The view from here
Why Data Should Be a Part of Your Fantasy Football Draft StrategyWhether you are drafting your fantasy football team or allocating marketing budget, the BKV Analytics team can create an actionable, data-driven analysis that will put you ahead of the competition. Big Data”) that making informed decisions has become a cumbersome task; so, we need to get smarter about how we make decisions. Data-driven strategies can give you a serious advantage when drafting your fantasy football team or optimizing digital media. At BKV, we can’t help but apply our analytical skills to a data-rich hobby like fantasy football. Although the complexity of our fantasy football analysis is relatively trivial compared to the predictive modeling work we do for our clients, there are several common elements that are part of any data analysis. Any successful analysis starts with reliable data and wise variable selection. For BKV’s Analytics department, that means evaluating available data sources to find appropriate, accurate, and granular records of marketing spend, customer data, media interactions, orders, et cetera. In the fantasy football world, every major sports outlet has some sort of player performance projections that differ on an individual level but tell roughly the same story with roughly the same accuracy – there are definite tiers among players at each position. The projections as useful in quantifying those tiers, but they are not reliable enough to base your entire draft strategy around. In addition to the projection data, we include a few additional variables to help gauge a player’s risk and opportunity including: age, strength of schedule, points above replacement, and average draft position (ADP). In order to make an analysis worthwhile, it needs to be actionable. At BKV, we strive not only to unearth cool and unexpected findings in our marketing analysis projects but also to ensure we can implement those findings in our clients’ marketing strategies. In our fantasy football analysis, we join all of that data to create an ‘actionable deliverable’ – a sortable, searchable spreadsheet used to identify which players we would be willing to take where they are usually drafted, based on ADP. (You can download this here). After reviewing the deliverable, we categorize players into five buckets: Studs, Sleepers, Value, Lotto, and Filler. Here’s a detailed breakdown:
Studs (1st-30th pick): We want proven, top-tier players with the least risk for our largest investment (i.e. early rounds picks). Regardless of where you get projections, there are about 30 players that are considered the best and safest bets.
Sleepers (31st-60th pick): There are always a few middle tier guys that outperform their projections; so drafting those players is a great way to maximize ROI. When looking for sleepers, we primarily look for young players in a situation where they will be a clear starter.
Value (61st-90th pick): The difference between top and middle tier QBs and WRs is smaller than that of RBs (per points above replacement), so we use our middle rounds to pick safe-bet QBs and WRs with some upside rather than the left-over risky RBs.
Lotto (91st+ pick): Lotto players are guys that have a lot of potential (like Slepers) but their playing time prospects are unclear. We’re willing to accept their risk with late round picks.
Filler (91st+ pick): Kickers, Defenses, and TEs are very tough to predict and are not significantly different from one to the next; so we wait until the late rounds to pick them. We’ll take a Kicker on strong offensive team, a TE with a good QB, and a decent Defense with an easy strength of schedule.In any analysis, there will be ‘outliers’. Even the best analyses will not be able to explain some events, whether it be an unexpected boost in sales, a drop in email opens, or a player that comes out of nowhere to become a fantasy football stud (like Alfred Morris did last year). Unfortunately, this is the simple reality with analysis—some things cannot be predicted and some things do not follow the trend lines. Over time, variables change and you must refine your analysis. During the preseason, a handful of guys will get injured opening the door to the next man up; and a few guys will unexpectedly win a position competition. To ensure we have up-to-date data, we update the ADP data every couple weeks and the projection data whenever the sports outlets update their projections. For our clients’ analytics projects, we recommend refreshing the model each business cycle (e.g. quarterly, annually, bi-annually) to ensure that our analysis is based on current, relevant data. At the end of the day, analytics doesn’t make the decision; it helps you make the decision better. Every single factor that can impact a football player’s performance—or a business’s marketing campaign—cannot be contained within even the most comprehensive analysis. No matter what, part of your decision making will still be an experience-based gut feeling. A fantasy football analysis helps narrow down the players you should take, but you still need to make a judgment call when deciding on which of the several viable player to spend your one pick. If you’d like to learn more about how BKV Analytics can help your business make better marketing decisions, click here for more information on our capabilities and offerings. If you’d like more info on fantasy football, watch the series The League.