News for hackers. Stuff that matters.
Re:How Big Banks Became Our Masters by marcosdumay on Wednesday September 27, 2017 @02:27PM (#15350784)
Scale should certainly be taken into account. Still, differences in scale are no excuse to completely invalidate the data points without a good reason why it should matter this much.

The examples we have of countries letting their banks fail recovered¹. The onus is on anybody disagreeing to show how the US differences are relevant. Just pointing at them and screaming does not make it.

1 - AFAIK, all of them. What is interesting because it includes Brazil at the late 90's (the one single time Brazil grew in recent history).

Re:How Big Banks Became Our Masters by jpttsn on Friday September 29, 2017 @01:31AM (#15362953)
Differences in scale are a great reason to throw data out.

Past experience gives us lessons for similar cases. The onus is on anybody using a model to ensure the use case is similar to the training data.

The US and Iceland are nations but it’s unreasonable to get hung up on that similarity of category. The nation is not a label for grouping economies that are comparable.

I don’t like bailouts either, for the record. My gripe is only with the scale question.