Michael Shashoua: Figures, Guts and Glory

Taking part in my second fantasy football draft last month for the popular interactive competition based on drafting virtual teams of real players from the US National Football League and then tallying scores based on players’ stats during the season, it occurred to me that the process of the fantasy draft has many similarities to securities trading and applications of data and information supporting trading and risk management processes.

In my league’s draft, participants got one minute to make a pick when their turn came around. This resembles the high-pressure, snap-decision trading environment. Between picks, as references, I had my own previously created watchlist of players of interest; a live updated feed of players still available to choose from, ranked by perceived value, including an average of when they were picked in similar drafts; and my own qualitative information added to this mix—a New York Times fantasy football evaluation that had tweet-length comments on individual players, skewing positively or negatively with few middle-of-the-road or non-committal comments.

Key Data
The live feed included a key piece of data—a number for the average position at which that player had been picked in previously held drafts on the system administering our league. So, for example, in a much later round I had the 139th overall pick. Consulting quickly with the live feed, I chose Baltimore Ravens running back Bernard Pierce, who was on average chosen 122nd, but was still available in our draft. At that late stage, not knowing some of the more obscure players, that piece of data was a good reason for this choice—hoping I was getting a “steal” of some sort.

This is akin to reading data and concluding that a security is, in effect, undervalued, and worth buying. However, you cannot discount qualitative analysis. Again, for example, at pick number 42, I was looking for a wide receiver for my roster, and had two choices still available who were close in rank—Roddy White and Larry Fitzgerald. On average, Fitzgerald had been drafted earlier than White had, but as I checked that New York Times analysis, I saw these comments:

• On White: “Top 10 WR once fully healed in ’13, poor defense will force tons of air time in ’14.”

The object lesson is that risk management invariably includes more than binary choices dictated by data.

• On Fitzgerald: “31 and likely final year with the Cards. Hasn’t cracked 1,000 + yards since 2011.”

I picked White. This was like researching a company and finding some piece of information about the product it is developing or the management culture that isn’t necessarily evident in the stock price on a given day. You may say this requires some instinct and gut feeling as well—because one’s interpretation of the facts can be subjective.

And in the securities reference data world, the object lesson is that risk management invariably includes more than binary choices dictated by data. Experience, expertise and market knowledge should—or ought to—play a part.

Late Pick
To illustrate this, let’s revisit the late pick of Bernard Pierce. I had also put him on my watchlist because his teammate Ray Rice was suspended for the first two weeks of the season for off-field issues. So Pierce may have gotten more yards as a result, and thus appeared to have more value than his average draft position showed. Seeing him still available at that late round, I chose him above others that may have seemed like smarter picks if you only looked at raw data.

As you read this, the US football season will already be under way and all these choices may have proved disappointing for unforeseen reasons, but the decision-making processes that led to them, through a combination of hard numbers with qualitative evaluation and knowledge, show how to produce a more useful form of reference data.