ADP vs Player Performance Part 3

Slide4

By John Bush

ADP vs Player Performance Part 3

I circle around into the RB vs WR debate in ADP vs Player Performance Part 3. We present the sector data for all positions with all years combined. We colorized the entire table and the highs and lows are apparent!

Please see

ADP vs Player Performance Part 1 http://www.fakepigskin.com/?p=36752

ADP vs Player Performance Part 2 http://www.fakepigskin.com/?p=36771


Figure 6-14 A and B. Preseason Draft Values of all Positions.

Note to begin in the Top 6 sector (Top 6 means the first 6 players drafted). You had RBs and WRs only over 2013 to 2017. The RBs had 75 vs WRs 95 in value. The take-home was the variation in the RB starts out high right in the first 6 picks as shown in Part A.

QBs

Nice QBs last into the 42nd draft pick. Note that historically we had nice showings of QBs in draft picks 61 to 72 as well (85 value). These two sectors have given rise to the late QB drafting.

TE

Note here that from draft pick 7th to 30th, those TEs have done very well. Also, scan the remain TEs and we see a sameness of value down into late TEs in draft picks to 126th. These value data map my “barbell” approach of TE drafting. Take top 1 2 or so or wait late! Forget the mid-TEs as they do not deliver enough value.

RBs

Top 6 stand alone in a tier (75 value), next we go into the draft round of 42nd (Median value of 59 ish), next tier starting at 43th to 66th RBs (45 value), the fourth tier as defined by draft picks begins at pick 67 to 132th (median of 38 ) and then the last tier for RBs from 133th to 168th is a median of 29. In truth then the actual data drive tiers for RB vs most pundits projections/made up tiers is clearly 5 basic tiers.

Try in your mocks to use 5 tiers based on data.

  • Tier 1 RBs 100% value     1st to 6th
  • Tier 2 RBs 78% value       7th to 42nd
  • Tier 3 RBs 60% value       43rd to 66th
  • Tier 4 RBs 50% value       67th to 132th
  • Tier 5 RBs 37% value       133rd to 168th

WRs

The top WRs if in the first 6 picks are a treasure (Brown in 2018) and they return 95 % value in tier 1 WRs. Tier 2 WR go from 7th draft pick to 54th returning a median of 67% value. Tier 3 WRs begin at 55th pick into 84th pick having a median of 55.3% value. Tier 4 WRs begin at 85th pick to the 108th returning a median of 46% value. Tier 5 WRs start at 109 draft pick to 168th and give back 38% value.

Try these in your mocks

  • WRs Tier 1 100%  value          1st to 6th
  • WRs Tier 2  70%   value           7th to 54th
  • WRs Tier 3  58%   value           55th to 84th
  • WRs Tier 4  48%    value          85th to 108th
  • WRs Tier 5  40%    value          109th to 168th

ADP vs Player Performance Part 3 Slide28


In Part B, the trendlines show a distinction between the RB. WR are higher in the average of EOS PF into pick 168. Another observation is after the 90th pick the value gets choppy. We see that RBs at that late coming closer to WR! We see the distance between those positions decreases across the landscape.


 Slide29


Running Backs

Figure 6-15 A, B, and C. Running Back Positional ADP by Year vs EOS PF

Part A focuses on the first 24 RBs picked from 2013 to 2017. Analysis of the tabular data, it’s very clear of the high failures within the top 6 and 12 RB. RB was the TOP ADP Pick in all except 2016! 2 were at 57 and 2 at 78 at the EOS! That is not what you need in production for a 1st ADP pick! Red Flags for 2018 and extra research is suggested! See the failure rates data in the preceding figures.

The trendlines give a more complex view with 2 better years 2014 and 2017 vs the rest of the years (2013, 2015 and 2016) with a similar trendline. The 2014/2017 seasons had almost a straight line of EOS RB PF numbers. Thus in those years, all drafters were successful vs the other 3 years where we see a decrease across the 24 draft picks and thus drafts would be less successful. 2018 can be a repeat of 2014/2017 or not? Caution is suggested.

In a broader view into the 24 ADP picks only 13 RB were at 80 in their end of Season Values and above vs 48 RBs were below the EOS 80 PF metric. So, the odds are at 20% to catch a good RB (above 80).

 ADP vs Player Performance Part 3 Slide30


In Part B we investigate the draft picks 25th to 96th for RBs. The RBs from this draft sector were only at the 15% level of who was above 80 EOS PF average.

Finally looking at Part C,  97th position and out, only 3 RBs out of 184 RBs drafted were above the 80% level at the EOS!  Only 20 RBs out of 184 were above 50 EOS as well.

Do not expect miracles after the 48th RB! Keep this in mind for 2018 drafts!

ADP vs Player Performance Part 3 Slide31


ADP vs Player Performance Part 3 Slide32


Reference Class Forecasting

To get useful data in the range of outcomes for RBs, I determined the 5-year highs and lows at each ADP pick. What is the range of possibilities in your 2018 draft?

This approach is a method called “reference class forecasting,” which achieves accuracy by basing forecasts on actual performance in a reference class of comparable projects and thereby bypassing both optimism bias and strategic misrepresentation. People are subjected to being overly optimism**.

Psychological explanations account for our inaccuracy in terms of optimism bias, that is, a cognitive predisposition found with most people to judge future events in a more positive light than is warranted by actual experience. **

** From Nobel Prize to Project Management: Getting Risks Right. Bent Flyvbjerg, Ph.D. & Dr. Techn. Aalborg University, Denmark. Project Management Journal, vol. 37, no. 3, August 2006, pp. 5-15

We want to give a reference forecast for last year’s drafts within the different positions. See cognitive biases and solutions (Fantasy Football Planning  http://www.fakepigskin.com/?p=36562 )  and the above reference.


Figure 6-16 A and B. 5-Year UL and LL levels of RB EOS PF Average from Picks 1st to 48th

Part A gives us a look into the first 24 RBs drafted vs their EOS PF at both the upper limits (UL)  and lower levels (LL). The UL has 7 EOS PF below 70% EOS PF. Even if you draft the best RB at ADP position each year you still would have had 30% fails!

Thus, the best-case scenario in your draft of the first 24 RB you can only expect 70% success! That is, your upper reference class figure!

In the worst cases, you can only expect 3 RB above 80% at a 12.5% success for your Lower reference class. The range is 70 to 12% success on average for RBs! I assure you that your great “feeling”  in RB drafting early in drafts is not on average accurate! Dial that feeling back and use caution to do extra research for your 2018 drafts.

ADP vs Player Performance Part 3 Slide33


Part B includes the data for RBs down to the 48TH RB. The trendlines reveal the distance (RISK!) goes down as you approach the 48th draft pick! We see the bottom or floor is close to 40% EOS PF value at 46th level and the upper level is at 50%.

Thus your 2018 RB drafts must juggle the early high reference classes vs the lower RB drafts. The difference is more critical early vs late! The later differences suggest your drafting for later RBs have more room for error vs the early RBs with their high UL vs LL differences.

ADP vs Player Performance Part 3 Slide34


 

Wide Receivers

Figure 6-17 A, B, and C. WR Positional ADP by Year vs EOS PF

 In a landscape view (see 6-17 Part A and B) looking and analysis of the 24th ADP picks, I found that only 8 WR were 80 and above vs 3 RB (Figure 6-15 A). So, the odds are at nearly 3x greater to catch a good WR in the 5 years average. That is a great selection % vs the RBs. You have a 2.7X greater chance to get a nice WR in the first 24th picks!

The yearly trendlines are also giving us some valuable insights. Firstly, the 2014 yearly data was very different vs the other 4 years in WRs. The orange trendline is increasing across the 24th picks. The EOS PF for WRs was going up and that year drafters were finding high success in the 2nd WR ADP based round players.

The other four years for WRs across the first 24 ADP picks had expected trendline dropping all below 80 but staying above 60 EOS PF. That is very consistent and leads to the WR safety concepts driving ZERO RB drafting. Note that 2017 was a strong year for WRs as was 2013. Those trendlines are starting near the 100 EOS PF level and dropping. That is compared to 2016 and 2015 whose starting WR EOS PF was closer to 90.

Takehome, 2017 had a great RB and now WR run that was not typical of the preceding years. Few if any pundits are focused that. That 5-year pattern screams caution to me in early rounds for 2018!

ADP vs Player Performance Part 3 Slide35


 16-17 Part B I investigate the 25th to 96th.

WRs for 25th to 48th only were at 8/56 or 14% above 80 EOS PF average. That compares to 15% for RB! We see a very equal success rate in drafting both RB WRs in rounds 3 to 8. That moves us to consider either RBs vs WRs in that draft level!

ADP vs Player Performance Part 3 Slide36


Finally looking at Part B and C, only 10 WR out of 193 RBs drafted were above the 80% level! RBs were at 4 players.

Do not expect miracles after the 48th WR! Keep this in mind for 2018 drafts but in the WR you have a 2.5X higher chance of getting a nice WR. The rate is still low at 5%!

ADP vs Player Performance Part 3 Slide37


Figure 6-18 A and B. 5 Year UL and LL levels of WR EOS PF Average from Picks 1st to 48th

The reference class limits for WRs is shown for the first 48th draft picks. If you check the RB data, you remember the high range was from best to worst case situations. We see in here that the range between the 2 UL and LL is much closer together. The floor is closer to the ceiling vs the RB position. Thus, you have a great chance for player success with your WR picks within the top 24th picks! (Zero RB again supportive).

In the first 6 drafts picks, any WR in that zone has an 11/12 shot at being above 85%. Nice!

Note the 7th draft pick to 24th, it is quite different as the WR floor drops as only 2/13 WR LL are above 80%. The risk has arrived in this range and for 2018 caution is suggested!


ADP vs Player Performance Part 3 Slide38


In the Part B plot, we went out to the 48th pick for WRs to see the reference classes for top and bottoms. The trendlines reveal the distance (RISK!) increases as you approach the 48th draft pick! We see the bottom or floor is close to 40% EOS PF value at 46th level and the upper level is at 65%. Compared the RBs which are 40 to 50 tops to bottom. Thus, your chances near the 48th pick are 15% better in the WRs pool vs RBs.

ADP vs Player Performance Part 3 Slide39

 


Figure 6-19 A, B and C. Reference Classes for RB vs WRs Upper and Lower Limits to the 48th Draft Pick.

We used the two positional UL and LL limit data to directly compare the two positions in Figure 6-19.

Part A gives the landscape view by the trendlines with red is WR and green is the RB. We see the top and bottoms into the 48th draft picks. The take-home is the RBs UL and LL come together as the WRs limits become closer. The distance can be thought of as a risk.

RBs risker early and WRs risker later!


ADP vs Player Performance Part 3 Slide40

In the first 24th picks the best case, ULs, between the RB and WR is almost dead on. The RB drops to worst case is very severe! Conservative approach leans to WRs over RBs.


ADP vs Player Performance Part 3 Slide41

.In Part B we show the upper limits or best-case situations. The data goes from equal early to about 10% value difference of WR over the RB position. If you are the best drafter, then these limits do not apply to you but for the rest of us, these set the reference class upper limits in our 2018 draft plans!


 ADP vs Player Performance Part 3 Slide42

In Part C, we have the worst-case limits in WR vs RB drafting into the 48th pick. The RBs floor decreases into the WR floor. It seems that close to the 24th pick the distance between them is too close to call. The levels of risk are the same by the 48th pick or so!

Note the worst case for RBs is almost a straight line into the 48th pick. Your floor is fairly solid but at a lower level vs WRs. WRs decrease and the floor lowers significantly going into round 3 and 4.


Quarterbacks

Figure 6-20 Part A and B. Quarterbacks EOS PF vs ADP Pick from 2013 to 2017

I investigated the QB position down to the 84 ADP pick and that tabular data from 2013 to 2017 is shown along with a Scattergram with trends. The trends for 2016 vs 2017 were shown below and analysis suggest these were 2 difference type of years and that may be important to 2018.

In 2016 the trend increases and drafters were very successful with later QBs. In 2017, however, drafter had lesser success as the trends decreased downward. It’s a small note of late QB drafting caution here but I elect to continue with that approach in the face of these data bump in the road.

ADP vs Player Performance Part 3 Slide43

In Part A, we plotted the ADP picks from 2013 to 2017 QB selected. Looking into the summary data we spot that first 11 picks only 1/12 QB drafted ended below 80% EOS PF level. QBs are a very safe play on average.

2016 vs 2017 trendlines were very different for an ascending pattern vs a slight decrease in scoring. These data as compared to the other positions was much flatter across the ADP picks. Late QB drafting is supported by the data but caution is suggested!


ADP vs Player Performance Part 3 Slide44

In Part B, we lined up the 5 years of QB picks and the average EOS PF. The table illustrates that 1/10 in QB picks in the top 2 QB have a 90% EOS success. The first 30 picks or down to the 6th QB each year, the failure rate (Below 80) is 13/35 or 33% failure.

You have a 66% chance in the top 6 QBs in landing a very nice QB. In the next 30 picks then 23/30 QBs failed! 

The 6th pick for QB is the breakpoint for your 2018 Drafts. If you go past that as I do, I grab multiples QBs for a streaming approach! 


Tight Ends

Figure 6-21 TE EOS PF vs ADPs 2013 to 2017

The trendlines of the 5 years show that 2016 was a different year vs the other 4 years! We assume in 2018 the trendlines will match the 4 years vs the outlier 2016 data. What are the failure rates? The top 12 picks of TE we see 3/12 failures (25% failures). The next 20 we have 16/20 failures. The breakpoint for 2018 is the 40th ADP pick or before. After the 40 ADP pick the failure rate is high. I use late TEs in streaming fashion. 

ADP vs Player Performance Part 3 Slide45

In review

  1. Read all three sections
  2. Write the conclusions that seem to make sense to you.
  3. Plan mocks to test
  4. Do your drafting and keep a diary to backcheck next Jan or so. 
  5. Caution and slow thinking are better than quick reactionary picking.

Other Articles of Interest

More Information Increasing Accuracy?  http://www.fakepigskin.com/?p=35319

Redraft ADP Pattern Analysis 5/9/18   http://www.fakepigskin.com/?p=36403

Non-PPR Player Rankings Part 1  http://www.fakepigskin.com/?p=36513

Non-PPR Rankings Part 2  http://www.fakepigskin.com/?p=36541

PPR Rankings Part 1   http://www.fakepigskin.com/?p=36549

PPR Rankings Part 2   http://www.fakepigskin.com/?p=36554

Fantasy Football Planning  http://www.fakepigskin.com/?p=36562

Why Late QB Drafting?  http://www.fakepigskin.com/?p=36601

Best at Best Ball Part 1   http://www.fakepigskin.com/?p=36666

Best at Best Ball Part 2   http://www.fakepigskin.com/?p=36677

Best at Best Ball Part 3   http://www.fakepigskin.com/?p=36708

Best at Best Ball Part 4   http://www.fakepigskin.com/?p=36726

Best at Best Ball Part 5   http://www.fakepigskin.com/?p=36734

FF Players Biased Decisions Part 1 http://www.fakepigskin.com/?p=36738

FF Players Biased Decisions Part 2  http://www.fakepigskin.com/?p=36743

ADP vs Player Performance Part 1 http://www.fakepigskin.com/?p=36752

ADP vs Player Performance Part 2 http://www.fakepigskin.com/?p=36771

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>