By John Bush
Why Late QB Drafting?
Why Late QB Drafting?. If you played in many leagues over the years, you know there are multiple ways to win. Last year in week 16 if you Gurley you won. However, you managed to get to the finals because of one player one game you either won or lost.
The question before us is:
Why Late QB Drafting?
I lead you through the arguments for this aspect of drafting.
Figure 1. 2014 to 2017 Passing Yards for Top 24th QB with 4 Year Average.
The data certainly shows a clear blurring of the top 7 QBs. The data line would be steady and would not decrease in a linear fashion. In fact, the 4200 from the end of the season (EOS) is seen in the 1st QB. Variation is the key.
I take away as a group if I draft early I would take one of the top 7 QBs. However, it is also seen the next 4 who have a media of 4050 average passing yards which is 95% of the 1st QB’s 4252 and 92% of the median passing yards a (4359) average of the top 7 QBs.
Losing 5 to 8% in passing yard value by a pick in the 8th to 11th is thought to be compensated by taking a better RB/WR/TE at the draft position of the early QB.
Can you get that 5-8% of the value back? Obviously yes, if your in-place-of-early QB pick is good. Thus, a low risk shot at the improvement in the team.
Note the 12th to 18th had a median of 3756. This is 86% of value lost waiting. Can you collect 14% or more back? Personal records (see the previous article on planning).
Figure 2. 2014 to 2017 Rushing Yards for Top 24th QB with 4 Year Average.
The rushing data reveals the results for the top QB in that the 2 years of low passing was compensated by high rushing for FP replacement. Can we count on rushing?
Note that the 13th, 15th, 16th, and 17th QBs have had nice rushing. So we have the top QB group high pass and high run combo, followed by a low run and high pass, followed by low pass, high run. I suggest in late QBs we must have a rushing potential. Separate the QB by ADP and last years rushing.
This year late QB targets would be Mariota, Prescott, Smith, Trubisky, Bortles, and Taylor based on their rushing potential. Thus, the opportunity exists for upside if their passing improves if they maintain their rushing. Your risk is only -14%.
The final point is the below the 16th QB might be the line in the sand for late QB drafting. The floor seems to be 27% lost waiting too late. If you are that good, then great!
We now some evidence for Late QB drafting.
Figure 3. 2014 to 2017 % Rushing Yards of Total for Top 24th QB with 4 Year Average.
The table below converts the raw rushing yards into a % of rushing yard contribution to total year. This data explains the idea of including rushing for later QBs. It also gives a view of the 3 flavors of QBs. Early. Mid to Late based on Rush/Passing skills.
Figure 3. Scattergram of 2014 to 2017 % Rushing Yards of Total for Top 24th QBs with 4 Year trends.
Hard to gauge but 2014 QBs were increasing, 2016 decreasing and the other years steady.
Figure 5. 2014 to 2017 Passing vs Rushing Yards for Top 24th QB with 4 Year Average.
Analysis of the visual plot support the conclusions from the last figures.
Total Yards Analysis
The next 3 figures give a view of total yards analysis.
Figure 6. 2014 to 2017 Total Yards for Top 24th QB with 4 Year Average.
The top 7 QBs are still above the field. (breakpoint cluster). The 8 to 11 QB are clusters followed by the 12 to 16QBs. 3 flavors of QBs are still apparent.
TOP 7 QBs median yards is 4604
8 to 11 QBs median yards is 4285 (-7%)
12 to 16 QBs median yards is 4014 (-13%)
17 to 24th QBs median yards is 3371 (-26%)
These losses are mirrored by the passing yards as well (see figure 1 discussion)
Figure 7. Scattergram with Trends for 2014 to 2017 Total Yards for Top 24th QB with 4 Year Average.
Nice clear drop in yardage but the steepness of the line is shallow. Hence, the losses discussed above.
Figure 8. Scattergram with Trends for 2014 to 2017 Average Total Yards for Top 24th QBs.
The plot with error bars for the 4-year averages by EOS QB rank is shown. The analysis reveals the existence of 4 group of QBs. These are highlighted by the red double arrows. The lines of demarcation are clear. They are 7th, 11th, 16th and 17 to 24th QBs.
You could rank within each to including rushing in the latter groups. Also, compare last years data to this year’s rankings (ADPs). What is changing as guessed by the public?
I have found a simple tool for this. Just write one sentence. This QB is at this ADP because and fill in. Now go research that question!
PRE Seasonal (PS) ADP vs EOS of Season (EOS) Average for QBs.
A major confounding factor is the PS ADP is not accurate. If it was then this article would not be needed. ADP vs EOS is skewed and can be 100% relied on. The PS ADP is a current metric of the “wisdom of the crowd’s thinking”.
We must include this difference into the future drafting of QBs.
Figure 9 A and B. Tabular and Graphical Plot of PS ADP vs EOS Yearly and 4 Year Average.
The data of PS ADP vs EOS QB yardage is shown in the table. Note the inconsistency in the 4-year averages. The clear takeaway, the 2016 year was the tops for QB scoring. 2017’s data fell back down. Not consistent!
The 4-year trends show the 2016 high level vs the other 3 years. Note the trend of seeing a barbell formation. High early to QB 4 or so followed by a drop in mid rounds and another peak at the 14th QB. This fits into the previous discussion of 4 QB groups and the low drop in value in the 11th to 16th QBs. These conclusions a major reason for drafting late QB.
EOS 4 Year Total Yards and Average vs EOS Ranking.
Setting up the second viewpoint for a final comparison.
Figure 10 A and B. Tabular and Graphical Plot of EOS Yearly and 4 Year Average of QB Yards.
The EOS data is very clear and a decrease in averages into the lower QB.
The plot is so clean and textbook. Nice linear regressional trends. We do see year to year slop changes but the average is clear with the trend.
Comparision of PS ADP Based EOS Averages vs EOS Averages.
Variation of ADP predictions.
Figure 11 A and B. Tabular and Graphical Plot of EOS Yearly and 4 Year Average of QB Yards.
The tabular view of these two data streams is color coded and it is such that differences are present. These difference are the root of late QB drafting. If the public was perfect then these data would be the same.
The 2 data lines are shown. EOS is red and EOS based on PS ADP is in green. The second peak (discussed before) is easily seen. That is the sweet spot for late QB with strong rushing and 2018 passing improvement upside!
The Difference of EOS vs ADP EOS QB Averages last 4 years.
Figures 13 A, B, and C. Final Comparisons of EOS vs ADP EOS Averages, DIfferences, % Difference, Trendlines and Early vs Late QB (Peak Areas).
I calculated the raw numerical differences and % effect of the ranking differences. The plot is shown below. Note the sweet spot differences are the second peak 12th to 17th or so.
Plotting the FP Differences and % Difference on the 2 different Y- axes, we can conclude the late QB peak has less variation and that is shown by it having the highest peak to almost 0. Remember the axes are negative from 0 (best) vs -50% (worst).
If we data transform and plot the 1 to 11th QBs differences vs the 12th to 21st, we can “see” the sweet spot of QBs which is nearly zero (12th to 17th). Little variation vs a late first group of -35% (QBs 8 to 11th). Valley spot. Early peak one 1 to 7, valley 8 to 11 and peak 2 (sweet spot) 12 to 17th.
Conclusion toWhy Late QB Drafting?
Why does this second peak exist? What elements are we blind to as a public?
I propose the mixture of rushing and passing confuses the public. The public can not separate these later QBs to parse out the passing side. Backward view of last year landscape!
I propose the focus of pundits (regardless of articles/tweets/podcasts/radio shows) to the contrary is less in the late rounds. Data blindness to future change. How good are you?
I propose that average talent in QBs is hard to separate out. The distances are small and hard to see.
The side of the barn vs 2-knot holes in one plank on the barn. Accuracy vs precision arguments! Can you hit the barn with accuracy and at the knot holes with precision? Do pundits with “push it out” have time for System 2 thinking? (see my articles and textbook on thinking in FF).
I will not be a slave to data but I have researched these ideas in the article and have been doing late QBs in best ball formats.
Please visit and read my other articles
Redraft ADP Pattern Analysis 5/9/18 https://www.fakepigskin.com/?p=36403
Non-PPR Player Rankings Part 1 https://www.fakepigskin.com/?p=36513
Non-PPR Rankings Part 2 https://www.fakepigskin.com/?p=36541
PPR Rankings Part 1 https://www.fakepigskin.com/?p=36549
PPR Rankings Part 2 https://www.fakepigskin.com/?p=36554