Targets, Targets Per Game, and Average Draft Position Analysis

Introduction

We need to understand the nature of Fantasy Football drafting in terms of projections, targets, positional differences and player values. In PPR the points per reception aspect force drafters to envision Targets, Targets Per Game, and Average Draft Position Analysis. Targets per game equate opportunity which is the basis of success in your drafts.

Additionally, we all try to master the average draft position by understanding what it means to us and how it may direct us to choose a player. Everyone wishes to draft a player that gives our team more value than we used our draft capital to acquire that player.


Targets, Targets Per Game, and Average Draft Position Analysis

Do player total targets correlate to targets per game metrics? It seems an obvious question and answer.  The answer is yes but how much given injuries and role changes over the season.

In the figure below, I present the graphical plot of the number of 2018 targets vs targets per game. Analysis reveals that a polynomial trend process yields a correlation of 0.88 which is close but not perfect. This plot includes RB, TE, and WRs.

I conclude that all drafters will be subjected to more uncertainty early (steep curve) in this exponential decay distribution of targets vs targets per game.  The level of target per game will eventually hit a level (plateau at 2 to 3 targets per game)  where overall little change is expected for each draft pick. The expected return is lower late in the draft for average player value however, there are clear exceptions as outliers.

Slide1


Positional Specific Targets vs Targets Per Game

Are there positional differences between the RB, TE, and WR in Targets, Targets Per Game, and Average Draft Position Analysis? I plotted the 2018 targets vs targets per game for each position. Overall an exponential decay model explains the data.

I note the differences

  • RB – no RBs were above 8 targets per game
  • RB – the break in the data is annotated a purple arrow at near 3 targets per game. KEY tier break for RBs.
  • TE- only 3 TEs were above 8 targets per game.
  • TE- These 3 are the top 2019 TEs by ADPs.
  • TE has a natural 2 tier breaks below 7 targets per game and 3 targets per game.
  • WR- 6 WRs in 2018 had 10 plus targets per game.
  • WR- break at near 5 targets per game.

As expected WRs are > RBs but seeing the TE population was informative as many are more RB-like in targets per game than WR-like. Factors that influence RB target success may be associated with TEs as well! (More research to be done).

Slide2


Statistical Analysis of Positional Specific Targets vs Targets Per Game

Are the positions targets per game really different? I turned to ANOVA and Tukey Ad-Hoc means-testing to analyze the data. Is the difference seen in the populations real or random? (more samples needed?)

In my textbook (kindle), I use statistical testing when possible. I use to build my confidence in the data conclusions I am noting. I begin by using an analysis of population variance (ANOVA)

Analysis of Variance (ANOVA) is a statistical method used to test differences between two or more means. It may seem odd that the technique is called “Analysis of Variance” rather than “Analysis of Means.”  The name is appropriate because inferences about means are made by analyzing variance.  see link** analysis_of_variance

ANOVA and Tukey Ad-Hoc Means-Testing Results

The right side of the figure below is the ANOVA result using RBs, TEs, and WR data in a grand comparison. The means of Targets Per Game (Light Purple) are 1.99 T/G RBs, 2.28 TEs, and 3.67 WRs. The results from the ANOVA suggest there is at least one difference between the means of these 3 position’s targets per game. (Probability of being wrong is minuscule)

Next, we turn to the Tukey analysis to compare the means. The right side has those results and comparison groups. RB and TE are not different. RB and TEs are different from WRs. Thus we can now with confidence deal with WRs differently from TEs (normal TE leagues). Research using TEs and pass-catching RBs should be consolidated at some levels.

Slide3


Inspired by my stats of RB, TE, and WRs, I  wished to visualize \

  • the population structures of each position,
  • 2018 End of Season Player Tiers, and
  • highlight players that in 2019 are potential values. 

Running Back’s Targets Per Game

Below is the population structure of the 2018 RB targets per game. This box and whisker graph gives us a snapshot of the position. The 5 RBs above 6 targets per game are the super pass-catcher types (yellow dots). The box with x and a line represents about 66% of the RBs which fall between 3 and 0.5 targets per game (edges of box). The line in the box is the 50% percentile for RBs and targets per game with the X being the average at 2 targets per game (see ANOVA)

Slide5


Below is the plot of the 2018 RB population of Total Targets vs Targets per Game. The RBs are fairly predictable from their targets (R-.89). Note also the players and their 2018 tiers to the right. The yellow highlights note players that have high 2019 ADPs. Interesting, group of RB values.

James White is the highest 2019 value based on the 2018 data! He is followed by

  • Cohen
  • Hines, Richards, Yeldon, Riddick
  • Johnson
  • Booker, McCoy, Montgomery and McGuire.

Slide6


Tight Ends Targets Per Game

The population dynamics in TEs is very interesting as it points to a level of extreme players at the top. 5 TEs are from 7.3 to 10 targets per games. The 2019 ADPs support that the public is picking at the top theses players. What about the average TEs? 66% of TEs are at 3 targets and below. Thus the extreme top players are generating 2 to 3X of the average. I use this information to draft in 2019 either the top or I wait and get ready to stream with 2 TEs. (Bar-Bell Drafting).

Slide8


Below is the plot of the 2018 TE population of Total Targets vs Targets per Game. The variation is the highest of all 3 positions (R -0.799). I do note a few deep TEs late in the total targets that still produced near 4 targets per games. (injury in 2018?). Thus some deeper TEs might be valued check on their 2018 metrics. Hence the bar-bell drafting and I look for these 2018 injured TEs to gather 2 of them to stream.

I also present the 2019 value players (see right yellow highlights) within the 2018 targets per game tiers.  Reed and Graham are the highest possible value later TEs! Note the deeper TEs in the targets per games.

Slide9


Wide Receivers Targets Per Game

The population of the 2018 WRs has a broad 66% group. The range is from 6 to 1 Target per game. On average your later WRs are worth more than RB and TE!

KEY- If you can find a sleeper RB or TE late then you have an advantage. If your choices are seemly not sleeper draft picks in RB/TE then your default pick should be a WR! Do not force your picks into uncertain RB/TEs! 

The upper WRs are 6 to 11.5 Targets per Game. Most drafters know this fact thus no advantage. However, I report below on some WRs that seem to be higher in target per game vs their ADPs.

Slide11


The plot of all 2018 WRs targets per game positions are shown by pink circle data points. I note the trend line has a more pronounced liner segment after about 24 WRs (exponential slope). The WRs were shown in the ANOVA/Tukey testing to be different from RB and TE.

The potential WR sleepers are shown to the right in my 2018 Tiers.  Note the yellow highlights of these “sleepers”. Landry is an extreme bargain from the 3rd Tiers. Next, we see 3 WRs in the 6th WR tiers that can be sleepers. (Sanders, Davis, Fitz). The next 7th and 8th Tiers have many WRs of note. Does this mean the public has made more mispricing errors than you make think late? 

Slide12


ADP 2018 vs Players Targets per Game (TGT/G)

I start with an overview of all RBs, TEs, and WRs’ ADP vs their end of season (EOS) targets. Can we predict by ADP a player’s targets per game?  It seems there is little correlation of a player’s pre-season ADP vs their EOS targets per game. Given that R=0.12 is the value, I predict no correlation. What are the metrics players use to set the ADP?

( future research)

Slide14


Below is a box and whisker landscaped plot of 2018 ADP vs TGT/G! The average is near 5 TGT/G while the upper 66% is 7 TGT/G and 99% level is at 11.5 TGT/G.

Slide15


% of Players in 2018 that achieved 6 TGT/G vs Draft Round

If you can pick players that can produce 6 targets/games, those players vs draft round based on ADP, are correlated to draft round. R=0.79 is a nice correlation. See the tabular data that shows 50 % or so or players drafted in the first 3 rounds will produce 6 targets per game.

The early drafts are critical >50%, rounds 3 to 8 have 40% of players and finally, in draft rounds, 9 to 11 we find on 20% of players are going to get 6 TGT/G. Research in the 4 to 11 rounds should be heavier than in the first 3 draft rounds on average. I conclude the uncertainty of our picks is highest in round s 9 to 11 vs 1 to 8 draft rounds. This is the truth of drafts and you must use this knowledge to find those rare players.

Slide16


ADP 2018 vs RB Players Targets per Game (TGT/G)

To begin with, I turn to the RB position. We must tease out the positions to “see” the truth! The trend-line below in looking at RBs by draft pick (1 to 60 picks and R=0.25). Note the almost flat trend-line at pick 12 to 60 RBs.

Next, I then looked at RBs round by round. The extreme RBs that get 4 targets per game are mainly the first 12 RBs followed by 1 RB from then next 12 RBs, 3 in the next 12, 1 in the next and 2 RBs above 4 TGT/G in the final 12 RBs. In PPR the public “sees” the pass-catchers and hence the nature of the trend-line.

Slide18


RBs above 4 TGT/G vs RB groupings by 12s.

The plot below cleans the noise and gives us the reality of the top RBs spread across 60 RBs by ADP. It’s all about the top 12 RBs. Your research needs to focus on finding superior RBs in the next 48 RBs. In 2018, in those 48 RBs, only 7 RBs achieve a nice 4 TGT/G!  The idea that it will be easy to draft those sleepers is an illusion. I predict drafting handcuff types that can be elite.

However, look at the top 12 RBs and consider which ones have injury issues and who might be draftable RB handcuff! *

2019 PLAYER POS TEAM Public Handcuff Public Handcuff
1 Saquon Barkley RB NYG Wayne Gallman Rod Smith
2 Alvin Kamara RB NO Latavius Murray
3 Christian McCaffrey RB CAR Cameron Artis-Payne Jordan Scarlett
4 Ezekiel Elliott RB DAL Tony Pollard Darius Johnson
5 David Johnson RB ARI Chase Edmonds
6 LeVeon Bell RB NYJ Ty Montgomery Elijah McGuire
7 James Conner RB PIT Jaylen Samuels
8 Todd Gurley RB LAR Darrell Henderson
9 Joe Mixon RB CIN Giovani Bernard
10 Dalvin Cook RB MIN Alexander Mattison
11 Nick Chubb RB CLE Kareem Hunt
Dontrell Hilliard
12 Damien Williams RB KC Carlos Hyde Darwin Thompson

*Given evidence that individuals high in intolerance of uncertainty (IU; the tendency to experience unknown outcomes as unacceptably threatening) exhibit poorer decision-making strategies and are more behaviorally inhibited in unpredictable situations than those low in IU. Intolerance of Uncertainty link

You must learn to be low to intolerance of uncertainty. Embrace uncertainty in your research on Targets, Targets Per Game, and Average Draft Position Analysis

Slide19


RBs from 2018 vs ADP and TGT/G

Note the green annotation of those players that were above 4 TGT/G!  Note the wasteland and the key RBs above 4 TGT/G in the later rounds.

  • Drake and Lewis
  • Thompson and Cohen
  • White
  • Bernard,  Conner, and Riddick

Slide20


ADP 2018 vs TE Players Targets per Game (TGT/G)

In the TE position, we have fewer viable players in typical leagues (not TE premium). The plot below looks across at 2018 pre-season ADP vs end of season TGT/G, The trend-line only gives us an answer of 50% of the variation in the data (R=0.5).

Interestingly, the data structure looks flat with a few exceptions. The TEs were near 5 TGT/G on average across the population of 18 TEs. This can suggest late TE drafting as you will get the same results. Therefore, I have a bar-bell approach to drafting. I draft the top few or I wait!

Slide22


TEs from 2018 vs ADP, TGT/G, Scaled to TGT/G Average

The table below is sorted from the 2018 ADP top to bottom of the first 18 TEs. Those players whose TGT/G were above 6 were noted by green highlights. These players are listed below along with their 2019 ADP.

Furthermore, in my drafting I take one of the first 3 or wait while watching for  Ebron/Walker/Reed. The public has injury concerns for Walker and Reed and TE committee issues for Ebron with Doyle back.

  • Kelce – 2019 1st
  • Ertz  – 2019 2nd
  • Kittle – 2019 3rd
  • Ebron – 2019 8th
  • Walker – 2019 13th
  • Reed – 2019 18th

Slide23

I also have scaled the 2018 TGT/G metric to the league average and subtracted to get the scaled to average number. The plot above suggested a “flatness” to the TE population. These numbers were then subjected to a Box and Whisker plot analysis (below).

The 0 point is the league average. 66% of the TEs were found near the zero points (flat) and just a few TEs were anywhere close to a significant TGT/G number above the other TEs (top of the figure near 3.9 TGT/T!

This analysis has driven my TE drafting of Early vs Late TEs.

Slide24


ADP 2018 vs WR Players Targets per Game (TGT/G)

The plot below includes the WRs from 2018 vs their ADP (WR 1 to 65). The R-value is 0.59 and that is double the correlation of RBs! RBs have more injuries and inconsistency vs WRs and comparing the two positions will establish that.

WRs have the highest TGT/G within their populations. The ANOVA analysis has established this position is different from RBs and TEs. 6 WRs were at 10 TGT/G in 2018! No RB or TE was above this level. 2 TEs were close though.

Slide26


WR Success Level Landscapes from 2018 ADP vs Success

Landscape analysis led to the question: What was the success level within drafted WRs by 12s? The levels of interest to me were 8 TGT/G and 6 TGT/G for WRs. I calculated the % of the WRs within the 12s segments covering 60 plus WRs.

The table below presents the numbers of 2018 WRs above 8 TGT/G and 6 TGT/G followed by the % of successful WRs within the 5 12s segments.  Analysis of the 8 TGT/G level reveals that in 2018 all of the top 12 WRs averaged 8 TGT/G for 100%, followed by an extreme drop in WR performance of 25% in the 2nd 12 WRs, 17% of WR in draft segment 3, 8% WRs in the 4 segments and no WRs in the 5th WRs segment generated 8 TGT/G.

Dropping the standard to 6 TGT/G per game we see 100%, 58%, 83%, 41%, and 33% over the 60 WRs. I list the 2 levels for comparison as well for further thinking. You must work harder the deeper into the draft. Debating the 1st vs 5th WRs is not as important as finding the 4 of 12 WRs in the 48 to 60th WRs that will return 6 TGT/G! Now is the time to get this done.

Slide27


A plot of the WR 2 levels of Success 8 vs 6 TGT/G vs ADP Segments.

The take-home is deeper into the WRs the murkier the choices become. There will be metrics pointing in opposite directions. Using team level analysis try to find which WRs are going to get the opportunities?

Does the team have a record of supporting 3 WRs at nice levels (LAR 2018) or just one WR1 at a high level (NO 2018)? Gather the historical data vs current 2019 ADP thinking. Keep this % in mind. If you are seeing 8 WRs in the 48 to 60 level that seem good that is probably not true! Look for high target numbers/ high impact offensives/ weaker defenses leading to more passing.

Slide28


WRs from 2018 vs ADP, TGT/G, and Scaled Difference from League Average (6.7 TGT/G)

The table below can serve as a historical reference for the research suggested above. I have included the TGT/G (Above 8 – purple and 6 to 8 -green), Draft Segment 1 to 6, and Scaled Difference in TGT/G vs 6.7 TGT/G league 2018 average (Blue to red annotation).

Key players in the Later ADP Segments looking at Targets, Targets Per Game, and Average Draft Position Analysis

  • Sanders
  • Edelman
  • Woods
  • Golladay
  • Miller
  • Allison
  • Westbrooke

Slide31

Slide32


WR Scaled Differences TGT/G vs ADP

I plotted the scaled differences metric and it can be seen below. The league average would be at 0 and players above or below are seen within the WR segments. Note by the 48th WRs, all but 1 WR were below the average of zero.

Continuing with that theme, I used a blue circle to surround the WR outliers from segments 3 to 5. These players were shown in the table to the right of the plot graph. I added the WR 2nd segment outliers as well. Thus we have 7 WRs after the first 12 that can be labeled as extremes as compared to their ADP.

Slide29


2019 Predicted WR Outliers by TGT/G Analysis.

The important aspect of the above data was that the outliers all were their Team’s WR1 at the end of the 2018 season. Thus, I wanted to gather the 2019 ADP for WRs and highlight players predicted to be the Team’s WR1 coming from the 13th WR outward.

I listed them interesting players in the table below and have highlighted the current 2019 ADPs.  Research all of these to find the outliers (sleepers) relevant to their ADP. Draft them often!

Slide30


Targets, Targets Per Game, and Average Draft Position Analysis

Fun Research in my textbook!

If you liked a metric-driven view of Fantasy Football check out my 2000 page textbook!

Winning Your Fantasy Football Drafts: A Comprehensive Textbook: June 2019 Edition [Print Replica] Kindle Edition

Click Link Below

Textbook on Kindle


Please Read

Part 1 ADP to FP/G QBs and RBs

Part 2 WR and TE ADPs to FP/G

AFC ADP to FP

NFC ADP to FP


My current PPR rankings

ppr-power-rankings-part-1/

ppr-power-rankings-part-2/



Defense Against the Positions Part 1, 2, 3, 4, and 5

DAPs Part 1 Link

DAPs Part 2 Link

Analysis of DAP Part 3

DAP Best and Worst Players Part 4

DAP Analysis 2019 Playoffs and Bye Week


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: