Where Does Edge Come From? The Value of Unique Data

Only Three Ways to Win

There are hundreds of different ways to win wagering on sports. Yet, at its core, every winning strategy requires at least one of the following to be true:

  • 1.      You have better data than your opponents.
  • 2.      You can use and interpret that data more effectively than your opponents.
  • 3.      You can act quicker or more efficiently than your opponents.

I think it is good practice to ask yourself which one of the three statements is true before you make any type of bet or enter any market. If you cannot pinpoint your edge’s source, you might still have an edge. However, it’s much less likely.

None of the strategies “above” are better than others, it depends on your strengths and goals. I discussed some of these pluses and minuses in my article about scaling in sports betting. In this post, I focus on the distinction between having better data and interpreting data better.

Where Do You Want to Compete?

Which of the following would you rather be:

  • 1.      A 6/10 modeler with data nobody else has access to
  • 2.      A 10/10 modeler with publicly available data

For me the answer is clear. I’d 100% rather be a 6/10 modeler with data nobody else has access to.

My background is in math, statistics, and data science. I’ve participated in markets where I’ve felt I had better data as well as markets where I felt I had better modeling. However, I currently focus on areas where I can have a data advantage. I want to avoid markets where modeling would be my only advantage.

The reason is simple. Having data that’s 10% better than my opponents give me a much bigger advantage than being a 10% better modeler than my opponents.

Interpreting data and modeling has never been easier. Today’s software enables a second-year college student to apply a random forest or neural network with a few lines of code. Of course, being able to run a model is not what solely differentiates a good analyst from a bad one. However, the modeling playing field is very crowded and competitive. Ask yourself, is this where you think your edge will come from?

Data Everywhere

So, you’ve decided you don’t think your most effective way of gaining an edge is through data analysis. What is the alternative? It’s to look for markets or bet types where you can generate an edge by having unique and valuable data. Many people think that in today’s world all the data is already out there. They believe there are no stones left to uncover. I wanted to share a couple examples of places to look. There are examples for both the technically inclined and those that are less so.

Having better scraped data sources has helped me in the past. One example is target data for college football. Until the past couple of years, it had been very challenging to come by receiver target data for the college game. You can project receiving statistics much more accurately if you include targets and air yards. Before this data became publicly available, it required scraping play-by-play data. Then, you had to parse the play descriptions to get the targeted player. Cleaning and parsing the data was a large undertaking. But it was very beneficial for CFB betting. For those who are technically inclined, the harder something is to scrape and turn into a usable format, the more valuable it will be.  

Here’s some other examples that are more accessible to those without coding skills. You create your own data. You can collect all sorts of data in all sorts of different ways. Here are some examples to give you some ideas on what types of way you can create your own data.

  • 1.  You can manually collect ownership data in DFS or some Pick’Em peer-to-peer contests. This can help you improve ownership projections or understand your opponents’ tendencies. These games are very much about predicting your opponent’s tendencies. If you have data that allows you to do this, you’ll have an advantage over the competition. Peer-to-peer games offer many chances to create unique, actionable data.
  • 2.  Collect historical odds data for obscure markets or props. These aren’t well tracked or publicly available. Historical lines and odds databases can be extremely helpful for a variety of reasons.
  • 3.  You can also collect data on more sport-specific stuff. For example, team tendencies around the coin toss and first plays. Or any other data not easily captured in a box score or game summary. Especially with the rise of SGPs using uniform pricing algorithms, having data on unique things teams or players do can be very valuable.

The options are really endless. It is no doubt that the collection of unique datasets can be a substantial undertaking of time. However, this is often where the biggest opportunities for me have come and where I will continue to look.

Please enable JavaScript in your browser to complete this form.
Email Address

Leave a Reply