Teams that win fantasy championships have one, often multiple, player(s) that significantly outperform ADP. Conservatism rarely pushes teams to the top of their leagues’ standings. In fact, it often works against them, anchoring them to the middle of the pack. This isn’t a new idea and we often talk about upside on the site. Historically, we’ve offered a number of tools that provide snapshots of a player’s low and high-end outcomes. Over the last couple of years, we’ve started to include visualizations, outlining a player’s entire distribution, but I don’t believe that I’ve spent enough time championing the importance of focusing on these visualizations or the historical distribution scores that I’ve used to describe them. We can talk about upside and why we need to chase it, but if we’re not thinking about it correctly, we’ll be chasing the wrong thing.

To get the most out of this article, I’d recommend first reading about the range of outcomes app and the process used to create it.

Taking A Step Back: What Is A Distribution?

If you didn’t read the article linked above, here’s the condensed version. To create a realistic range of outcomes for Dalvin Cook’s 2020, The Range of Outcomes App reviews the stats that he produced in 2018 and 2019.^[1]For 2019 rookies, it only considers 2019. It then finds players that produced similar two-year stretches in previous seasons. It does this by looking at the stats that are useful in predicting future performance. It then compiles the stats that the matched players produced in the season following the one that matched with Cook’s 2018/2019 stats. Next, the app gathers the fantasy points that these players produced in the subsequent season and uses them to build a histogram.

A histogram is a chart that groups values into ranges represented by bars. The height of a bar shows the number of values that fall into that range. In our fantasy football context, each bar represents a range of fantasy points. The height of each bar represents the number of matching players that scored point totals falling within the range associated with the bar.

Cook’s histogram tells us that the most common fantasy points scored by his matches lie somewhere between 17 and 20 PPR points per game (PPG). In fact, 12 of his matches scored between 17.5 and 19.5 PPG. Further, 90% of his matches, or 90% of his distribution, scored between 11 and 23 PPG.

Compare Cook’s distribution to Leonard Fournette’s. Which player owns a “better” distribution?

Why Distributions Are Useful

It’s pretty clear that Cook owns a stronger range of outcomes. In addition to having more matches score north of 20 PPG, a smaller number of his matches scored less than 10. While both player’s matches were most heavily concentrated around 17 to 20 PPG, Cook had more matches falling into this bucket. Assuming that the process used to create these distributions is sound, it’s reasonable to feel better about Cook’s 2020 outlook. He has a higher ceiling, lower floor, and a more favorable concentration of matches.

However, if we only select a handful of data points, sourced from the player’s matches, we might not get a picture as lucid as the one painted by reviewing the above histograms.

The table does indicate that Cook has a stronger floor when reviewing 25th percentile points, but it might give the reader the sense that baseline expectations for the two players are more or less the same. However, from absorbing the distributions above, we know that a higher percentage of Cook’s matches scored more than 15 PPG. In fact, only 56% of Fournette’s matches scored more than 15 PPG, whereas, 70% of Cook’s did. Without including the distributions in our analysis, this nuance and useful context are lost.

Why Does This Matter?

There a number of reasons that this is really important. For starters, we’re often not considering players in a format as tangible as the one presented by the Range of Outcomes app. When talking about subjective projections — ones built by assigning market shares and efficiency — we generally focus on