The NCAA Tourney & Automating the Portfolio Process

I am very good at eliminating myself from the equation in many facets of my life.  That is, when it comes to things where I know I am deeply fallible and perhaps even ignorant, I try to automate it thereby reducing the impact of my own worst enemy – ME.  The NCAA tournament bracket is just one example of something I do every year that I know I don’t know much about so I turn to an automated process.  I know that winning the bracket is part luck and part playing the odds.  And if you can increase your odds the luckier you’ll get.

I often talk about knowing yourself, knowing you’ll be wrong, knowing you don’t know everything, etc.  When harnessed, this is a powerful understanding.  Understanding your own ignorance will actually make you MORE intelligent. With the tourney, I know I don’t know everything so I automate the selection process through a wonky quantitative process that takes my personal analysis out of the equation.  For instance, if I just went with my own biases I’d pick Georgetown every year and I’d lose every year.  This year I have them losing to Florida in the third round of the tournament.  I don’t really care.  It would be great if my alma mater won, but you don’t increase your odds of winning anything by letting your emotions and inherent biases cloud your judgement.

Portfolio construction shouldn’t be much different.  You should have a process guided by rules within an overarching plan.  Of course, discovering a plan that works is easier said than done, but letting your own worst enemy constantly change the rules of the game as it goes along is a certain road to catastrophe.

PS – Lousville will beat Florida in the final.

-------------------------------------------------------------------------------------------------------------------

Got a comment or question about this post? Feel free to use the Ask Cullen section or leave a comment in the forum.

Cullen Roche

Mr. Roche is the Founder of Orcam Financial Group, LLC. Orcam is a financial services firm offering research, private advisory, institutional consulting and educational services.

More Posts - Website

Follow Me:
TwitterLinkedIn

Comments

  1. Automating decisions is an extremely fascinating argument. I’ve spent some time thinking about it but I don’t have an aswer. Is it really possibile to do it ? I think in normal circumstances an automated process is by far superior but in extraordinary circumstances it will fail miserably. The reason is the nature of the human brain. It is fallible and sub optimal in normal circumstances but on extraordinary circumstances there are some circuits that are activated, programmed by mother nature for our own survival. I don’t have a definitive aswer, still thinking, really fascinating stuff. So may be, let the machines do some work but keep the finger on the exit button.

  2. Would you consider instinct to be part of your own personal analysis, and as such would you consider it too as flawed? Maybe it’s your instinct that Georgetown won’t advance because they haven’t done so in the recent past. Maybe you just would rather not be disappointed. If they lose and are not advancing in your bracket, no problem, the bracket lives on. If they win, your bracket suffers, but your team wins so you are happy. Either way, you win. The point is that as a human being, it’s hard to remove yourself from the equation. Emotion, instincts, facts – all should be part of the equation.

  3. Cullen – Have you ever built any tools using Google Drive (spreadsheets, apps script, etc.) ? They actually have built in Tournament function that can retrieve all types of tourney stats and real-time data. Not a bad starting point for building an expert system.

    http://support.google.com/drive/bin/answer.py?hl=en&answer=1388875

    I have not built an automated NCAA expert system, but am working on tactical trading system with Google Drive as the basis.

  4. Hello Alberto,
    Those are some interesting thoughts. I see the same thing happening in chemical plants when controlling the process. The best operators are those that leave the controllers in automatic for the most part and only take control when something out of the ordinary happens. The worst “chase the chart” and constantly mess with the controls.

  5. That sounds like a classic problem with any feedback control system. As long as you’re in the operating region that your controller was designed for (very often approximated by linear ODEs… i.e. the “linear region”), then everything runs well… it’s when it get’s outside this region that your automatic feedback controller may actually start to make things WORSE!

    As an example consider a case where negative feedback helps reduce the “error signal” (the difference between where you’d like to be and the place you are), but if for some reason something should get out of whack and the time lag for this feedback data is increased enough to change your negative feedback signal into a positive feedback signal, then your control signal may be adding energy at the wrong times (you start to resonate) and you end up blowing up the system you were trying to stabilize.