The Numbers Don't Lie: My Journey in Understanding Statistics

I've been at this blog for sometime and I have been really trying figure out why I' am so bad at my picks.  I've been recognizing my bias  I think jiu-jitsu specialist will always pull off an amazing submission and my other problem is my fandom also gets in the way. I've started to go back on my picks to see why I thought X was going to beat Y.  What were my preconceptions. Where did I get those preconceptions and were those thoughts valid.  One fight in particular that I went back to analyze was the Clay Guida vs Erik Koch.  I thought based on Clay Guida's last fight with Brian Ortega that Guida's chin was gone and he was aging out of the sport.  I was wrong in this assumption but the odds makers made the same assumption. But I went back and looked at that Ortega/Guida fight.  Guida looked great in the first round.  He dropped Ortega in the 1st. The second was up for grabs but I gave Guida the nod for the 2nd round.  Ortega did his thing and pulled out another late round finish after pressuring Guida landing a devastating knee to put Guida away.  Understanding how dangerous Ortega was off his back, Guida really didn't commit to takedowns out of fear of being submitted.  Take that into account Guida not having one of his tools at hand and winning the striking for 2 out of 3 rounds.  Guida had a better shot than expected and proved that during the Koch fight. I learned that I really need to take stock of the fighter's ability based on their performance rather than inserting my own narrative with any analysis.

After recognizing that another problem with my picks is my lack of total recall of every round in a fight.  There are a lot of fights.  The UFC puts on fights every week and a lot of time I've got a decent buzz on so its hard to remember every round of every fight.  I do re-watch parts to jog my memory but that's also time consuming as well.  So I decided to go on to FightMetric and look up the actual stats for each fight.  I noticed that they gave access to the fight data for academic purposes.  They had an article titled "Predictive Modeling of Mixed Martial Arts Fights Using Novel Fight Variables" I started reading the article and quickly found that I had no idea what those formulas meant.

After that I was a man on a mission. Understanding that building my own model combined with my actual fight knowledge. I would eliminate my bias and increase the accuracy of my fight picks.  I found an online course through the UT Austin called Fundamentals of Data Analysis.  Its great course I'm learning the basics right now but it uses a data modeling program called "R".  So by the end of the course I should be well on my way to building my own model.   I'm quite excited about learning this material. This has probably been the first time I've enjoyed learning in a long time.  I went through college begrudgingly especially graduate school.  The only thing that kept me motivated in Grad School was understanding the government was paying for my degree and if I did not pass I would have to pay for it on my own.  The self pacing of the course is a gift and a curse.  It gives me flexibility but had a bit of a break as of late.  Work has been busy but its going to slow down again and I can start teleworking again which will allow me to work on me to wrap up these stat classes.