- Fantasy Baseball Recap: Tanaka Rebounds
- Fantasy Baseball recap: Nick Pivetta continues good start
- Week 4 Fantasy Baseball Adds and Streamers
- Fantasy Baseball Review: Jeff Samardzjia is Back!
- Fantasy Baseball Waiver Wire: Jed Lowrie is Smoking Hot
- Fantasy Baseball Recap: Luke Weaver Clobbered in Chicago
- Fantasy Baseball Review: The Return of Jose Bautista
- NHL DFS-Wednesday Rivalry Playoff Edition
- DraftKings Valero Texas Open Preview
- Fantasy Baseball: Rough Nights for Tanaka and Ohtani

# WR Crowding Analysis: 2017 Predictions

### by John Bush and Dave Cherney

## WR Crowding Analysis Introduction

WR Crowding Analysis remains a popular research area in both redraft and dynasty fantasy football as to the real effects on WR 1s playing on teams having other nearly similar WRs (WR Crowding Analysis). We wished to feature and apply a scientific based approach to study this question. All FF players currently have an endless slate of pundit based opinions which will be chock full of biases.

This work represents the first statistical study to quantify the “crowding effects” on WR Ones .

While this material initially pertains to the redraft world, we will dive into the cause and effect of this data as it pertains to the Dynasty demographic.

## Materials and Methods

When considering FF questions, we must use past performance data as a reference for this upcoming season. We gathered the end of season PPR points scores by all major WRs from 2016. Using excel we could then calculate Team by Team’s % of the WR Player Crew vs. the Team WR average. WR1s were defined as the highest scoring WR from 2016 and the WR2s were the second highest scoring WRs and so on!

## Data Analysis and Discussion

Figure 1 presents that data used for the 2016 analysis. We sorted all 32 NFL WR1s as defined about by the difference in the WR 1s % of the Team’s PPR 2016 Total vs. the next highest scoring WRs (WR2s) % of team total. Analysis of this spreadsheet sorting (high to low % differences) revealed 4 obvious groupings to use in statistical testing.

**Group 1 WR1s as an average represented 47.9% of their total Team’s % WRs PPR Production and their Team’s WR2s were on average 26.8% less of the 2016 total Team’s PPR points.****Group 2 WR1s were worth on average 41.6% of their Team’s 2016 PPR total points. Group 2 WR2s were 17.8% less than their Team’s WR1s totals.****Group 3 WR1s were worth on average only 36.4% of their Team’s 2016 total WR PPR points and on these teams, the WR2s only differed by 7.9% which was much closer than the preceding Groups WR**s.

Finally, as we finish our survey of Figure 1 we find that Group 4 WR1s were at 34.2% average of their Team 2016 PPR WR points. The WR2s from these final teams were only different by 1.8% essentially no difference.

This data certainly seems very suggestive and the overall conclusion is not disputable that at least four groups of WR1s will exist for us to draft in 2017.

We next turn to statistical analysis of these data. Did these four different WR1s group form by random sampling or is this a obvious pattern that can be confirmed by scientific statistical procedures. If confirmed we may then use this concept to predict the 2017 WR1s four groups.

## Figure 1. 2016 WR1s Analysis by Groups and WR2 Differences

The data shown in Figure 2 highlights the data summaries derived from our initial analysis from the data shown in Figure 1. The level of WR2s closeness in PPR point production within their team defines the 4 WR1 groups. The analysis of the graph emphasizes the loss of WR1s team value as the WR2s % team’s total goes up. This graph visually shows the WR1s value drops moving down from groups 1 to 4.

## Figure 2. Overview of WR1 Groups and their Tabular and Graphical Landscapes

## Statistical Testing Confirmation

To complete our data analysis research must determine if the differences seen as suggested in WR1s groups 1 to 4 were due to random chance or not. We first use ANOVA testing to determine if any of these 4 WR1s groups are different from each other and if a difference exists only then we can use another test (TUKEY’s) to test whether the group’s averages in PPR points are really different. We will accept only a 5% chance of having a false positive. The ANOVA testing was done and using the 4 groups, the overall p stat value was well below the 5% threshold. (p=.00009%).

We can conclude that these WR1s groups are real and there are differences existing within these WR1s groups. This is the first report of such a statistical finding! Next, we then used the Tukey test to compare the multiple groups to each other and the conclusions are stated below.

## Conclusion from 2016 WR Crowding Analysis.

Competition between the WR1s and 2s cost the WR1s that amount of team value. These WR1s should be highly considered and most likely their ADPs modified lower to account for this negative relationship. This effect will be more as your move bottom to top. Expect a drop of 5 to 6 in positional ADP units on the average as determined by WR crowding analysis.

## Analysis of the 2016 Data From a Dynasty perspective.

We’ll begin with Group 1 in Figure 1.

Terrelle Pryor is the standout here. The former quarterback turned wide receiver really stole the show in Cleveland last year. If you happened to have him on your squad, you were heavily rewarded. He has now taken his show to the Nation’s Capital. He will likely not carry as big of a difference between himself, Jamison Crowder and the multitude of other targets, he should be featured early and often with huge upside. He’ll likely be drafted in the 4rd round of your Dynasty startups, but if you’re a believer, that price isn’t too steep.

The other standout here comes from Tennessee. Rishard Matthews remains under the radar while quietly producing 65/945/9 in 2016. Many experts are predicting a large regression with the new receiving additions including Eric Decker along with rookies Corey Davis and Taywan Taylor. I don’t happen to be one of them and Matthews should come at a fantastic value.

The surprise in Group 2 would be Tyreek Hill. I would have suspected the underachieving Jeremy Maclin, even playing only 12 games for the Chiefs, to have been the leader here. Hill’s Dynasty ADP has fluctuated throughout the year yet with Maclin’s departure to Baltimore, you could be richly rewarded as he once again should dominate the percentage of throws and still be used on punt returns.

An interesting value play within Group 3 comes from the situation in Chicago in the form of 24-year-old Cameron Meredith. Last season, Meredith broke out to a degree hauling in 66 balls for 888 yards and 4 scores. However, the landscape has changed in 2017. Gone is the former No. 1 receiver Alshon Jeffrey who packed his bags for Philadelphia. With only injury prone receivers Kevin White, Markus Wheaton and Victor Cruz as his nearest competition, it is nearly certain Meredith’s workload will greatly increase.

Equally changing is the quarterbacking situation as the Bears bid farewell to Jay Cutler and added an interesting competition between Mike Glennon and first round pick Mitchell Trubisky. Yet Mark Sanchez still ranks second in the pecking order and getting more reps that Trubisky hindering his development.

Lastly, looking at the upside play in Group 4; we will look at an older player that is in a fantastic situation as we welcome Pierre Garcon. Having left Washington, DC via free agency, he has now reunited with Kyle Shanahan in San Francisco. During his last four seasons with the Redskins, Garcon totaled 332 receptions for 3,916 and 17 touchdowns. Breaking it down further, the season he and Shanahan were together in 2013 he had his highest production with 113/1,346/5. Even better news could be on the horizon regarding the quarterback. While Brian Hoyer is serviceable, there is a strong belief that Kirk Cousins will be steering the ship next season. Fortunately, many Dynasty owners don’t covet older player so he may just come at a bargain.

## 2017 Predictions for Redraft

We now have a basis to judge the 2017 WR1s as to their potential grouping and ADP penalties to apply. The current MFL ADPs were used to generate 2017 seasonal PPR point projections.

These projections for each Team’s players were summed to get the predicted Team’s WRs total. Next each player’s projections were used to calculate the % of the Team’s value. The groupings were determined by the simple subtraction of the WR2s % of team value from the WR1s.

These finalized difference numbers were sorted from highest to lowest. The Group 1 WR1s had WR2s whose projections were nearly 20% or greater. Group 2 WR1s had WR2s on their teams that were only different from 11% to 20% The Group 3 WR1s were nearly different by 10% to 3% and Group 4 WR1s were the same as their WR2s or different up to 3%. The actual projections and % calculations are presented in Figures 6 to 9.

## Figures 6 to 9. MFL ADP Based Projections of Team WRs, % of Team WR Totals and WR2s Differences

## 2017 Redraft WR1s Penalties

A summary of the main WR1s from 2017 and what groups they are currently are in is presented in both Figures 10 and 11. Each group lists the players, their PPR projections, % of their team’s WR total and our crowded WRs group assessment along with ADP penalties.

The conclusions from our analysis for redraft 2017 follow at the end of Figure 11.

## Figure 10 and 11: 2017 WR1s by One of 4 Groups with Projections, Team Values, Differences and ADP Penalties.

## Conclusions for 2017.

- All Group 3 and 4 WR1s should be dropped by 5 or so ADP units on average.
- All Group 1 WR1s should be highlighted and considered better than their current ADPs
- Group 2 WR1s are priced right on their ADP average.