Aug 27

TABM 2013 NFL Preview: Five Factors, AFC South

From year to year, the NFL experiences rapid and dynamic changes. As we explained upon the creation of this website, many of these changes have to do with regression to the mean. Because of this regression, a majority of these changes can actually be predicted. With proper research of the NFL stat books and a keen understanding of the game’s context, you can successfully predict the changes.

This is exactly what we did when creating last year’s NFL preview, as six spectacular articles foretold some major change in the NFL from 2011 to 2012. We look to do that again this year. While we already discussed some of the major changes, we want to expand the prognostications to the whole league. Therefore, I’ll spend eight August days discussing five factors of regression in the 2013 season for each team in the NFL. This will help to paint the picture for our season predictions, which will be made at the end of the month.

Today, we look at the AFC South.


Houston Texans (12-4, 1st place in 2012)
Houston TexansFor much of the 2012 season, the Texans looked like they were primed for a championship. They started the season 11-1 and held the top seed in the AFC for much of the year. However, the team faded down the stretch in disappointing fashion. The downfall began with a 42-14 loss in New England on Monday Night Football. The team later lost by 17 at home to Minnesota and loss by 12 at Indianapolis to drop down to the third seed. Houston was able to beat Cincinnati again in the Wild Card round. However, the Texans were once again blown out by the Patriots to fall short of their Super Bowl expectations. Perhaps there were warning signs before the downfall, as the team needed a 5-0 record in one-possession games to get to 11-1. Will the disappointing December and January be a sign of what’s to come in 2013? Or will the Texans be once again fighting for the league’s greatest prize? Much of that will rely on how much regression this team will face.

Down1. Close-game success: As already mentioned, the Texans won their first five one-possession games to get off to their great start. However, they actually finished 6-0 in one-possession games when you include the postseason. Therefore, that makes the Texans 7-5 in games determined by more than one possession. Say the Texans break even in those close games, which is what you would have in a mean situation, and the Texans are 10-8 when the postseason is included. No wonder the Texans with 10.16 Pythagorean wins in the regular season, according to our adjustments to Bill James’ formula. The general rule of the thumb is that teams tend to regress when their real wins and Pythagorean wins are notably different. Several publications have suggested that the general threshold for potential regression is ±1.0 win. I’ll say that this merely a starting point, because regression can sometimes show how the difference is explained by rarely-occurring factors. (This is the case for the 2012 Giants, as explained yesterday.) As we’ll show below, regression doesn’t explain the -1.84 Pythagorean win difference. Regression results: worse W-L record

Down2. Run defense: Both the offense and defense performed quite well in Houston. For the defense, the best success came in stopping the run. Therefore, regression will primarily come against the run. Houston allowed only 1560 yards and five touchdowns on the ground. Successful teams are much more likely to allow fewer than 100 rushing yards per game, as proven by the list of 26 teams since 1983 to allow no more than 1600 rushing yards or five rushing touchdowns. Only one finished .500, while only one finished with a losing record. No team achieved this in consecutive seasons since the 1995-97 49ers accomplished a three-peat of the feat. With regression suggesting the Texans will drop a few games in the standings, opponents will have more opportunities in gain yards and touchdowns. Regression results: more rushing yards and touchdowns allowed

Down3. Andre Johnson: Along with Calvin Johnson and Brandon Marshall, this Texans receiver had a historically good season statistically. The three each totaled at least 100 receptions for 1500 yards, something that’s now only been done 21 times (per Pro Football Reference). This marks the third time Houston’s Johnson did this, which ties an NFL record with Marvin Harrison. Of the previous 18 times, only once did a player improve in both receptions and yards, and that was Harrison in 2002. As for Johnson, he pulled off the feat in back-to-back years in 2008 and 2009, but he regressed in both categories in both of the following seasons. Looking at Johnson’s numbers from 2012, as he collected 112 receptions for 1598 yards, it’s going to take a lot to top those numbers. Regression results: fewer receptions and fewer receiving yards

Down4. J.J. Watt: Another Texans superstar put up some mega numbers in 2012, but this one came from one of the young guys. Second-year defensive lineman J.J. Watt put together one of the best seasons in recent memory, as we mentioned earlier in this preview. He officially collected 20.5 sacks and 16 passes defensed. He became just the ninth player to collect 20+ sacks since the stat became official in 1982. Notice that no player did this in multiple seasons. Meanwhile, passes defensed more recently became an official stat, and it seems J.J. Watt is the new king of the swat among defensive linemen. It seems like double-digit swats would’ve been impressive enough to get our attention, but Watt went far beyond that. Finally, according to Football Outsiders’ stats, Watt finished with 56 defeats. The next best total since 1996 was 45 defeats, achieved by Ray Lewis in 1999. That’s too awesome to replicate in 2013, if ever again. Regression results: fewer sacks, fewer passes defensed and fewer defeats

Down5. Offensive health: According to Pro Football Almanac 2013, the Texans had one of the most healthiest teams in the league. No team finished with a better Adjusted Games Lost on offense than Houston’s 6.7 total. By all means, the Texans had an extremely peaceful season on offense when it came to roster adjustments. Seven starters played all 16 games, while the H-back James Casey and second tight end Garrett Graham missed only one game combined despite swapping back and forth in the starting lineup. Besides that, two other starters in tight end Owen Daniels and right tackle Derek Newton missed a combined three starts. Finally, there was a mid-season switch at right guard between Ben Jones and Antoine Caldwell. That luck will not happen again, as the AGL numbers are volatile from year to year. Regression results: more missed games by projected starters


Indianapolis Colts (11-5, 2nd place in 2012)
Indianapolis ColtsJust one year after a nightmarish 2-14 campaign, the Colts bounced back to grab another postseason berth. By allowing the release of Peyton Manning and drafting Andrew Luck with the first overall pick, it seems like Indianapolis hit the fast forward button on the rebuilding process. In his rookie season, Luck led the Colts on seven game-winning drives, including four fourth-quarter comebacks. This was the highlight for an up-and-down season. It also suggests that when Luck begins to reach his peak, he has the intangibles to win big playoff games. On the down side, with those frequent late and close wins by Indianapolis, it diminished the talk about the team’s shortcomings. The team did not put up the quality stats or efficiency stats you’d see from a normal 11-5 team. The Colts finished just 22nd in our 2012 TABRankings, and they totaled just 7.23 Pythagorean wins. Will this lead to significant regression in Indianapolis?

Down1. Andrew Luck: As we already mentioned, Luck showed some serious clutch ability as a rookie. His seven game-winning drives (GWD) rank in a tie for third in a single season (per Pro Football Reference). As discussed during the first article of this preview, that is an incredibly unusual feat that is nearly impossible to repeat for a second consecutive season. Considering Luck’s 7-3 record with GWD opportunities, and the team’s 9-1 record in one-possession games, any sort of regression to the mean could result in at least one or two fewer wins in 2013. Even if the Colts play significantly better, and they secure more chances to win before the fourth quarter, Luck could go 3-3 with GWD opportunities and repeat his 11-5 record as starter — or worse. Simply, the Colts had some luck in Luck. And although luck doesn’t take away what the Colts earned, it does mean that regression is looming. Regression results: fewer GWD

Down2. W-L Record: As we already mentioned the 9-1 record in one-possession games, this case of regression should only be quite apparent. When the score is decided by one possession, they are normally are closely-competed game. It should follow that a .500 record is likely in those situations, give or take one game. However, Indianapolis won 90 percent of those games, with means the team probably played more like an 8-8 or 7-9 team. Remember those Pythagorean wins we mentioned during the team’s introduction? That resulted in a -3.77 Pythagorean win differential, which is by and far the worst total in the league. This happened with the Colts facing the second-easiest schedule in the league, according to the TABRankings. Finally, let’s remember what we said earlier in the season preview about team that improve by at least seven games. Remember that the Colts went from 2-14 to 11-5. There will be some elasticity involved with the Colts heading into 2013. So let’s say the Colts improve enough to be two or three games better in 2013. That might not even be enough to maintain their 11-5 record. Regression results: worse W-L record

Up3. Big-play run offense: As we already said, the 11-5 record diminished the talk about the team’s shortcomings. That doesn’t mean the Colts struggled in certain areas — and were downright poor in other areas. One of the worst areas for the Colts was the rushing offense. The team never had a rush longer than Vick Ballard’s 26-yarder. In fact, for the whole season, Indianapolis totaled just five rushes of 20+ yards — doing so on 440 carries! Only the Chargers, Jets and Lions held a lower percentage of big runs. The Colts did enjoy some more moderate success in racking up 45 runs of 10+ yards (per Sporting Charts). Therefore, we can’t say there’s a guarantee of significant regression, although the addition of a few big runs will go a decent way. There will be some improvement if the ranking for 10+ runs stays in a similar area. Regression results: more yards per rushing attempt

Up4. Run defense: On the flip side of the field, the run defense had problems that went much deeper. The team allowed 2200 yards on 428 carries, averaging 5.14 yards per attempt. That puts Indianapolis among a very select group of playoff teams. The 2012 Colts became the fourth such team to allow at least five yards per rushing attempt (per Pro Football Reference, who has a glitch by including the 1944 Giants and 1945 Rams — both teams actually allowed fewer than three yards per attempt). Each of the first three teams improved their run defense. Oh, and remember that one of those teams was the 2006 Colts, who turned around their run defense by the postseason. Here’s one area the Colts will see significant improvement after they finished dead last in rush defense DVOA (per Football Outsiders). Regression results: fewer yards allowed per rushing attempt and better rush defense DVOA

Up5. Defensive fumble recoveries: One other small factor that should help the defense is simple regression to the mean for recovering fumbles. Not only did the Colts force an abnormally low amount of fumbles (11 in total), but they recovered a low percentage of fumbles (giving them three fumble recoveries in total). Indianapolis ties the 1974 Rams, 1995 Packers and 2011 Dolphins for the fewest fumble recoveries in a single season. The team will definitely see a boost from regression, as the team also had only 15 takeaways for the season. Regression results: more fumble recoveries and more takeaways


Jacksonville Jaguars (2-14, 4th place in 2012)
JaguarsFor a franchise that faces a potential threat of moving away from Jacksonville, the 2-14 season by the Jaguars might mark an all-time low. Jacksonville recorded a five-point victory in Indianapolis and a five-point victory at home against Tennessee. However, that was coupled by seven losses of 17+ points. The horrific season sees the beginning and end of the Mike Mularkey era. However, it doesn’t see the end of the underwhelming Blaine Gabbert era. He will seemingly get another shot at the team’s starting gig with new head coach Gus Bradley, who was last seen as the fiery defensive coordinator for the Seahawks. Gabbert has not come close to fitting the part of “franchise quarterback,” but he does only have two seasons under his belt. Bradley will hope the third time is the charm. Beyond that, Bradley will be able to enjoy some significant amount of regression. The question that lies is “how much regression will happen?”

Up1. Rushing offense: Considering that the Jaguars were so often down and out of games, it’s not much of a surprise that the team didn’t have time to establish the run. As more progressive football publications explained, teams don’t establish the run to win games. Instead, teams establish the run when they are leading. Actually, winning the passing efficiency battle most effectively gets a team the lead. For the entire year, the lowly Jaguars totaled 1369 yards and five touchdowns on the ground. Excluding the strike-shortened 1982 season, only 17 teams since the merger have failed to surpass 1400 yards and five touchdowns for the season (per Pro Football Reference). No franchise had this happen to a team in consecutive years. We don’t expect the Jaguars to become the first franchise to do so. Regression results: more rushing yards and more rushing touchdowns

Up2. Defensive efficiency: The Jaguars finished 5-11 in 2011 despite finishing with worse yards per play (4.14, compared to 4.82 in 2012) and a worse scoring offense (243 points scored, compared to 255 in 2012). That’s because the defense performed much better. According to the play-by-play efficiency metrics created by Football Outsiders, the Jaguars dropped from a -11.3% defensive DVOA in 2011 to an 11.7% defensive DVOA in 2012. Jacksonville was ranked fifth in that category, only to fall to 28th. One statistical theory semi-related to regression is the elasticity theory, which states that a vast improvement or decline will lead to a “regression” towards the middle of the two in the following season. We qualify this as a “vast” decline, so at least the Jaguars can enjoy getting one step closer to being average on defense. Regression results: better defensive DVOA

Up3. Drive success: The Jaguars reached a rare level of futility when it came to sustain drives and stopping opponents. Jacksonville finished dead last in the league with -9.23 net yards per drive (per Football Outsiders). That marks the first time since 2009 that a team had worse than -9.0 net yards per drive. (In 2009, the Raiders had -10.40 net yards per drive, the Rams had -10.51 net yards per drive and the Browns had -10.76 net yards per drive.) Since 1997, only twice did a franchise finish below that threshold in back-to-back seasons. Both time involved the Browns (1999-2000 and 2008-09). With the rushing offense and overall defense set to improve, this regression will follow. Regression results: better net yards per drive

Up4. Sack differential: One other factor will play a role in the net yards per drive. That will be the sack differential. Jacksonville allowed 50 sacks in 2012, ranking 30th in the league. Meanwhile, they ranked dead last by racking up only 20 sacks. The 2012 Jaguars became the ninth team since 1940 to have that few sacks and that many sacks allowed (per Pro Football Reference). The list included two franchises in their inaugural season. Of the first eight teams, they held a near-perfect track record for improving on both sides of the ball. Regression results: fewer sacks allowed and more sacks

Up5. Overall record: On the flip side of what happened with the Colts, the Jaguars are subject to improvement, according to Pythagorean win differential. While going 2-14 overall, Jacksonville finished with 3.34 Pythagorean wins. This can be explained by a 2-5 record in one-possession games. It can also be explained by the notion that NFL teams normally really aren’t as bad as 2-14 would suggest. With those two explanations working with the other four factors of regression, it’s safe to say that the Jaguars will enjoy some more favors so they can get a few more wins. Regression results: better W-L record


Tennessee Titans (6-10, 3rd place in 2012)
Tennessee TitansIt wouldn’t be fully accurate to say the Titans experienced the good, the bad and ugly in 2012. However, it’s quite close. Tennessee did have some interesting moments go in its favor. This includes a wild 44-41 overtime win against Detroit, a Thursday night upset against Pittsburgh, and three other wins that featured a combined turnover advantage of 12-0. However, the Titans looked downright terrible at other points in the year. The eyesores included a 38-10 loss in San Diego, a 38-14 loss in Houston, a 30-7 loss in Minnesota, a 51-20 loss to Chicago and a 55-7 loss in Green Bay. Then there’s the 24-10 home loss to Houston that featured a 6-0 turnover disadvantage. The Titans allowed 308 points in their first nine games. With all those huge defensive problems, and the offense in limbo with quarterback Jake Locker still trying to become a franchise-caliber player, the 2013 season could be one to forget in Tennessee if regression won’t help out favorably.

Down1. Return touchdowns: Part of the reason the Titans experienced such a wild year is that they were able to get so many return touchdowns. Tennessee returned four interceptions and one fumble for a touchdown on defense. On special teams, the Titans returned three punts and one kickoff for a touchdown. Only the Bears (10) totaled more return touchdowns. Tennessee got to this point in part due to unusual big-play production in the punt return game. Since 1940, only 23 teams scored at least three punt return touchdowns in a season (per Pro Football Reference). Only the 1951-52 Lions (four touchdowns in 1951 and three in 1952) and the 2006-07 Bears (three touchdowns in 2006 and four in 2007) pulled off this feat in consecutive years. Meanwhile, the five defensive touchdowns seems incredibly unique for a team that nearly 30 points per game. This year, the Titans and Saints became just the second and third teams to allow 450+ points and have at least four interception return touchdowns (per Pro Football Reference). The first team was part of the original Baltimore Colts franchise that disbanded after the 1950 season. Therefore, we have absolutely no precedent for what follows on defense. Also, we can’t tell the regression impact of these returns on a team that was outscored by -141 points. We can only see the impact on special teams. Regression results: fewer return touchdowns and less punt return average

Up2. First-quarter scoring defense: This specific subset can play quite a big role in a team’s success. If the defense struggles in the first quarter, it could set the tone for the whole game. Trailing the ballgame tends to set the tone out of your favor, too. While the Titans allowed 308 points in their first nine games, putting them on pace to set an NFL record for points allowed, they saw a strong portion of it came in the first quarter. Even when the overall scoring defense got better enough to avoid that dubious record, it still set one record of futility. Tennessee allowed 155 points in the first quarter, which surpassed the 137 points allowed from the 1984 Vikings to set an NFL record. With the 9.69 points per game, the Titans allowed at least three more first quarter points per game than any other team, which means we’re talking about at least a 30 percent increase on every other NFL team. That horrific defensive play in the first quarter is something we may never see again. Regression results: improved scoring defense

Up3. Overall scoring defense: Thanks to the horrific start in overall scoring defense and the continuing horrific first-quarter play, the Titans became one of just 31 teams since 1940 to allow at least 450 points (per Football Reference). Only the 1999 Bengals allowed more points after allowing 450+ points the previous year. Considering that the first quarter subset is already going to help, this regression will happen twofold. Regression results: improved scoring defense by at least 20 percent

Up4. Offensive line health: According to Football Outsiders Almanac 2013, the Titans finished with the second-worst Adjusted Games Lost along the offensive line. Center Fernando Velasco was the only player to get on the field for all 16 games, starting each game. Left tackle Michael Roos, by starting all 15 games he played, also enjoyed a generally healthy season. Beyond that, the line didn’t see much continuity. Right guards Leroy Harris and Deuce Lutui had to split the season with eight starts apiece. Meanwhile, left guard Steve Hutchinson and right tackle David Stewart each missed four starts. The offensive line continuity is generally an important factor. However, it’s arguably more important for a running back like Chris Johnson. The speedster doesn’t create his own lanes or holes behind the line of scrimmage, as his playmaking ability comes at the second level. Johnson did enjoy his best yards per attempt average his historic 2009 season, but that was influenced heavily by three touchdowns of 80+ yards. (See what we’re talking about involving the second level?) In fact, 20.68 percent of his total yards came on those three plays. He averaged 3.61 yards on the other 273 carries. That showed in Johnson’s 33rd-ranked DYAR and the team’s 26th-ranked rushing DVOA. Better continuity from injury regression should help that. Regression results: better Johnson DYAR and better team rushing DVOA

Down5. Pythagorean win differential: The Titans went 6-10 despite totaling only 4.78 Pythagorean wins (-1.22 differential), according to our adjustment to Bill James’ formula. This is extremely unusual to see from a losing team that got the benefit of nine return touchdowns. We tried to look for a comparable team to this situation, and the most recent good example we could find was the 2010 Cardinals. The team benefited from 12 return touchdowns, including a ridiculously lucky seven from fumbles. That team only went 5-11 and totaled just 4.38 Pythagorean wins. Take out all return touchdowns (12 for and four against) involved with Cardinals game, and the team totaled somewhere around 2.60 Pythagorean wins. The team got better all around the following year, as expected. It also saw the return touchdowns regress, as all four came from the punt return game during Patrick Peterson’s electric rookie season. Arizona finished the year 8-8. So, with the Titans, we can’t guarantee the team will be worse, especially with the defensive regression. We’ll save any talk of team record for the season predictions. Regression results: closer Pythagorean win differential

Leave a Reply

%d bloggers like this: