Stadium Impact on IDP Statistics

  Posted 8/24 by Aaron Rudnicki, Exclusive to Footballguys.com

By their nature, IDP statistics can be very subjective at times. When a RB is tackled by a single player in the open field, it's obvious that a solo tackle should be awarded and nothing else. But how does an NFL statistician award a tackle on a play when there are several defenders involved? When a DE puts his hand up and bats down a pass, it's clear that he should be credited with a PD (pass defensed) since the receiver had no chance to make a completion. But what happens if the deflection happens in the secondary with a CB and WR both fighting for the ball? We've all seen instant replay decisions in the NFL where it's nearly impossible to determine whether a player caught the ball or not. While you might feel certain that the play was a catch, someone else looking at the same exact video evidence might come to the opposite conclusion. With 11 defenders rushing toward the ball carrier at top speed, trying to keep track of who made each tackle is probably a bit like making those same types of replay decisions on nearly every play. Unlike with the reviewable plays, however, the spotter probably won't have nearly the same amount of time and resources to make each decision. When people have to make decisions quickly, it becomes much easier for bias to enter the process.

A couple years ago, I posted some data from the 2005 season in which I tried to determine how much of an impact the stat crew at each stadium was having on IDP statistics. One of the most interesting findings at the time was that the stat crews in Philadelphia were handing out 12.6 passes defensed per game, which was 2.5 times the league average of 5. In contrast, the New England crew handed out just 2.8 per game, which was about half of the league average. Since the crew in Philadelphia was awarding nearly 10 more PDs per game than the crew in Foxboro, when drafting your team or making weekly lineup decisions, you'd obviously want to favor Eagles players over Patriots to take advantage of this. There were similar differences apparent in the assisted tackle numbers as well. For example, IDPs playing in Miami or Pittsburgh were being credited with over 20 assisted tackles per game combined, while IDPs playing in Atlanta or St. Louis were being credited with fewer than 5 assisted tackles per game. While that difference was offset somewhat by a reverse trend in solo tackles (i.e., crews hand out more assists and fewer solos), there clearly seemed to be a different standard being used depending on where the game was being played. Although these examples were the most extreme, they do paint a pretty clear picture of the importance that a stadium statistician can have.

I compiled the same numbers for the past two seasons and took another look at the data to see if the earlier trends help up over time or whether there might be any new trends emerging that IDP owners could potentially gain an advantage from.

Here are the assisted tackle numbers from 2006:

This is the total number of assisted tackles for each team divided by home and away games, and it's sorted by the percentage that were awarded at home. You can see that Steelers players got almost 70% of their assisted tackles while playing at home. The 69.3% total for Pittsburgh here was nearly identical to the 69.7% they had in 2005. The Patriots, Giants, and Titans were the only other teams with home assist percentages higher than 60% for both seasons. Meanwhile, the Chiefs, Bears, Rams, Jaguars, Cardinals, Falcons, Raiders, and Vikings all ranked in the bottom 8 for both seasons, suggesting that the stat crews in those cities may be stingier than most when it comes to awarding assisted tackles.

Here is the same table for 2007:

Some of the earlier trends are still apparent here. All four teams who were above 60% in 2005 and 2006 (PIT, NE, NYG, TEN) were still all at 58% or higher in 2007 as well. Although we don't even know if the same statisticians are involved from year to year and they could likely change in the future, the fact that these patterns have held up over 3 full seasons does suggest that there is something going on here. If your scoring system happens to weight assisted tackles favorably, then it might not be a bad idea to target players from these teams or at least start them when they are playing at home. As for the teams that ranked near the bottom in the previous two years, the only ones that still ranked low again in 2007 were the Bears, Rams, Jaguars, Cardinals, and Falcons. So, all other things being equal, it might not be a bad idea to avoid starting players from these teams when they are playing at home since they seem to be at something of a disadvantage when it comes to assisted tackles.

To illustrate the scope of these differences, the home and away columns each include a sum of assisted tackles over 8 games. So, for example, all Rams defenders combined only averaged about 5 assisted tackles per game when playing at home while defenders from the Lions and Jets combined to average over 20 assisted tackles per home game. That's a pretty wide range and gives you an idea of how much impact something as simple as where the game is played can have.

For comparison purposes, here are the solo tackle numbers for 2006.

And for 2007.

You can clearly see that the range is much smaller here with nearly all teams coming in between 45 and 55%, and the overall league average very close to 50%. Since these stats are much less subjective than assisted tackles, it makes sense that we'd see less variation here. It's worth mentioning that some of the teams which seemed to rank highly in terms of assisted tackles are showing up on the low end of the solo tackle rankings. For example, the Steelers ranked on the lower end in home solo tackles for all 3 seasons. This could mean that the stat keepers in Pittsburgh are using a different standard whereby they choose to hand out multiple assists to several players involved in a play rather than a single solo to the primary tackler. The other 3 teams that ranked very high on assisted tackles at home, however, did not show a corresponding drop in solo tackles here, indicating that those stat keepers might be more likely to hand out a solo tackle to the primary tackler and an assisted tackle to the secondary tackler, potentially inflating the numbers a bit. If true, you may want to target players from the Patriots, Giants, and Titans.

As mentioned earlier, another area where subjectivity can play a big role is the passes defensed (PD) statistic. In 2005, the Eagles players ranked 1st in the league with 127 PDs combined, and an amazing 95 of them (75%) were awarded while playing in home games.

Here are the numbers from 2006, once again sorted by the percentage awarded in home games:

Note that the Eagles players still ranked first in the league in 2006 and still received over 70% of their PD stats while playing in home games. The Eagles were awarded 88 PDs in 8 home games, which was 34 more than the 2nd place Chargers were awarded in their 8 home games. Clearly, the stat crew in Philadelphia was very generous with these statistics.

And for 2007.

In the 2007 season, the Eagles actually looked like an average team, with just about 52% of their total PDs awarded in home games. This was a pretty big shift from previous seasons so the days of loading up on Eagles DBs and expecting plenty of PDs to follow appear to be over. The teams which ranked among the lowest in the league for all 3 years were the Vikings, Cowboys, and Chiefs, but they didn't appear to be too far below the league average.

The next question I was interested in looking at was whether the stat keepers in these locations are actually biased towards the home teams, or do they simply hand out more assists to both teams. Many have speculated that the crews in places like Baltimore and Miami were often crediting every possible tackle to players like Ray Lewis and Zach Thomas, thus inflating their numbers compared to what we might see from a more objective observer. To look into this, I divided the data up by stadium where the games were played, and then looked at the stats awarded to the home team vs. the away team (all 8 road teams are grouped together).

Here are the assisted tackle numbers by stadium for both seasons, sorted by the total number of assists handed out per game (for both teams combined):

And for 2007.

First thing that stands out is the overall drop in the number of assisted tackles handed out in 2007. The average awarded per game dropped from 28.4 in 2006 to 25.1 in 2007 (it was 27.3 in 2005). That's a pretty substantial drop and there's actually a good reason for it. As discussed in an earlier FBG article by Jene Bramel, the drop was not related to a decrease in the number of plays from scrimmage but rather to an effort by the league to improve the consistency and reliability of these defensive statistics. The league sent out detailed guidelines, including a 25 minute video, which provided uniform standards for all stat crews to use. The new overarching guideline for all defensive stats, including sacks, is to credit only those players who had a material impact on the play. Also, with regard to the bias question, the overall numbers from 2006 showed nearly 53% of awarded assisted tackles going to home teams, but that percentage dropped down to 51% in 2007, so it looked like the stadium crews might be awarding these statistics more fairly as well.

While the overall numbers showed a drop, a closer look at the assisted tackles handed out by each stat crew suggests that the impact may not have been as great as expected. For example, the stat crew in Saint Louis continues to hand out a ridiculously low number of assists with just 9.4 per game going to both teams combined in 2006 and 2007, which was comparable to the 9.8 from 2005. The same could be said of the crews in Atlanta and Jacksonville as both awarded fewer than 14 assisted tackles per game in all 3 years of data that I looked at. The statisticians in these stadiums certainly seem to be using a different standard when tracking tackles compared to the rest of the league and the new guidelines don't seem to have had much of an impact. For a stadium like Kansas City, however, there did seem to be a move in the right direction as we saw the average number of assists handed out go from 13.6 in 2005 and 11.4 in 2006 all the way up to 19.5 in 2007. In contrast, Miami went from awarding 48.2 assists per game in 2005 and 42.8 in 2006 all the way down to just 11.0 in 2007. This suggests that Zach Thomas may have benefited from some inflated assist numbers during his time in Miami, and that same advantage might no longer be available for Channing Crowder.

There's some interesting things happening with the Steelers numbers as well. Earlier we saw them ranked first in the league in 2006 with almost 70% of all Steelers player assists being earned in home games. Here, we see that the statisticians in Pittsburgh weren't really biased in favor of Steelers players since the % of assists going to home teams (52.2%) was right around the league average. In 2007, however, the Pittsburgh stadium ranked dead last in the percentage of total assists handed out to the home team (39.1%) while the total number of assists given to both teams combined ranked first overall by far with an average of 44.8. So, it looks like the stat crew in Pittsburgh really prefers to hand out multiple assists over a single solo tackle, but they also were very generous with the road teams in 2007 for some reason.

For comparison purposes, here are the solo tackle numbers broken down by stadium:

And for 2007.

One important thing to point out here is that the number of solo tackles didn't drop nearly as much as the number of assisted tackles did, suggesting that there were some real changes in the way that stat crews awarded tackles in 2007. In other words, it looks like the overall trend shifted from awarding a solo tackle plus an assist on a single play to now just awarding a solo tackle where appropriate. So, IDPs who rely heavily on assisted tackles might be losing some of their value under these new NFL guidelines. It's also interesting that the Pittsburgh stadium crew has consistently ranked among the lowest in the league in terms of solo tackles awarded per game, going from 81.6 in 2005 to 83.8 in 2006 to 75.9 in 2007. Some other stadiums that consistently ranked near the bottom of the league in all 3 years were San Francisco, Cincinnati, Detroit, and Dallas. One theory would be that these stadiums are home to offenses that focus heavily on the running game, which leads to fewer plays per game and thus fewer tackle opportunities, but doesn't really seem to fit for the 49ers or Lions. It seems likely that crews in some of these stadiums were less likely to award a solo tackle on an out of bounds play while others were more generous in those types of situations. With the San Francisco stadium ranking dead last in solo tackles awarded last year with an average of just 74 per game, which was nearly 14 per game lower than the league average, it's pretty amazing that 49ers rookie Patrick Willis still managed to lead the league in that category. The only stadium which clearly ranked among the league leaders in solo tackles awarded each year was the Jets, who ranked 1st in both 2006 and 2007. Over the past two seasons, the crew for the Jets awarded an average of nearly 98 solo tackles per game, which was about 10 solo tackles higher than the overall league average. This certainly looks like it could help the value of a player like David Harris, who had a monster second half after taking over for Jonathan Vilma last year.

In terms of passes defensed, the Philadelphia stadium stood out in 2005 by awarding 25.2 per game, which was 2.5 times the league average of 10. Here are the PD numbers from 2006:

You can clearly see that the Philadelphia stadium crew continued to award an unusually high number of PDs compared to the rest of the league but they had dropped by 5 per game from the year before. They were closer to the norm in 2006 but they still awarded nearly 3 times as many per game as the Indianapolis crew.

Here are the numbers from 2007:

Notice that the Philadelphia crew was actually below the league average in 20007, a huge shift from prior seasons and another sign that the league guidelines were actually being followed. Additionally, the range between the most generous stadium in Baltimore (11.8/game) and the least generous in Indianapolis (7.3) was only 4.5 per game. Split up among both teams, that's really just 2 extra PDs per game between the 1st and 32nd ranked stadiums and probably not worth worrying about.

To get a sense of the overall numbers, here is a table of all statistics awarded by each stadium for the past 3 seasons combined, ranked by fantasy points (using FBG scoring) per game:

The Jets stadium seems to have been the most IDP friendly place to play over the past 3 seasons. The stat crew there has handed out 17 fantasy points more per game than the league average, and a full 40 points more per game than the league's stingiest crew in Dallas. The only other stadiums that were at least 10 points above the league average were Tennessee, New England, and Baltimore. If these trends hold up, it might not be a bad idea to target some key IDPs from these teams like David Harris or Keith Bulluck. Don't expect the stadium effect to turn a scrub into an elite player, but it certainly could have an impact over 8 home games. Furthermore, if you look at the Jets, they are in the same division with the Bills and Patriots, plus they also have road games scheduled at San Diego, Tennessee, and Seattle this year. So, Jets IDPs are likely to play 13 out of 16 games in stadiums that ranked among the top-10 in fantasy points allowed per game over the past 3 seasons.

Some of the least IDP friendly places to play over the past 3 seasons were Dallas, Jacksonville, Cincinnati, Miami, and Minnesota, with all stadiums awarding at least 10 fewer fantasy points per game than the league average. For those expecting big bounce back seasons from players like Zach Thomas and Mike Peterson, it might be reasonable to temper your enthusiasm slightly as they could have an additional hurdle to overcome. However, as we saw with Patrick Willis last year, it's certainly possible for a great player to overcome any limitations of the stadium he plays in.

While some of the variation on a year to year basis is likely explained by the changes in personnel and coaching, along with just the raw number of plays (which is something I plan to control for in future analyses), taking a look at the entire 3 year period likely reduces the impact of those differences substantially. Obviously, there were some important changes in the way that these statistics were tracked prior to 2007 so these patterns are likely to keep changing and it might be a mistake to rely too heavily on numbers from 2005 and 2006. However, I think there was still quite a bit of consistency from year to year in the way that these stadiums track these statistics that is worth investigating. The difference between a top-10 player at his position and a backup might only be about 2 or 3 fantasy points per week over the course of a season. Also, when making a tough decision on who to start each week, if you know which stadiums tend to be more IDP friendly, that could wind up being a key piece of information that helps you win your game and eventually make the playoffs.

Feel free to e-mail me with any questions on this article at rudnicki@footballguys.com.