Link
Big, big update today to expand upon my last two posts. First, as a proof of concept, we looked at passing stats for last season in DSFL and NSFL. Then, we went backwards chronologically to cover the history of both leagues from a passing perspective. Now, I’ve gone laterally to give you every stat that exists in the index! This involved learning how to do some fun web scraping with tables, but it was well worth the time investment to get this big database stored. There are way too many metrics to go through them one-by-one, so I’ll just cover something interesting from each category.
Passing - This has been covered in a previous post, so we’ll skip that one.
Rushing – Contrary to what we’ve seen with passing data, rushing in the DSFL is not so dissimiliar between what we see in the NFL. It’s generally evident that more rushing equals more wins. This makes sense—teams tend to run the ball when they are winning at the end of games, so rushing attempts and rushing yards tend to be correlated with success. However (and I think this speaks volumes to how passing-oriented the league is), there is no significant correlation between rushing yards and wins in the NSFL! It looks like this is slowly starting to change over time though, as you can filter by more recent seasons to see an increasingly positive impact.
[div align=\\\"center\\\"]![[Image: zsHtuIF.png]](https://i.imgur.com/zsHtuIF.png)
[/div]
Defense – This one was easy. The importance of sacks is practically jumping off the screen at me, especially in the NSFL. The regression coefficient for sacks in the NSFL is about ~0.1—in other words, for every 10 sacks that your team accrues compared to another team in the league, the data suggests that you should have an additional win. For comparison, the coefficient for fumbles and interceptions is about ~0.14. This suggests that recording three sacks has relatively the same value as forcing two turnovers. I found this particularly astounding for a couple of reasons—1) turnovers are a guaranteed drive-killer by definition, whereas an offensive can theoretically continue after a sack, and 2) turnovers will generally result in much better field position for the defense than a drive that includes a sack.
[div align=\\\"center\\\"]![[Image: sNJHX1W.png]](https://i.imgur.com/sNJHX1W.png)
[/div]
Kicking – There’s not a lot to be learned here, but it is a little humorous that the rate at which you make field goals doesn’t matter all that much. The amount of field goals and extra points you convert are indicators for winning (because they directly correspond to points being added to the scoreboard), but overall, the kicking game is more of a measure of health for an offense. If you attempted a lot of field goals, it simply means that your offense is good and has little bearing on the quality of your kicker.
[div align=\\\"center\\\"]![[Image: OUYGSAL.png]](https://i.imgur.com/OUYGSAL.png)
[/div]
Punting – Interestingly, a punter’s longest punt for a season is a minor predictor of winning. After pondering on this for a bit, I suspect this be because the sim is exporting gross punting yards—how else can you explain seasons with 70-80 yards punts? Because of this, I’m inclined to say that a punter’s longest punt is directly correlated with their punt power and therefore the punter’s overall skill.
[div align=\\\"center\\\"]![[Image: umyFJXa.png]](https://i.imgur.com/umyFJXa.png)
[/div]
Offensive Line – There are only two stats for linemen and we covered sacks in the defense blurb, so let’s talk about pancakes. They matter! I’m not sure how pancakes are treated in the sim—does it just take the defensive player out of the play entirely? I was a bit surprised to see that pancakes matter a bit more in the NSFL than in the DSFL—my intuition was that offensive linemen are quite rare in the DSFL, so I would have expected linemen performance to be a bit more differentiated in the lower league.
[div align=\\\"center\\\"]![[Image: pDz5Xit.png]](https://i.imgur.com/pDz5Xit.png)
[/div]
Special Teams – Finally, special teams. Unfortunately we have to essentially just discard kick return and punt return totals. Lots of kick return yardage is bad for your team because you’re giving up lots of points, and lots of punt return yardage is good for your team because you’re not giving up lots of points. So with that sorted, let’s discuss kick return average, particularly in the DSFL. Here, the regression coefficient is ~0.35—or, for every three yards your team’s kick return average increases, wins increase by one game. That seems like a massive improvement for a three-yard difference! I’m curious to see if this is solely due to roster composition, or if there is something unique happening in the sim on kick returns where teams could be able to exploit some kind of advantage.
[div align=\\\"center\\\"]![[Image: 0prX2Nl.png]](https://i.imgur.com/0prX2Nl.png)
[/div]
Lots of thoughts about where to go from here (individual stats, TPE, sim stuff), so let me know if there are any ideas out there!
Big, big update today to expand upon my last two posts. First, as a proof of concept, we looked at passing stats for last season in DSFL and NSFL. Then, we went backwards chronologically to cover the history of both leagues from a passing perspective. Now, I’ve gone laterally to give you every stat that exists in the index! This involved learning how to do some fun web scraping with tables, but it was well worth the time investment to get this big database stored. There are way too many metrics to go through them one-by-one, so I’ll just cover something interesting from each category.
Passing - This has been covered in a previous post, so we’ll skip that one.
Rushing – Contrary to what we’ve seen with passing data, rushing in the DSFL is not so dissimiliar between what we see in the NFL. It’s generally evident that more rushing equals more wins. This makes sense—teams tend to run the ball when they are winning at the end of games, so rushing attempts and rushing yards tend to be correlated with success. However (and I think this speaks volumes to how passing-oriented the league is), there is no significant correlation between rushing yards and wins in the NSFL! It looks like this is slowly starting to change over time though, as you can filter by more recent seasons to see an increasingly positive impact.
[div align=\\\"center\\\"]
[/div]Defense – This one was easy. The importance of sacks is practically jumping off the screen at me, especially in the NSFL. The regression coefficient for sacks in the NSFL is about ~0.1—in other words, for every 10 sacks that your team accrues compared to another team in the league, the data suggests that you should have an additional win. For comparison, the coefficient for fumbles and interceptions is about ~0.14. This suggests that recording three sacks has relatively the same value as forcing two turnovers. I found this particularly astounding for a couple of reasons—1) turnovers are a guaranteed drive-killer by definition, whereas an offensive can theoretically continue after a sack, and 2) turnovers will generally result in much better field position for the defense than a drive that includes a sack.
[div align=\\\"center\\\"]
[/div]Kicking – There’s not a lot to be learned here, but it is a little humorous that the rate at which you make field goals doesn’t matter all that much. The amount of field goals and extra points you convert are indicators for winning (because they directly correspond to points being added to the scoreboard), but overall, the kicking game is more of a measure of health for an offense. If you attempted a lot of field goals, it simply means that your offense is good and has little bearing on the quality of your kicker.
[div align=\\\"center\\\"]
[/div]Punting – Interestingly, a punter’s longest punt for a season is a minor predictor of winning. After pondering on this for a bit, I suspect this be because the sim is exporting gross punting yards—how else can you explain seasons with 70-80 yards punts? Because of this, I’m inclined to say that a punter’s longest punt is directly correlated with their punt power and therefore the punter’s overall skill.
[div align=\\\"center\\\"]
[/div]Offensive Line – There are only two stats for linemen and we covered sacks in the defense blurb, so let’s talk about pancakes. They matter! I’m not sure how pancakes are treated in the sim—does it just take the defensive player out of the play entirely? I was a bit surprised to see that pancakes matter a bit more in the NSFL than in the DSFL—my intuition was that offensive linemen are quite rare in the DSFL, so I would have expected linemen performance to be a bit more differentiated in the lower league.
[div align=\\\"center\\\"]
[/div]Special Teams – Finally, special teams. Unfortunately we have to essentially just discard kick return and punt return totals. Lots of kick return yardage is bad for your team because you’re giving up lots of points, and lots of punt return yardage is good for your team because you’re not giving up lots of points. So with that sorted, let’s discuss kick return average, particularly in the DSFL. Here, the regression coefficient is ~0.35—or, for every three yards your team’s kick return average increases, wins increase by one game. That seems like a massive improvement for a three-yard difference! I’m curious to see if this is solely due to roster composition, or if there is something unique happening in the sim on kick returns where teams could be able to exploit some kind of advantage.
[div align=\\\"center\\\"]
[/div]Lots of thoughts about where to go from here (individual stats, TPE, sim stuff), so let me know if there are any ideas out there!
![[Image: Maglubiyet.gif]](https://sig.grumpybumpers.com/host/Maglubiyet.gif)
![[Image: TRRHgeA.png]](https://i.imgur.com/TRRHgeA.png)
![[Image: tTMB4jP.jpg]](https://imgur.com/tTMB4jP.jpg)
![[Image: 55457_s.gif]](https://signavatar.com/55457_s.gif)
