[Statistiques avancées] Can we draft better? | Touchdown Act (NFL Act)

In the 1930s, Bert Bell, the owner of the defeated Eagles franchise, proposed a new idea to balance his game. He wants to set up a lottery to prevent the best university players from being attracted to the best professional projects, which will allow less good teams to get priority over the most promising talents.

The draft was born out of his idea, and 6 years later the first choice in history will be chosen by his right.

But was the draft really balanced in the way he wanted? Some teams seem to be subscribing to first selections for years, while others seem to be constantly asking for play-offs. Can this difference be explained by the ability to draft better than others? This is what we will try to determine today.

L Approximate Value (AV) D Pro Football Note

It is difficult to calculate the contribution of players playing at different levels due to the lack of common statistics. Pro Football Reference AV (Approximate Value) attempts to evaluate each player’s contribution by looking at the points scored and collected by the team and distributing these points to each player according to their playing time.

This statistic cannot be used to accurately compare two players, but it does give a general idea of ​​their status. For example, you can find players like Aaron Donald, Aaron Rodgers, Fred Warner or Xavier Howard in the top 10 this year. Its use is appropriate for a study like ours, which has a very large number of players.

Review Jimmy Johnson’s charter

Jimmy Johnson, one of the first to calculate the value of each choice in the draft, coached the Cowboys in the early 90s. Its charter, which sets a point value for each exam, is still used today to negotiate trades.

In his model, the first choice worth 3000 points is worth 5 times as much as the first choice of the second round. After the second round, the value of each test decreases slightly.

But does this 30-year estimate agree with reality? To verify this, we will subtract the total VA of each player drafted between 2000 and 2017 and reduce the number of games with a common size.

In the blue curve (trendline), early choices show better returns on investment than later choices. It promises NFL teams. They are generally the best at assessing each player’s theoretical space throughout the draft.

By taking the equation of our trend curve, we can get an idea of ​​the expected value of each exam depending on the exam number, so assign a value to each exam similar to Jimmy Johnson’s model. This value is called xAV or AV.

To compare this model with Jimmy Johnson, it is necessary to keep them on the same level. We will express each choice (arbitrarily) according to the value of the 33rd test of each model. In other words, the value of the 33rd test is 1.

Despite the uncertainty of our soft sample draft, Jimmy Johnson is incredibly optimistic about the value of the first round selections. And, conversely, tends to underestimate the outcome of the draft. If it turns out that teams are still using this method, they overestimate their ability to fall right in the first round.

Evaluate each draft

Now a model gives us an idea of ​​how a player should act to justify his choice, by which choices and by extension we can determine which drafts are successful. For that we will create AVOE (AV Over Expectation, above AV expectations). This data is obtained by subtracting a player’s VA from the expected VA. We know if a player has acted beyond expectations regarding his selection number.

More precisely, the model is modified to suit the level of the players. A team will not have the same expectations when drafting a receiver than a quarterback in the third round. This prevents a draft from being weighed by selecting a special team player. Taking this into account, we get an AVOE value, which depends on the level and the player’s choice number.

Here are the 5 best and worst drafts of the century according to this model. Note that the simulation does not say which team selected the best players, but rather selects the best players according to their draft number. For example

The famous Seattle draft of 2012 has been installed above, with multiple positions to choose from and all bias at the end of the process. The 2009 Broncos have 2 first rounds and 3 second rounds for a total of 10 exams. With the exception of Robert Ayers (18th) and Noshon (12th) Moreno, who had passable careers, the other 8 players played a total of 24 games. This strengthens the reliability of the model.

Predict future success

Thanks to the work done earlier, we now know how to calculate the capacity of teams with bare noses. Let’s see if this skill will be repeated over time. For a given group, let’s see if there is a correlation between its ability to draw well in a year and beyond. The total for a particular year on the X-axis is in AVOE and the total for the y-axis is the same for the next year.

The correlation with a coefficient of 0.05 is almost non-existent. Looking at the 3 year cycle, we get more or less the same thing. There are definitely trends of teams drafting more or less. But overall no team stood alone. The relationship between two drafts over two consecutive years is determined by the inconsistency rather than the ability of the teams to accurately predict the status of future players.

This is why it is dangerous for owners to blindly believe Jimmy Johnson’s charter, which in no way likes variations on the drafting process. Sometimes the first choice 33 is worth 5 times, but it is blindly believing that it ignores the uncertainty of the draft. Owners exaggerate the ability to design well when this is a volatile event.

Conclusion

The drafting process is very complicated and over time it becomes a hassle that can be copied. All the teams that were able to maintain high quality for a few years fell apart. After 3 drafts in the top 5 places of the year, John Schneider and Pete Carroll fell in line. No matter how talented the staff is, a player’s success is unlikely to be accurately predicted each year. Draft in this sense is really a lottery, no one can win it emotionally.

Bert Bell wanted justice when creating the draft, and he was built. Every year in the battle room, all teams can greet each other because when luck comes there is no better than the other.