Following the theme of data analysis and statistics, today we will be looking at the NFL Draft. Rather than analysing the 2019 Draft, we will be looking at a few questions surrounding the draft in general, and see if we can answer them by analysing historical data:
- What was the Greatest Draft of All Time*?
- What was the Worst Draft of All Time*?
- Do high draft picks mean more success?
- Are there certain positions more likely to contribute to a successful draft/unsuccessful draft?
No doubt as we go through the data other questions will arise! All of this data is from the 2003 Draft onwards, as that’s when the 32 team NFL had it’s first 32 team draft.
How to measure Greatness?
For this analysis, we will be looking at a few variables:
Average Draft Pick (ADP): The teams average draft pick for that year. For example, if a team had Picks 1, 33, 67, 89, 104, 143, 167, then their ADP would be (1+33+67+104+143+167)/7 = 73.6. Ideally, a smaller ADP = the team had higher draft picks.
ADP Rank: The Rank of the ADP, in all drafts since 2003.
Win% Last Year: The team’s win % in the year preceding that draft.
Win% Next Year: The team’s win % in the year following that draft.
% Change to next year: The difference between Win% Last Year and Win% Next Year
Avg Win % Next 5 Years: The Teams Avg. Win % over the 5 years following the draft.
5 Year % Change from Prev. 3 Year: The difference between Win% over the previous 3 years, and Avg. Win% Next 5 Years.
Looking at Average Draft Pick
Common sense tells us that teams who pick higher will get better players, and therefore increase their Win % for the next year. When we run a correlation test, we get the following:
What this graph tells us is that as ADP increases (teams have lower draft picks), the % difference between last year and the next year decreases (teams get worse). As we can see by the slope of the line, it is only a slight correlation, so we cannot say conclusively from this data alone that it is worth trading down if you want to win more games the next year, but it is a piece of the puzzle.
The Greatest Drafts of All Time
If we took ADP at face value, we would expect the greatest drafts of all time to have the lowest ADP, and the worst drafts to have the highest ADP. This is not the case, as we can see below:
Both Green Bay and Washington got worse the following year, whereas New England got 21% better, and went unbeaten until the Superbowl. New England are the most difficult team to assess in the Draft, as Belichick consistently drafts low, and produces winning seasons out of relatively no-name players. Due to this, New England are somewhat of an outlier in this analysis.
If we were to look the highest rank ADP, we get the following:
Denver got worse even though they had the 2nd highest ADP since 2003, and Cleveland and NYG both improving by similar amounts. Both these teams were coming off poor seasons (CLE = 25%, MYG = 18.80%), so you would expect an increase due to regression to the mean, and the players taken in the draft include Johnny Manziel, Saquon Barkley, Justin Gilbert, and Joel Bitonio, which in my opinion doesn’t make a terrible draft, but not an incredible, team changing draft (although for NYG it might be too early to assess this).
If we can’t look at ADP as a reliable indicator for greatness, what else should we use? How about if we rank teams by the amount they improved in the year following the draft? By ranking by % Change to next year, the Top 10 are:
Looking at these actual drafts, it is clear to see that the draft is not the be-all and end-all of success. Let’s look at the Draft before the greatest turnaround in NFL History, the Miami Dolphins in 2008:
Jake Long was an All-Rookie and made the Pro Bowl a few years later. Phillip Merling had some contribution, notably a pick-6 in Week 17. Chad Henne got limited playing time, and the rest had small contributions as well. So what made the Dolphins turn their franchise around? Probably several things, including a new VP Operations (Bill Parcells), GM (Jeff Ireland), and HC (Tony Sparano), and the introduction of the Wildcat Offense, among other factors.
Going through the list of draftees from these 10 Drafts, we see that many of the teams held the #1 Pick, or at least their first pick was high. This goes against the ADP correlation with Win% Increase. Perhaps this is because the #1 Pick is so valuable? The players first off the board in these 10 Drafts for these teams were Jake Long (1), Andrew Luck (1), Eric Fisher (1), Ezekiel Elliott (4), Ryan Shazier (15), Baker Mayfield (1), Eli Manning* (1), Leonard Fournette (4), Jadeveon Clowney (1), Gerald McCoy (3), all of which are (arguably) franchise altering players.
*Didn’t play for the Chargers, got traded on draft day for Philip Rivers (4).
If we rank the teams by %Win Change as we have done, it seems a recipe for success is to hold a high 1st round draft pick, and then mid-range picks for the rest of the draft. The exception to this is the 2018 Cleveland Browns, who had the 8th Highest ADP since 2002, and a 46.7% Win% Increase, however this was from a 0% Win season. My personal opinion is that out of these 10 Drafts, the Colts, Cowboys, and Browns drafts are the Top 3, as the Colts got a franchise QB and 3 other quality players (Luck, Fleener, Allen, and TY Hilton), the Cowboys got a franchise RB, LB, and QB (Elliott, Smith, and Prescott), and the Browns got a franchise QB, CB, RB and a quality G and WR (Mayfield, Ward, Chubb, Corbett, and Callaway). Feel free to disagree!
The Worst Draft of All Time
Using the same % Change in Win % measurable, let’s look at the worst ranked teams, i.e. the teams who got worse by the largest % following the draft:
Let’s look at (by this measurable), the worst draft of all time, the 2013 Houston Texans. This also confirms that even if a team drafts well, they can still play badly due to coaching/scheme/a million other variables:
In my opinion, this is a good draft. Hopkins is the best WR in the NFL (coming from a Texans fan!), DJ Swearinger was solid, but perhaps didn’t match up to a 2nd Round Pick, Quessenberry unfortunately was diagnosed with non-Hodgkins T Lymphoblastic Lymphoma but was a contributer, and Ryan Griffen has contributed much more than most 201st overall picks. So why did they regress from a 12-4 to 2-14 record? Gary Kubiak was fired after Week 14 for his performance, and the trio of Schaub/Yates/Keenum didn’t play well, resulting in Schaub being traded and Keenum being waived before the start of the next season. This perhaps indicates the importance of solid QB play to being a successful team!
Again, looking through the picks of these teams, the majority of first picks are in the lower 20s. Lets compare the average pick for each round between the Top 10 and Bottom 10 Drafts:
This table shows us that the ‘Good’ drafts (where the team improves the most the next year) have a much higher 1st, 2nd, and 3rd pick, and slightly higher 4th – 6th pick. We should somewhat expect this, as with the way the NFL works, it is hard for a team to be consistently bad, so if an average team has a bad season, and gets high picks, they are going to be an above average team with the additions they have made.
However, is the metric ‘% change to next year’ really a good metric to use? How many rookies have an immediate impact their rookie year? 1st round picks, perhaps, but many must wait until years 2, 3, or 4 until they can impact the team. In the next section we will look to see how drafts match up when we rank by % increase in the 5 years following the draft.
5 Years after the Draft – Good and Bad
Before we start this section, it is important to note only drafts from 2013 and earlier can be judged in this, which eliminates some of the drafts we looked at in the previous section. We are comparing to the previous 3 years here, instead of 1 year, to try and exclude good teams who have had one bad year.
Let’s look at the ‘best’ draft by this measurable, the Seattle Seahawks in 2011.
This was arguably the start of the transcendent Seahawks defense, with Sherman, Wright, and Maxwell being drafted here, and looking back, Sherman at #154 is an absolute steal. It’s important to note that Russell Wilson was taken the following year, which no doubt contributed to % increase in the next 5 years. Compared to when we rank teams by win% in the year immediately following the draft, the 1st selection in this draft is #25, instead of in the Top 10.
This draft doesn’t tick many of the boxes of the good drafts when we rank by % increase immediately after the draft, for example, the first 3 picks being a tackle, guard, and linebacker, with only 1 skill position taken and no QB. This suggests to me that you can’t turn a team around with 1 single draft, rather, it takes a couple of years at least to fill gaps, and turn into a great team (confirmed when Seattle took Wilson the following year).
Let’s look at 2 more Top Drafts, the Broncos in 2010 and 2011:
Although there are some quality selections here (Von Miller, Julius Thomas, Demariyus Thomas, Eric Decker), there is the notable absence of a franchise QB (Although I still believe Tebow could have succeeded), and outside of Von Miller, there aren’t really any ‘flashy’ picks. Obviously, the lack of QB, but high level of success is the result of signing HOF QB Peyton Manning in 2013, which again shows the Draft isn’t necessarily the #1 way to get great talent.
Reverse Engineering the ‘best’ seasons since 2003
What happens when we look at the drafts preceding the ‘best’ (highest win%) seasons? This table shows the highest win% seasons since 2003:
The obvious team to look at here is New England. Let’s look at their individual draftees 2003-2007, which would include their 2003, 2004, and 2007 seasons:
These drafts obviously don’t include the ultimate gamechanger in Tom Brady, who was drafted in 2000, or Superbowl MVP WR Deion Branch (drafted in 2002). There are a few great players in these drafts (Vince Wilfork, Lawrence Maroney, and Benjamin Watson, among others), but no truly amazing players, apart from perhaps Wilfork. So, if they didn’t draft amazing players, how were they so successful? The simple answer is Belichick is a football genius, Brady is (arguably) the greatest QB of all time, and they made some quality free agent and trade acquisitions in this time (Randy Moss anyone?).
The 2015 Carolina draft tells a similar story, with some good players (Kelvin Benjamin, Devin Funchess, Kony Ealy) being taken, but the true difference makes being Cam Newton, and the scheme designed for him to use his freaky athleticism.
So, having ranked teams by lots of different measurables, reverse engineering great seasons, and looking at individual draftees, do we have a recipe for success? The answer is no. No draft and no team are alike, with each team having different needs, schemes, free agents, coaching, scouting, and priorities on different players. If I was to summarise my take on this data, it would be:
- If you are QB needy, and there is a generational QB, take him (e.g. Andrew Luck). Obviously, you can get great QBs in lower rounds (Russell Wilson, Tom Brady), but most good QBs are first round picks.
- Caveat – unless the QB you want has no question marks, don’t reach, otherwise you will end up with a Ryan Leaf/Vince Young scenario.
- Another caveat – make sure you have a solid OL if you are going to start a rookie QB, otherwise you end up with a David Carr scenario.
- If you aren’t QB needy and have a high pick, either trade down, accumulate picks, and fill your needs (ideally from inside (OL/DL) out), or if there is an extremely talented skill position player (Barkley/Fournette/Hopkins) and you have the framework for them to succeed, take him (see Cowboys taking Ezekiel Elliott when the have the #1 OL).
- If you are going to tank, tank hard, and try and get a top 4 pick. You more than likely aren’t going to find generational QB talent outside the top 4 (there are, of course, outliers like Brady/Wilson).
- Have Bill Belichick as your HC.
Feel free to disagree with me, and have your own takes! There is no one true recipe for success, and if there was, you wouldn’t have the 2016/2017/2018 Browns.