The Lending Club goal is to minimize the investment risk of providing bad loans, the loans with a high probability of defaulting or being delayed, but also to avoid rejecting good loans and hence losing profits. Here, the main criterion is driven by accepted risk - how much risk Lending Club can accept to be still profitable.

Furthermore, for prospective loans, Lending Club needs to provide an appropriate interest rate reflecting risk and generating income or provide loan adjustments. Therefore, it follows that if a given loan has a high interest rate, we can possibly infer that there is more inherent risk than a loan with a lower interest rate.

In our book, we can benefit from the Lending Club experience since they provide historical tracking of not only good loans but also bad loans. Furthermore, all historical data is available, including final loan statuses representing a unique opportunity to fit into the role of a Lending Club data scientist and try to match or even beat their prediction models.

We can even go one step further-we can imagine an "autopilot mode". For each submitted loan,we can define the investment strategy (that is, how much risk we want to accept). The autopilot will accept/reject the loan and propose a machine-generated interest rate and compute expected return. The only condition is if you make some money using our models, we expect a cut of the profits!