Traders matter more than their code

I really love Twitter. Yes, I've said it. True, it seems to absorb time, like a never ending sponge. However, on the plus side I have learnt a huge amount about finance by using Twitter. Recently, I started following @krscapital and I recommend you follow him too, if you're a finance Twitterer. He recently tweeted a link to an article entitled Wall Street Techs Take Secrets to Next Job at Their Peril (on Bloomberg) alongside the following quotation taken from the same piece: "strategies that were once in a trader’s head have taken physical form as stored code"

Rather than attempting to fully discuss the piece, I instead wish to focus more on the quotation above. The gist of the quotation is seems to be the that code can replicate a trader's brain. To some extent that is true. Within code, we can create a model which trades on the behalf of a trader. Simply flick a switch and it starts trading all by itself and the money comes in. However, is it that simple? No.

Who decides how much capital to place on a trading model? The trader of course. Hence, the code does not totally remove a trader from the business of trading. It is simply, that he or she, has a slightly different role to play. Hence, a trader still has considerable responsibility.

Yes, the code which encapsulates a model is important. This is particularly true, when it comes to very high frequency trading. In this instance, the speed of the code becomes a crucial factor in its success. However, if we think more broadly, far more important is the trader who actually builds, understands and continually hones a trading model. Without a deep understanding of a model, it becomes impossible for a trader to have the confidence to place risk on it.

The value of traders is not what they have achieved already, but what they will develop in the future. A trader who builds a single model in his (or her) whole career and then keeps it like a museum piece, by contrast seems less valuable. After all, trading models generally don't work "forever", particularly those which are more "alpha" in their approach, which have higher risk adjusted returns.

If models did last "forever", maybe, just maybe, a piece of code could be more valuable than a trader. Code is a tool for a trader to make money. By contrast, code without a trader holds far less value. Furthermore, the notion that there is some "secret sauce" a special code for successful trading seems pretty unlikely in my view and having talked with many systematic traders over my career. If anything, I suspect the notion of a "secret sauce" is more marketing exercise than anything else.

To read more about this topic please see my book Trading Thalesians which has a chapter on the notion of whether there is a "secret sauce" to trading (mixed in with a bit of ancient history).