06-26-2022, 07:16 PM
(This post was last modified: 06-26-2022, 07:55 PM by RocketMan13. Edited 4 times in total.)
6. Write 600 words or more on something about anything in the league that interests you. It could be related to statistics, a league issue that you take seriously, or a niche part of history that doesn’t fit neatly into either of the above categories. This must be directly related to the league, so don’t wax 600 words about your team’s participation on a Werewolf server or something.
When breaking down last season's performances statistically I found it most interesting to try and determine what statistics could reliably predict winning percentage. The first comparison made was winning percentage as a function of passing yards. This yielded surprising results in the shape of a bell curve (see below) , with the highest winning percentage near the middle. Completion percentage, attempts, completions, and passing touchdowns all had zero correlation.
Equally poor fit factors came when comparing winning percentage to passer rating and number of interceptions (these statistics are highly related), with an R^2 of around 0.1 (see below).
Conclusions are hard to draw for such a limited data set, but the quadratic form of the passing yard data would seem to suggest that a moderate amount of success through the air is optimal for winning games. When examining game conditions it becomes clear that a team that trails is far more likely to try and pass the football, leading to inflated passing stats for some teams. It is equally valid to suggest that those teams who struggled throwing the football had a difficult time winning games. Perhaps an examination of the other half of offensive football will enlighten us.
Data from the run games across the league did not appear to illuminate any trends in the ISFL game. all 14 teams averaged between 4 and 4.9 yards per carry. While the number of attempts varied considerably from Austin's league-low 296 attempts to Philadelphia's league-leading 506, no data trend exists with winning games and running the ball.
Defensive stats started to prove out the same way. No correlation with winning games and tackles, fumbles recovered, interceptions, or even sacks. That is, until we compare winning percentage with passes defensed. Defending balls after they're thrown had a correlation R^2 value of 0.46, the strongest yet.
Both Honolulu and Arizona (the teams with the highest winning percentage) were excellent defending passes in the air. It might come as no surprise that their corners are some of the highest-rated defenders in the league. Conversely, the Colorado Yeti had a horrific year defending the pass (or just in general, as they surrendered 585 points last season, good for last in the league). Thus the first aphorism of football can once again be held up: "Defense wins championships". This also applies to point totals. We compared the correlation to points scored on the season against the correlation to points allowed. Points scored carried an R^2 value of just 0.23 (see below)
Compare that to the scatter plot for points allowed:
Here we see the strongest data set yet. Allowing fewer points was strongly correlated with winning more games, with an R^2 value of 0.722. It would appear based on the data that preventing points is more important than scoring them, and that defending passes in the air appears to be the best way to do so.
If there are any other conclusions that one can draw from a look at the data, it's that the best teams defensively are also the most consistent ones. Season 35 Champions Honolulu only gave up more than 30 points twice all season. Arizona's defense put up similar numbers during the regular season. With consistent high-performing results going deep into the playoffs it's not hard to see how the two teams at the top of the ASFC are likely favorites to hoist this season's Ultimus trophy.
So there we are, GMs of the ISFL. If you want to win the Ultimus trophy the way forward is to invest in your defense, particularly the defensive secondary. Scoring more points is a positive trait to have in your team, but it's not nearly as important as preventing your opponent from scoring.
When breaking down last season's performances statistically I found it most interesting to try and determine what statistics could reliably predict winning percentage. The first comparison made was winning percentage as a function of passing yards. This yielded surprising results in the shape of a bell curve (see below) , with the highest winning percentage near the middle. Completion percentage, attempts, completions, and passing touchdowns all had zero correlation.
Equally poor fit factors came when comparing winning percentage to passer rating and number of interceptions (these statistics are highly related), with an R^2 of around 0.1 (see below).
Conclusions are hard to draw for such a limited data set, but the quadratic form of the passing yard data would seem to suggest that a moderate amount of success through the air is optimal for winning games. When examining game conditions it becomes clear that a team that trails is far more likely to try and pass the football, leading to inflated passing stats for some teams. It is equally valid to suggest that those teams who struggled throwing the football had a difficult time winning games. Perhaps an examination of the other half of offensive football will enlighten us.
Data from the run games across the league did not appear to illuminate any trends in the ISFL game. all 14 teams averaged between 4 and 4.9 yards per carry. While the number of attempts varied considerably from Austin's league-low 296 attempts to Philadelphia's league-leading 506, no data trend exists with winning games and running the ball.
Defensive stats started to prove out the same way. No correlation with winning games and tackles, fumbles recovered, interceptions, or even sacks. That is, until we compare winning percentage with passes defensed. Defending balls after they're thrown had a correlation R^2 value of 0.46, the strongest yet.
Both Honolulu and Arizona (the teams with the highest winning percentage) were excellent defending passes in the air. It might come as no surprise that their corners are some of the highest-rated defenders in the league. Conversely, the Colorado Yeti had a horrific year defending the pass (or just in general, as they surrendered 585 points last season, good for last in the league). Thus the first aphorism of football can once again be held up: "Defense wins championships". This also applies to point totals. We compared the correlation to points scored on the season against the correlation to points allowed. Points scored carried an R^2 value of just 0.23 (see below)
Compare that to the scatter plot for points allowed:
Here we see the strongest data set yet. Allowing fewer points was strongly correlated with winning more games, with an R^2 value of 0.722. It would appear based on the data that preventing points is more important than scoring them, and that defending passes in the air appears to be the best way to do so.
If there are any other conclusions that one can draw from a look at the data, it's that the best teams defensively are also the most consistent ones. Season 35 Champions Honolulu only gave up more than 30 points twice all season. Arizona's defense put up similar numbers during the regular season. With consistent high-performing results going deep into the playoffs it's not hard to see how the two teams at the top of the ASFC are likely favorites to hoist this season's Ultimus trophy.
So there we are, GMs of the ISFL. If you want to win the Ultimus trophy the way forward is to invest in your defense, particularly the defensive secondary. Scoring more points is a positive trait to have in your team, but it's not nearly as important as preventing your opponent from scoring.