I read an article about your book and brief interview in the Kansas City Star this morning. Bravo!!

I would like to share a few observations.

I was a “buy side” advisor for several large hedge funds for about a dozen years. My position is anything that is predictable can be gamed.

The central problem with algorithms is the construct. While algorithms are based on “hard data,” there are two common flaws. First, there are often forecasts, assumptions and unproven theories included in the data. Second, the weighting of the data is often wrong. Garbage in, Garbage out (GIGO).

This suggests we should more carefully evaluate the data and its importance. The problem here is the people writing the algorithms seldom have the expertise in the various fields to do this. Let me share a very fundamental example.

When my daughter took her first Algebra course she was allowed to use a calculator. The problem is she had little ability to mentally estimate the logical outcome for even simple Algebraic equations. She simply didn’t have a good enough grasp on basic math to know if an answer was logical or not.

Due to this, if she entered the wrong data, she had no idea if the result was logical, and would assume that since it was produced by a calculator, it was right. We took the calculator away so she could learn the importance of the variables in the equation (get a feel for how numbers work together).

With the above thoughts in mind, I think you can expand your reach considerably by pointing to the fact algorithms are simply math, and math is fact. However, if the data are bad, or weighted incorrectly, the result of an algorithm will likely be wrong.

I read an article about your book and brief interview in the Kansas City Star this morning. Bravo!!

I would like to share a few observations.

I was a “buy side” advisor for several large hedge funds for about a dozen years. My position is anything that is predictable can be gamed.

The central problem with algorithms is the construct. While algorithms are based on “hard data,” there are two common flaws. First, there are often forecasts, assumptions and unproven theories included in the data. Second, the weighting of the data is often wrong. Garbage in, Garbage out (GIGO).

This suggests we should more carefully evaluate the data and its importance. The problem here is the people writing the algorithms seldom have the expertise in the various fields to do this. Let me share a very fundamental example.

When my daughter took her first Algebra course she was allowed to use a calculator. The problem is she had little ability to mentally estimate the logical outcome for even simple Algebraic equations. She simply didn’t have a good enough grasp on basic math to know if an answer was logical or not.

Due to this, if she entered the wrong data, she had no idea if the result was logical, and would assume that since it was produced by a calculator, it was right. We took the calculator away so she could learn the importance of the variables in the equation (get a feel for how numbers work together).

With the above thoughts in mind, I think you can expand your reach considerably by pointing to the fact algorithms are simply math, and math is fact. However, if the data are bad, or weighted incorrectly, the result of an algorithm will likely be wrong.

Best Regards,

Paul

LikeLike