Data Mining - How Accurate are First Impressions?
One week is just that, one game for each team. Not much to go on. Not time to panic on underperforming studs like Eli Manning & Maurice Jones-Drew. However, YARDS PER PASS ATTEMPT (YPPA) is one measure that does a good job at predicting a team's offensive success over an entire season (using only one week of data).
Week #1 YPPA is a better predictor than points scored, rushing yards and plenty of other measures. In fact, when you combine YPPA with pass attempt data, history shows you have a solid leading indicator. Again, YPPA is even a better leading indicator than points scored.
THE SECRET FORMULA
On opening day if an NFL team throws the ball 25 or more times and, in the process, is able to average 8.25 or more yards per passing attempt, 96% of the time, that team will end up having a reasonable offense over the following 16 weeks. In fact, 62% of the time, the team will end up being a high-end NFL offense.
Below I'll show you what teams met the 25-8.25 criteria in 2012.
Keep in mind, this is 100% objective analysis (no biases) based on a single, yet very powerful objective measure.
YPPA predicts good things for...
|Robert Griffin III||26||320||12.31|
Note: This study is NOT saying guys like RGIII and Christian Ponder are going to be BETTER THAN Tom Brady and other studs that did not make the cut. Rather it is saying, based on a 100% objective measure which is known to be an OUTSTANDING leading indicator, the odds are good one or more of the above QBs could be a very pleasant fantasy surprise this year.
To learn more about this finding, please read below...
Here is my original YPPA research that lead me to the 8.25, 25 finding....
NFL Week #1 is finally here. Now it's time to watch as much live NFL action as you can. In addition, after the games, you can check out all the highlights. If you're really working overtime, you might examine every line of every NFL Box Score.
After you take all that in, is it easy to spot which squads will have a fairly decent offense and which will have a hard time moving the ball? If an offense only generates a FG in Week #1, that's not a good sign. Clearly, if a team gets shut out on opening day, that's really a bad sign! Or is it? Before you answer, you might want to take a look at this study. The findings could surprise you!
We examined every NFL regular-season game played over a six-year period. We looked at Week #1 offensive output and compared it to rest-of-season production. That gave us 180 case studies to explore. The table below illustrates just a handful of our case studies.
|Case Study||Week #1
|Avg. Total Yards
In addition to the above, we reviewed 15 other variables for each Team-Year combination. Suffice it to say, we did a lot of number crunching. So what did we learn?
First, let's look at those teams that only generated 0-3 points on opening day. We'll pit those inept offenses against teams that really "showed us something" in Week #1. Specifically, offensive units that came out on opening day and put up 41 or more points! Let's review Table-II and see how the Week #1 mavericks and duds performed during the rest of the NFL season.
in Week #1
|Average Total Yards
|41 or more||325|
Guess what? As you can see from the above table, over the remainder of the NFL season, there is barely a difference between those teams that looked grand on opening day and those teams that just couldn't move the ball. It's amazing, but true. That leads up to our first finding.
FINDING #1: Simply looking at extreme Point Production Totals in Week #1 provides little help in forecasting total team offense in Weeks #2-17.
If offensive Point Production isn't a good a predictor, then what is? We've all heard about the importance of a good running game in the NFL. But, is offensive rushing production in Week #1 a solid predictor of team offense?
|Avg. Total Yards
|0 to 55||335|
|56 to 184||334|
|185 or more||337|
|Avg. Total Yards
|0.0 to 3.2||331|
|3.2 to 4.2||
|4.2 to 5.5||337|
|More than 5.5||331|
After consulting Table-III and Table-IV, we see rushing metrics offer us very little predictive capability. Amazingly, teams that rack up 185+ rushing yards on opening day, barely average more total yardage during the rest of the season, than teams that generate 0-55 rushing yards on opening day.
FINDING #2: Don't assume a team displaying a dominate rushing attack in Week #1 will have an outstanding offense the rest of the season. Conversely, if a team doesn't amass big yardage on the ground in Week #1, don't count that offense out.
If Point Production and Rushing stats aren't good guideposts for predicting offensive capabilities, what objective measures are?
FINDING #3: As it turns out, Passing Yards per Passing Attempt (YPPA) is one of the best objective predictors of future offensive production. However, even YPPA has limits.
Week#1 Yards Per
Avg. Total Yards
|0.00 to 4.50||316|
|4.50 to 5.70||326|
|5.70 to 8.25||338|
|more than 8.25||350|
If a team's YPPA is over 8.25 in Week #1, on average, that team will generate 350 yards of offense per game during the rest of the NFL season.
To further enhance forecasting accuracy, we will combine the YPPA statistic with Passing Attempts.
Finding #4: If, on opening day, an NFL team throws the ball 25 or more times and, in the process, is able average 8.25 or more yards per passing attempt, you should take note. It turns out, 96% of the time, that team will end up having a reasonable offense over the following 16 weeks. In fact, 62% of the time, the team will end up being a high end NFL offense.
In addition to identifying teams that should do reasonably well, the YPPA statistic can be used to find teams that will have problems on offense.
To find teams most likely to struggle on offense, locate ones that have a YPPA less than 5.1 and gained 0-80 rushing yards in Week #1. Usually, 3 out of 4 teams meeting the 5.1-80 criteria will perform below NFL norms. In addition, they will usually only average about 300-315 yards of offense per game the rest of the season.
Some Ways to Exploit this Research:
As Week #1 ends, that's the time to put these principles into practice. Think about these findings and see how they might work in your league.
First, look for some waiver wire pickups. Using these findings, pick up an available defense that will be playing a weak offense several times in the coming weeks.
Also, look for available player talent on a team that shows promise according to these guidelines. This research suggests opportunities exist to find sleepers and avoid duds.
In addition, look for value and try to trade for quality talent on a team that may have struggled based on traditional stats in Week #1, however, had a decent YPPA.
Also consider how you might exploit a 'big-name' player on your roster, if his team is likely to struggle on offense based on these findings. In making the trade, be sure to get quality for your 'big-name' talent.
Finally, after you use this research to fit your situation, be sure to strike while the iron is hot!