- Week 3 IDP Rankings: Sack Heavy
- Week 2 Waiver Wire: It’s Chris Carson Time!!!
- Angle of Pursuit: Fantasy Football Week 2 Review
- NOT IDEAL: Mike Evans vs. Xavier Rhodes
- Fantasy Football: TE Injury Replacements
- NFL Week 2 Picks Against the Spread
- Monday Night Risk Reward Week 2
- Sunday Night Risk Reward Week 2
- Sunday Afternoon Risk Rewards Week 2
- Fantasy Football: Fix Your Brain

# Missing Games Affects WR 2016 Part 2

**By John Bush**

This is a continuation of the previous article Missing Games Affects WR 2016 Part 1 (Missing-Games-…wr-2016-part-1). Please go to that first article before going into this part 2 article.

## Missing Games vs Seasonal Performance Scoring

In the next figures 1 to 3, I have grouped all the WRs by Games Played and highlighted their seasonal performance scoring averages. If you scan across the data, it became clear that the games missing cost does exists.

We will used these group’s data to test the main question. The question to be addressed is: **What is that average cost by missing games and are any differences between groups statistically relevant? **

This exploration should lead to the nature of penalties that 2017 drafter can apply to their projections. I suggest adding a missing game component to your pre-draft research and use this article are a way to formulate penalties to any players you think will miss time in 2017.

## Figure 1 to 3. 2016 Grouped WRs by Games Played

The next figure, Figure 4, presents a nice landscape view of the entire 2016 WR data by games played and numbers of WRs in performance scoring groups. Analysis of this data clearly shows the nature of the missing game cost. Note the pattern of group numbers as denoted by colorization. Follow the deep green colors downward to see the numbers movement to the lower scoring columns as you miss more and more seasonal games.

**Preliminary Conclusions**

**One half of the WRs that did not miss a game in 2016 were in the 40 to 60 performance average scoring group****Missing one game has a slight cost with a majority of players being seen in 30 to 39 performance average scoring group****Missing just 2 to 3 games will cost. A majority of those players are in the 20 to 29 performance average scoring group.****Missing half of the season and the highest number of those players are in the 10 to 19 performance average scoring group**

## Figure 4 Summary of Players by Games Played and Performance Scoring Groups

I now set up the statistics side of this article. Using the data and group averages shown in Figure 4 and plotted in Figure 5, I begin the analysis.

## Figure 5. Data Sets for Each WRs Scoring Groups to be Statistically Tested.

## Figure 6. 2016 Average Player Performance By Games Played

The data in Figure 6, highlighted the decreasing scoring averages for WRs as the numbers of games played decreases. That number starts at 40, hits 32.5 with 15 games played and hits bottom at 13.8 in WR that only played 4 to 6 games in 2016. Interestingly the WR average of WRs that played 1 to 3 games was at 15.1.

**Are these averages really different? Are these differences between the WR groups from random chance or not? **

In previous articles, I have glossed over the stats parts of my data analysis but wished to illustrate that process here.

The overall process is as follows.

1) Question

2) Sample

3) Group

4) Determine the Means (group averages)

**Steps 1 to 4 have been done and illustrated in Part 1 and in Figures 1 to 6 **

5) Do ANOVA Testing – Analysis of variance (*ANOVA*) is a collection of statistical models used to analyze the differences among group means and their associated procedures (such as “variation” among and between groups) * https://en.wikipedia.org/wiki/Analysis_of_variance

6) Set a probability level (usually p>0.05 or 5% chance or higher of being random) to determine if the Group Averages resulted from randomness.

7) Thus if the ANOVA P value is greater than 5% then we stop and move to another testable question, The Groups are coded A to G but represent those shown in Figures 5 and 6.

8) The ANOVA P value (Figure 8) is 0.0000000000000001. We can interpret this to mean that the chance of our group difference being by random chance is pretty much non-existent

## Figure 7 Group Means plus ANOVA Specific Metrics

## Figure 8. Generated ANOVA P Value

**** Tukey’s range test**, is a single-step multiple comparison procedure and statistical test. It can be used on raw data or in conjunction with an ANOVA (post-hoc analysis) to find means that are significantly different from each other. Named after John Tukey,^{[2]} it compares all possible pairs of means, and is based on a studentized range distribution (*q*) (this distribution is similar to the distribution of *t* from the *t*-test. See below).^{[3 }

^{** https://en.wikipedia.org/wiki/Tukey%27s_range_test}

9) Since the ANOVA P value was “good”, I next applied the TUKEY’s test to compare all groups means to each other.

10) P Values were set at p

11) All Significant Between Group Differences were determined and an inference of significance (green) or insignificance (red) was calculated.

## Figures 9 to 11. Tukey Test Results for all WRs in 2016 Games Played Groups.

I leave the scientific weeds to get you a take-home message for your 2017 drafts. The conclusions are shown in Figure 12. The Groups are listed at the top along with their PF scoring averages. Next below the groups are the statistical inferences.

Missing 0 to 1 game should be considered to have no effects on WRs PF scoring. However, beyond missing 1 game, there are costs in missing more than 1 game. Those costs are illustrated at the bottom of Figure 12

Missing 2 to 3 games apply a minus 33% penalty all the way to missing 10 or more a nd apply a 63% or more penalty to those WRs.

Thus if Zeke is sanctioned for 6 games then he should be on average 44% less of value to you in your projections and ADP levels. Currently, the Zeke penalty is not that large by ADP. Remember this data is on the average and ZEKE could rise above that level but that is the his average floor!