Going Deep: Touchdown Regression – Negative Quarterbacks

The Magnetic Pull of the Mean

 

Every year players on your fantasy team score you points. You have your league winners. These are the players that outperform what was expected of them based on where you acquired them, whether in the draft, waiver wire, or trade. Then you have your team tankers. These are the players that crushed your team’s hopes because they did not perform up to the expectations.

One of the most fun things about fantasy football is that every year you can start fresh and use what you learned from last year to make your best guess as to what is going to happen in the upcoming season. Regression to the mean is a way that you can gain an edge on your opponents. Think of the mean, or average, as a magnetic pull. All fantasy players are being pulled towards the mean. Some players, the league winners, have such a strong force towards above-average performances that they will continue to score above average points for your team. Some players, the team tankers, have such a strong force towards below-average performances that they will continue to score below average point totals for your team.

Eventually, just like father time (I’m sure there is a strong correlation), the magnetic force of the mean always wins out. Players performing above average start to creep back or even slingshot back toward the magnetic pull of average play. The same, thankfully, for the opposite. Players that are performing below average for fantasy teams will inevitably make their way back toward the magnetic center of the mean.

Identifying which players have been under-performing in regards to the mean will help you decide the players that you want to invest in, as other fantasy owners believe that their below-average performance of the past is permanent. Alternatively, identifying over-performing players will help you avoid those with no hope of keeping up that above-average pace and no hope of escaping the tractor beam that is the mean.

 

The Process

 

This series of articles will focus on arguably the most exciting part of both real and fantasy football, the touchdown.

Touchdowns are notorious for being “not sticky” or unpredictable from season to season. This means we can not look at Dak Prescott’s 30 passing touchdowns last season and pencil him in for the same amount in 2020. You can’t even rely on his touchdown rate to repeat itself. According to T.J. Hernandez at 4for4.com, passing touchdowns have a correlation season to season of just .36, and touchdown rate is even lower with a season to season correlation of .21.

Instead of looking for an exact touchdown total for quarterbacks in the 2020 season, let’s go back to what we talked about at the top. Let’s see which quarterbacks underperformed and overperformed in the touchdown column in the 2019 season.

We know that the magnetic pull of the mean is endless. Some players regress more than others. Yes, this even applies to unicorns like Patrick Mahomes. The measures that predict how many touchdowns a quarterback should score (what I call Expected Touchdowns) is measure through passing yards and rushing attempts. Looking at the tables below, you will be able to see the r-square values for passing yards per touchdown and rushing attempts per touchdown for the quarterback position. What this r-squared value means, in simplest terms, is how much one variable (passing yards) explains (correlates) to another variable (passing touchdowns). If you want to know a more in-depth explanation, you can read about it here.

Dating back to the 2015 season, here are the correlations between quarterback passing yards and passing touchdowns:

 

 

Dating back to the 2015 season, here are the correlations between quarterback rush attempts and rushing touchdowns:

 

 

In some areas of statistics, an r-squared value of .54 would be considered pretty weak, to very weak, but not in football. In football, there are a lot of variances and a correlation of .3-.7 is moderate. Any correlation that is greater than .7 would be considered strong. As you can see from the r-squared value, the variables of passing yards and passing touchdowns for the quarterback position are moderate. The same can be said for rushing attempts to rushing touchdowns for the quarterback position.

Using this information, we can feel comfortable using these conversions to help find 2019 touchdown overperformers and underperforming at the quarterback position. The next question becomes how much should we expect the player to be pulled back toward the mean. To try and put this into perspective, we will look at the player’s historical trends in their efficiency of converting their total expected touchdowns (eTD) into actual touchdowns (aTD). If 2019 looks like an outlier based on their previous season’s efficiency of converting passing yards and rushing attempts into touchdowns than we should expect regression to hit in 2020.

Without further to do, here are 2020’s top quarterbacks due for negative touchdown regression. For the quarterbacks due for positive touchdown regression, see Part 1 in the series.

 

2019 Over-Performers: Fewer Touchdowns in 2020

 

Lamar Jackson, Baltimore Ravens

 

You had to see this coming right? Lamar Jackson was one of, if not the, best fantasy value in recent memory. According to Rotoviz’s NFL Stat Explorer Tool, Jackson had nine 30+ fantasy point weeks in 2019. Using the same tool for comparison, Patrick Mahomes had just four of those 30+ fantasy point weeks. What helped generate those massive scoring weeks? I can’t point to a single factor and say that was the difference-maker, but touchdowns certainly helped the cause.

Jackson had a total of 43 touchdowns (aTD) in 2019. He was the definition of a duel threat with 36 touchdowns coming through the air and seven more on the ground. When looking at what a fantasy player can expect from a quarterback with Jackson’s passing yardage and rushing attempts, it’s clear that Lamar massively overperformed in the passing touchdown department.

 

 

Based on his yardage total, Lamar Jackson should have passed for something more in the ballpark of 19 touchdowns. Instead, he very nearly doubled that expectation by throwing for 36 touchdowns. When looking at actual passing touchdowns over expected touchdown rate (apTD/epTD) since 2015, Jackson’s 2019 season was the most efficient on record. The next closest rate was 2017 Deshaun Watson at a 1.8 apTD/epTD rate.

 

The other gaudy number we need to keep in mind for Jackson’s 2019 season is how many rushing opportunities he had. While he did execute at an expected rate when it comes to touchdowns, the number of rushing attempts he was granted was in another stratosphere. Looking back to 2015, Lamar Jackson’s 2018 (147 attempts) and 2019 (176 attempts) seasons are well ahead of the next closest quarterback which is 2017 Cam Newton.

 

So what can we expect in 2020?

 

 

The scary news is that Jackson still has a path to overperform in the touchdown department. His conversion of rush attempts to touchdowns was right at league average for a quarterback. There is the argument to be made that he will not continue to get as many carries based on the moves the Ravens have made in the off-season, (specifically drafting running back J.K. Dobbins) but there is some wiggle room there for more efficiency on the ground. What’s nearly guaranteed is that he does not convert passing touchdowns at such an ultra-efficient rate. In the 2018 season, the one after Watson’s 1.8 apTD/epTD, Watson came back down to league average in apTD/epTD rate. In fact, the next five players on the list behind Lamar Jackson’s 2019 apTD/epTD, all fell in their efficiency the following season. On average, those quarterbacks dropped approximately .58 points. If Lamar, sticks to that trend, he will still be converting touchdowns at an above-average, 1.28 apTD/epTD rate.

 

Verdict: Lamar Jackson made history in the passing touchdown efficiency column based on his yardage totals in 2019. That is not something to expect in 2020. He should still be highly efficient in apTD/epTD though, just don’t expect 17 extra passing touchdowns. There is a path to more touchdown efficiency on the ground. Baltimore’s offseason moves (and the logic in protecting your starting quarterback from taking excessive hits) make it unlikely that Jackson gets so many rushing attempts in 2020 which would raise his arTD/erTD rate if he maintains the touchdown totals.

TL;DR: He was really, really good in 2019. In 2020 expect him to be really good.

 

Drew Brees, New Orleans Saints

 

Despite playing in only 11 games because of injury, Drew Brees finished the 2019 season tied for fifth in passing touchdowns, finishing with 27. That had him throwing touchdowns on par with league MVP Lamar Jackson in touchdowns per game with 2.45. Needless to say, if you held on to Brees or if you grabbed him off the waiver wire before his return from injury, it paid off.

 

 

Given the yardage and rushing attempt totals, Brees was expected to score 19 total touchdowns last season. He scored at an extremely high rate for his yardage totals and even managed to score a rushing touchdown, vulturing one from Alvin Kamara, Latavious Murray, and Taysom Hill. A three-man vulture is truly impressive. Brees has always been known for his gaudy stats and his accuracy, but has he always been this efficient at converting touchdowns?

 

 

The answer is: kind of. While Brees has almost always been above league average in converting passing yards and rushing attempts into touchdowns, he has never been as efficient as 2019, although he was close in 2018. In those two seasons, Brees led the NFL in completion percentage in the red zone (and in 2017!) which resulted in 20 of his 27 total touchdowns in 2019. As mentioned before, Brees is known for his accuracy, so it makes sense that his completion percentage over expected (CPOE) was the highest of his career in 2018 and 2019 according to airyards.com. When in the red zone, the field shrinks, and quarterbacks are not expected to complete as many passes. Brees, however, has not had any issues.

 

Verdict: Expect Brees to back down his touchdown total some. The Saints threw the ball more in 2019 than they had in their previous seasons, so we should see some regression there as well. It is apparent, however, that the Saints’ red zone scheme over the last two seasons has paired quite well with Brees’ accuracy. Expect him to get his touchdowns in the red zone, however, it should not be sustainable at such a high rate.

 

Ryan Tannehill, Tennesee Titans

 

Ryan Tannehill was a top 10 fantasy quarterback in fantasy points per game in 2019. In fact, according to FantasyData, Tannehill led the NFL in True Passer Rating which calculates quarterback rating “on all unpressured throwaways and dropped passes.” He also led football in completion percentage on play-action passes, completing an amazing  75.9%. Despite the efficiency (and partly due to his starting only 12 games after relieving Marcus Mariotta of his duties), Tannehill should have only scored 19 total touchdowns based on his numbers.

 

Tannehill was most efficient on the ground. He converted 43 total rush attempts into four touchdowns. What’s most impressive is only two of those touchdowns came within the 5-yard line, where we see most quarterbacks not named Lamar Jackson score from. While it seems that Tannehill’s 2019 season came out of nowhere, there were hints of his touchdown efficiency just the year prior in 2018.

 

 

When looking at a 2018 season where Tannehill threw for just 1979 yards, he managed to somehow throw for 17 touchdowns. That yardage total is more likely to get a quarterback in the area of 12 touchdowns. In every other season of his career, Tannehill has hovered right around league average in touchdown efficiency. It is safe to say that we should see some regression back to those seasons in 2020, but with better talent and play-calling, he’s shown that he can still be above average in the touchdown department.

 

Verdict: Don’t expect the amazing touchdown efficiency that came in 2019. However, the play-caller and the weapons on the field are still intact, so I would expect an aTD/eTD rate of right around .1, or slightly above league average.

 

Honorable Mentions

 

Here are the rest of the 2020 starting quarterbacks that scored three or more touchdowns above expected and should be due for some touchdown regression this season.

 

 

Check back in next week for the top underperformers of 2019 at the running back position.

 

 

All stats for this article are from fantasydata.com, profootballreference.com, airyards.com.

Photo by Rich Graessle/Icon Sportswire