enCYCLONEpedia: Relating stats to wins
By Kirk Haaland: Who has been a contributor to Cyclone Fanatic for more than two years and has his own Iowa State Football & Basketball website, www.encyclonepedia.com. There, you can read more of his Cyclone related content including historical data and current day coverage. Follow Kirk on Twitter: @khaal53
I have a theory.
It is far from original or groundbreaking, but the thought has slowly solidified in my mind this season. The Iowa State offense is the weak link of the football team, the offensive line is the weak link of the offense and the rushing game is where that is most apparent.
In effect, the offense is struggling in both the passing and rushing games but the limited ability of the running game is the crux of the problem.
A small snippet into that point can be seen when comparing the average yards per rush on the season from 2009-2011. In 2009, ISU had a very good offensive line that was headlined by Kelechi Osemele, Ben LaMaak, Reggie Stephens, and Scott Haughton. KO is the only remaining player from that group and outside of Brayden Burris who was at right tackle, but is now out with a broken leg. The current group is very young.
Here is the steady decline in the team rush yards per carry category:
2009: 4.53 2010: 3.78 2011: 3.88 (Yes, I know 3.88 is a larger number than 3.78 but there are still five tough conference games remaining that will likely bring that number down further—especially with no Shontrelle Johnson.)
But the struggling ground game for ISU is only part of the problem.
The Cyclone defense has been gashed in recent weeks by the ground games of its opponents and is giving up 4.59 yards per carry. By comparing the first seven games of the 2011 season to full seasons of every year back to 1943, that is the 21stworst effort to date. At least for this season, I think many of us would conclude that the defense is performing okay and would be much better off if it weren’t for the inconsistent offense.
So, what does this all mean?
It got me to wondering, is the anemic offense to blame for the lack of wins, is it the poor defense that is the root of the problems, or is it the combination of the two? I took a look back at the Iowa State record book and sought after an answer.
Warning: Non-math geeks read carefully…
To find answers I used everyone’s old friend, the correlation coefficient. The correlation coefficient is a statistical tool that calculates the relationship between two variables. It calculates “r” and runs on a scale from -1 to 1. As the number approaches 1 the likelihood for a statistical correlation increases and as it approaches -1 the likelihood for an inverse correlation increases. The closer the number is to zero, it is less likely that the numbers have a statistical link.
Generally, if the value falls between -0.3 and 0.3 the relationship is considered weak. From 0.3 to 0.7 or -0.3 to -0.7 the relationship is considered moderate. Anything from 0.7 to 1 or -0.7 to -1 is considered a strong relationship.
If you want to read more to get the mathematical explanations and further detail into this read here.
I went back to 1943 (the first year that I had complete data to conduct the analysis) and computed the raw yardage numbers for passing and rushing for both Iowa State and the opponents played in each season. Total rush yards, rush attempts, rush yards per carry, pass yards, pass yards per attempt, total yards per play, and the margins of each stat for ISU when compared to its opponents. I then used the correlation coefficient calculation and compared those statistical categories to ISU’s win percentage in that year to find the stats that were most strongly linked to a greater win percentage.
Other than going back to 1943, I also used a few other seasons as benchmarks to compare the data. I went back to 1970 to include the Majors and Bruce successful years in the analysis and to offer up a near halfway point between now and 1943. I also used the years just back to 2000 to get the more recent numbers while excluding the total ineptitude of the Cyclones from the Walden years and early McCarney years that may skew the data.
In every case the three strongest statistical links to Iowa State’s winning percentage in a given season were (in no particular order): total yards per play margin, pass yards per attempt margin, and rush yards per carry margin. That is perhaps not a shock. Obviously, I didn’t take the time to run the analysis on points scored or allowed…so spare me the jokes.
Check the chart below that details the three season ranges I used for the analysis and the correlation coefficient for the three factors mentioned above:
From this we can gather a few things…The main one being that total yards per play margin typically has the strongest link to season long win percentage of any major yardage stat that is compiled. Perhaps that is or isn’t newsworthy to you because it is somewhat obvious. However, let’s use that conclusion to compare to how the 2011 season is shaping up.
Thus far in to the 2011 season, the Cyclone defense is giving up 456 yards per game and 5.68 yards per play, which places it at as the 112thbest defense in the NCAA. The ISU offense is averaging 370 yards per game and 4.81 yards per play, which ranks it as the 79thbest offense in the NCAA.
Further, for the comparison for the range of seasons from 1943-2010, this year’s Cyclones (again, comparing full seasons to just seven games for the 2011 squad) have the 27thworst rush yard per carry margin (-0.71 yards per carry). The best record for a team that had worse numbers was the 1985 team that went 5-6. The 2011 team also has the 17thworst pass yards per attempt margin (-1.63 yards per pass attempt) and the ISU team with the best record that had worse numbers was the 1973 team that was 4-7. For total yards per play margin, this year’s team is 24thworst (-0.87 yards per play) and the 1985 team is again the one with worse numbers and a better record.
The story is nearly identical for the range of years from 1970-2010, in fact the only thing that changes is the rankings of the team’s numbers in each season. Compared to the past 41 seasons the 2011 team has the 18thworst rush yards per carry margin, the eighth worst pass yards per attempt margin, and the 16thworst total yards per play margin. In each category the same teams with worse numbers show up when looking at the “best record with a worse ranking”.
As mentioned earlier, the 2011 Cyclones are currently averaging 3.88 yards per carry, to go along with that they are averaging 5.72 yards per pass attempt, and 4.81 yards per play. That is a stark contrast to the opponents of the 112thranked ISU defense that is giving up on average 4.59 yards per rush, 7.35 yards per pass, and 5.68 yards per play. Those numbers will surely only separate further for the remainder of the season as only Kansas appears to offer an opportunity to slow the widening of that gap—with an outside shot at Kansas State as they normally do not put up gaudy numbers.
There have been numerous instances this year with the offense backing the defense into a corner by recording three and outs, turnovers, and not converting drives into points. It all starts with the lack of consistency from the ground game. The Cyclones are likely on pace to have the worst rushing yard per carry average since Gene Chizik’s arrival in 2007 when they averaged 3.14 yards per carry.
Considering the strength of correlation between the total yards per play margin and the season’s winning percentage something has to improve for Iowa State. Whether it be the offense making more plays, the defense limiting the opponent’s big plays, or a little bit of both but the evidence is clear that right now, Iowa State has some work to do if it is to win more games in 2011 and beyond.