I just recently graduated from Flatiron’s data science bootcamp program and wanted to start a project to keep my skills sharp. I decided to combine two of my favorite passions: python/modeling and sports. I have done some research to figure out what sport is the easiest to predict to start with. Then once I perfect the first sports model, I will move on to try and predict the harder sports. The sports I am choosing from are both professional and collegiate level.
College Soccer, College Football, College Basketball, MLB, NBA, NHL, NFL, Professional Soccer
Through some reasearch I have found the following points to be extremely useful in deciding which sport to start with:
- Picking a sport with only two outcomes will be much easier to predict than a sport with three outcomes, 50% chance vs. a 33% chance. We obviously would like the 50%.
- Sports with 50% chance: MLB, NBA, College Basketball, College Football
- Sports with 33% chance: NHL, NFL, College Soccer, Professional Soccer
-
The type of sport should be taken into account. So what does this mean? Is the sport a team or individual sport? What is the team make up, is there a density of talent? The less moving parts and or players on the pitch, ice, field, or floor is usually more statistically in favor of predicting the outcome.
-
General professional gambler standpoint seems to be pointing to the NFL as being the easiest at predicting and has the least amount of abnormal events slash under dogs winning.
- Final point seems to be that there really is not one sport that is the easiest to predict and or bet on. The number one advice I have seen is to bet on the sport you are the most passionate about.
Sport Selection:
With all of the research and points taken into consideration, I am down to college basketball and college football. Both of these sports can not end in ties and are my two favorite sports by far. I believe just from my own sports knowledge that college basketball can be very unpredictable, but I would like to model this sport. As this sport is in season currently, and I figure I will want to model every sport eventually so why not start with college b-ball. March Madness after all is right around the corner.