Greetings everyone, I hope you have had a happy holiday season and are optimistically looking forward to the 2019 calendar year and DFS season. With the year coming to a close, I thought it would be a good time to publish a post reflecting on ASA's year past and looking ahead to the future.
I think the site finished 2017 and entered 2018 with a lot of momentum. We had a team of about four to five working on ASA's content offering; things like statistics-forward DFS content, back end site development, data acquisition and architecture, and social media engagement. We had a strong core team of Brandon, who provided experienced DFS strategy and site direction; Mason, who worked as the site's data architect; and myself, who put together the site's analytical content. But over the course of the year the team dissolved, and our overall investment in the site declined, essentially reaching zero over the summer after the end of the NBA season. I think a pivotal staffing loss was the departure of Mason, who after graduating college moved on to a full-time job. And we have since struggled to maintain the richness of data needed to provide a viable offering.
I've taken time in the last month or so to reflect on the decline of ASA's offering, and in my opinion there were three shortcomings that eventually led to the site's dormancy.
- Data quality + frequency - Following Mason's departure, we were unable to collect and leverage quality data. I was able to collect simple DFS data, but without rich box score and game variable data, it became difficult to develop profound content. Additionally, we struggled with provided up-to-date data on a daily basis. This problem arose from the combination of our inability to collect data at the necessary frequency and our inability to publish that data to the website in a timely manner.
- Real-time content - Related to our inability to provide fresh-daily data, our content offering and tools weren't as real-time functional as I would have liked. Even our more dynamic tools were running off of stale data, which in effect made their strategic information gain sometimes outdated.
- Content dynamism - I felt that our content lacked dynamic value, they tended more toward historic summary and less towards response to the present and prediction towards the future. While I do think evaluation of the past is important for processing information about the present in order to predict for the future, we didn't bridge this necessary thought gap as well as I think we could have.
I have been developing my own data skills, in particular in the areas in which Mason was so instrumental. I have developed data acquisition and management skills such that I envision a possible future in which ASA can return to being a viable offering. Over the pas month I've worked to compile historical NFL, NBA, and MLB data, and have processes in place such that we will be able to update our database daily across these three sports. I'm still working to set up an infrastructure that will allow us to access day-of pre-game data, which is so instrumental to process in lineup construction. I have already put together an application that will allow site visitors to access our repository of historic data, I will continue to develop our data collection process so that you will have access to increasingly rich and increasingly up-to-date data. It is my firm belief that in order to get ASA back to a viable offering, we need to start with developing a bulletproof and scalable data collection and implementation framework. This work is in progress.
It is my goal to get ASA back to subscription-viable offering. But in the near future, ASA will remain a free, open access site. Our intention for doing this is twofold: 1) we want as many people as possible viewing, using, and interacting with our site as we develop new tools content and 2) we want to get your feedback on our content offering. The second point is of extreme importance to us. We are constantly trying to come up with fresh and usable content offerings, but it is an imperfect process. Brandon and I try to think of content that will help our style and process, but we acknowledge different players have different processes and strategies. And as such, different players need different tools to help them realize their highest return. So it is my hope that by making the site an open access offering in the short term, we can get your feedback on what tools you find useful, what tools you don't find useful, what adjustments you would like to see made to our tools, and what strategies you are trying to implement that our content doesn't support.
We're always trying to find the right balance in the complexity of our content offering. I think we have tried to avoid being a "projections site". While I think good projection is a viable offering, it is a very saturated space, and it is hard to develop an offering that immensely different from what is already out there. It is my goal to develop some sort of projection model, because I do think it is a good way to levelset player expectations, but I would like to focus more of my time on developing content that explores more nuanced aspects of DFS strategy and projection such as anomaly detection, prediction of high-variance outcomes, and latent factors and player tendencies that contribute to fantasy production. I think since the beginning, our goal has never been to develop an offering of "play these players, or follow this algorithm, or pay attention to these variables... and you'll print money". Our mission has been to develop content that can be leveraged through an inexact and player-specific process to improve returns for all players of all levels of DFS skill and engagement. In the mantra "give a man a fish and you'll feed him for a day, teach a man to fish and you'll feed him for a lifetime", we're far more interested in teaching you how to fish (or giving you more fishing tools if you've already developed your own effective process).
And as alluded to in previous sections, our goal is to build dynamic and interactive tools, so that you can use them in a process that works best for you, even if that process is different from the process that works for others. I've already started to develop some tools (and have daydreamed on countless more), like the beta version of this NBA "inactive response" app. I think this tool aligns with the vision I have for future ASA content offerings; the app can be used to identify lineup targets who benefit most from injured teammates being ruled inactive. It is flexible: users can compare players across FanDuel or DraftKings scoring formats, they can compare players in terms of scalar margins or percentage margins. It is interactive and dynamic: users can quickly toggle across different teams and players, easily evaluating scenarios for the entire slate. And it is transparent: through the app, users can download the raw data that feeds into the app for their own personal exploration.
Lastly, we're hoping to re-vamp our site layout to improve user experience and flow. We've always struggled with how to embed our applications into the website, so that you won't be redirected to a non-ASA url (like I have don above, sorry, I hope the days of doing so are numbered). I've got what I think will be a solution to this problem, but it will likely require a bit of a site makeover, probably moving the site to a new web-hosting platform. The moves that will need to be made to improve site flow are a bit unclear, but while most of our focus right now is on developing a data infrastructure and building up our content offering, we are not ignoring the need to constantly improve site user experience.
If you have any questions, feedback, or input about future of ASA, don't hesitate to reach out to our team either through the site feedback widget or Twitter. Thanks for your support and patience, and I look forward to our journey in the coming year. Best of luck to you all in 2019.