often instruct their electronic trading systems to stop trading—and sometimes to close out their positions—whenever they recognize that they are entering uncharted territory. Many traders shut down when volatility spikes—both because highvolatility episodes are uncommon and thus not well understood and because even if such episodes were well understood, they represent periods of exceptionally high risk.
老师可以解释一下这个答案吗？
机器学习的过程可以理解为从一堆纷繁复杂的数据中学习一个惯性，比如在一堆天鹅里通过对特征的描述确定一只天鹅应该有的样子，以后可以从一群动物或者景物中确定一只天鹅，其特征是：白色、有翅膀等等。但一旦出现了一只黑天鹅，他就不能识别。这就是黑天鹅事件。