Does the following logic make sense? The more prolific an NFL quarterback is as a player the more popular they will become. Popularity then leads to more demand for a player's merchandise, such as jerseys. With a bigger demand on those jerseys in a free market, such as eBay, the jerseys of the better players will then sell for more than their inferior counterparts.
That string of thought seems to make sense on the surface, but under a scientific test how does it hold it up?
The following test was set-up to scientifically compare QB's skills and the price their jerseys sell for on eBay with the following rules in place.
1. Every QB's ability will be based on the last year's performance as judged by passer rating. There is no good scientific way to take into account a QB's running abilities or residual past popularity as part of his passer rating.2. Each NFL team will be given one QB to be used in the study. The QB used will be the player who started the most games for them last year. If both QB's started 8 games, the player still on the team's roster will be used.
3. A given player's eBay jersey sale price will be determined by taking the first ten authentic home or away with tags jersey sales of the player (size does not matter). If the player does not have ten jersey sales then he will be thrown out the competition. The auction can be either standard or Buy it Now.
All of the information on the QB's was gathered with the following players being used for each team (a player who did not reach the pre-requisite 10 jerseys a * will be by their name):
| Teams | QB Starting The Most Games | Season QB Rating | Average Jersey Price |
| Arizona Cardinals | Matt Leinart | 74 | 20.69 |
| Atlanta Falcons | Michael Vick | 75.7 | 28.94 |
| Baltimore Ravens | Steve McNair | 82.5 | 22.67 |
| Buffalo Bills | J.P. Losman* | 84.9 | 25.16 |
| Carolina Panthers | Jake Delhomme* | 82.6 | 20.65 |
| Chicago Bears | Rex Grossman | 73.9 | 20.26 |
| Cincinnati Bengals | Carson Palmer | 93.9 | 25.20 |
| Cleveland Browns | Charlie Frye* | 72.2 | 0.00 |
| Dallas Cowboys | Tony Romo | 95.1 | 13.00 |
| Denver Broncos | Jake Plummer* | 68.8 | 12.99 |
| Detriot Lions | Jon Kitna* | 79.9 | 0.00 |
| Green Bay Packers | Brett Favre | 72.7 | 17.69 |
| Houston Texans | David Carr | 82.1 | 18.57 |
| Indianapolis Colts | Peyton Manning | 101 | 21.54 |
| Jacksonville Jaguars | David Garrard* | 80.5 | 0.00 |
| Kansas City Chiefs | Damon Huard* | 98 | 0.00 |
| Miami Dolphins | Joey Harrington* | 68.2 | 0.00 |
| Minnesota Vikings | Brad Johnson* | 72 | 0.00 |
| New England Patriots | Tom Brady | 87.9 | 20.42 |
| New Orleans Saints | Drew Brees | 96.2 | 21.09 |
| New York Giants | Eli Manning | 77 | 25.98 |
| New York Jets | Chad Pennington | 82.6 | 22.64 |
| Oakland Raiders | Andrew Walter* | 55.8 | 0.00 |
| Philadelphia Eagles | Donovan McNabb | 95.5 | 27.94 |
| Pittsburgh Steelers | Ben Roethlisberger | 75.4 | 29.12 |
| San Diego Chargers | Phillip Rivers | 92 | 23.03 |
| San Francisco 49ers | Alex Smith | 74.8 | 25.29 |
| Seattle Seahawks | Matt Hasselbeck | 76 | 34.61 |
| St. Louis Rams | Marc Bulger* | 92.9 | 0.00 |
| Tampa Bay Buccaneers | Bruce Gradkowski* | 65.9 | 0.00 |
| Tennessee Titans | Vince Young | 66.7 | 31.20 |
| Washington Redskins | Mark Brunell* | 86.5 | 0.00 |
Now, staring at a large pile of random numbers does not usually tell much to the naked eye, but when the numbers are analyzed something very interesting occurs. When a graph is created mapping a player's passer rating against their average jersey selling value it looks like this:
What both simply looking at the data points and the linearization tells us is that there is really no correlation based on our data. We have an R-Squared value of only .08 (meaning extremely low correlation, go here for more details) and our limited data actually suggests as you become a better passer your jersey selling price decreases slightly. While that in all actually probably is not true since the correlation is so small, it is still a funny result to add in with the overall lack of correlation. With these results in though, we have to ask. Why did this happen?
Study Related Issues:1. A sample of 10 jerseys is likely to small to cancel out some of the strangely high and low selling jerseys. A sample of 20 or 30 likely would have made the results more clear.
2. When evaluating the quarterbacks their running ability was not taken into account. QB's like Michael Vick and Vince Young derive much of their popularity from the plays they make with their feet.
3. The number of jerseys for a player that remained unsold was never taken into account. A dedicated seller could decide to post the same jersey for a given player time and time again waiting for a big payday from a high Buy It Now or auction starting price while another seller may be content to sell cheaply in bulk. Auctions starting from a dollar or less my give a better indication of the player's true average jersey sale number.
4. Certain sizes of jerseys may be slightly more competitive in bidding for. Many people wear large and XL so those jerseys may end up selling for more than others, so if in our small sample size there was an overabundance of them for a given player then it may have skewed his results.
Real World Explanations:1. When a player's jersey is selling for less than $30 people stop being so picky about who the player is and how good they are. When a jersey is a few hundred dollars people are more choosy about the name written across the back.
2. With the plethora of jerseys constantly being sold on eBay there is never good motivation to pay a high price for any player. Most people when it reaches above a certain amount will just wait for the next auction to come around and try again until they get lucky.
3. Many people, especially in football as compared to basketball, will buy a player from their favorite team more than just buying a specific player. In basketball, Kobe, Lebron, and Wade jerseys fly off the shelves while in football many fans are more likely to buy their team's starting QB and be willing to pay for it rather than just buy elite player's jerseys. Some fan bases on more rabid than others even so they will pay more for their inferior than another more calm fan base will pay for their superior QB.
4. If a player one player is better than another than even with the "buy your team" effect he may still have more jerseys produced than the other player. If this is the case then if the end demand on eBay for the better player is not likewise greater than for the other then there is an oversupply of that players jersey driving down the price.
5. Many young players are not developed enough to produce high passer rating, but are popular and sell well because of there college days and pro potential. Players like Matt Leinert and Vince Young are not great passers yet, but they are very well known throughout the league and people see them as a couple of the next great NFL QB's which helps drive their sales.
6. If a player plays in a larger market he will likely have a higher demand for his jersey than those who play in a smaller market. Chad Pennington and Eli Manning both had worse passer rating than other players, but since they play in New York they were both able to outsell price wise even Peyton Manning who plays in Indianapolis, a much smaller market.
7. Even as a player ages and loses his abilities/has a down year, often his popularity still lingers in the minds of his fans. Matt Hasselbeck had a horrible statistical year in 2006, but coming off a Super Bowl appearance he was still able to have amazing jersey selling numbers.
When all these study factors and real world explanations in the books it, it actually makes alot of sense that the results came out how they did. The NFL is not the player driven league that the NBA is where individual players command much of the attention, so it is not surprising that individual player items fall into the pattern they did (I would/someone should do this with NBA jerseys as well, maybe take the top 20 scorers or PERers and do a comparison). It is smart to note that the logic that seemed to make good sense at their start of work did not stand up to more rigorous scientific testing that was done showing again we cannot always trust the first impression of a situation (omg... hes preaching, *back button*). Even with a more refined study the current data shows very little hope of completely bouncing in a new direction with the correlation where it is, letting us draw out the theory finally that, "A QB's skill level has no bearing on the price that his jersey will sell for on eBay". The hypothesis was wrong.
Look for another eBay article coming sometime this week along with other topics that will be discussed as new series evolve. Until then, I'll see you on the auction page (if you try to outbid me I'll block your IP address from this site...).
