09-10-2023, 09:37 PM
(This post was last modified: 09-12-2023, 10:36 PM by xenosthelegend. Edited 1 time in total.)
Hello everyone! I have been tracking data on penalties all season long in the ISFL and wanted to take some time to share my findings with all of you.
Do penalties even matter?
Penalties can be one of the most frustrating things to occur while watching a live stream. Always seeming to come at the most inopportune time, there are times it seems your team is playing the refs as well as the opponent! But every team deals with it right? Does the team that registers more penalties decrease their chance of losing? Or is it just a nuisance where we remember that one random time a penalty knocked us out of field goal range?
To answer these questions, I calculated "penalty differential" and "penalty yardage differential" and compared that to the outcome of the game. Just a quick summary, penalty differential is similar to turnover differential in that we take the losing team's number of penalties and subtract the winning team's number of penalties. The same was done with penalty yardage to create a penalty yardage differential. Below you will see every regular season game from season 43 plotted with penalty differential vs. point differential along with penalty yardage differential vs. point differential.
Just by looking at the data, it appears that maybe a team wins more if they have a positive penalty and penalty yardage differential but there doesn't appear to be a strong trend as teams win all the time with a negative differential. So a deeper analysis was needed.
First, let's take a look at Penalty Differential. During this season, the team with fewer penalties (a positive penalty differential) went 64-40 for a winning percentage of 0.615. To see if this is statistically significant, we can perform an inference test (specifically a one-sample z-test for a proportion for those you that want to know). This analysis takes the assumption that penalty differential doesn't matter (that the win percentage of a team with a positive penalty differential should .500) and see how likely our results would be if that was the case. Running this test, we get a p-value of .0093. This means that, if penalty differential doesn't actually impact which team wins, the winning percentage of 0.615 would only be surpassed 0.93% of the time. Or in other words, penalty differential does seem to influence which team wins!
One would expect a similar result with penalty yardage differential and that is indeed what happens. During season 43, teams with fewer penalty yards (a positive penalty yardage differential) went 65-42-1 for a winning percentage of .606. Using the same inference test, gave a p-value of .0131. Meaning that if penalty yardage differential didn't matter, we would surpass the win percentage of .606 only 1.31% of the time. This confirms the idea that penalty yardage differential also can be used as a metric to try and maximize to increase your teams chance of winning.
Ok so penalties do matter, but how much?
This question is a little harder to answer. However, let's take look at some commonalities that we saw this season. When posting a positive penalty differential the only teams without a winning record where New Orleans (0-5), Chicago (0-5), Sarasota (2-4), and Austin (5-5). Of those teams, New Orleans and Chicago would be hard pressed to have a winning record under any metric. Austin was at .500 so it barely counts. That leaves just one outlier team, Sarasota, that I am not sure how to explain. (Maybe the sim doesn't like them? Though results I show later may show the opposite!) Similar results were found as these four teams were the only ones without winning records with a positive penalty yardage differential as well. Below are each teams' record based on Penalty Differential and Penalty Yardage Differential.
However, maybe penalty differential isn't the best way to breakdown the importance of penalties to winning games. My documentation so far has not accounted for the difference between a -1 differential and a -10 differential. So instead let us take a closer look at some penalty volume stats.
Maybe penalty volume will give us a better indicator of success.
When recording the number of penalties for each team for each game, you can find that the a team had on average 4.53 penalties per game for a total of 37.04 yards. I went ahead and calculated the number of penalties and penalty yards for each winner and loser and found the record for each value. By splitting the results into four approximately equal groups, you can see a very obvious trend.
Teams that had two or fewer penalties: 40-16-0, .714 Winning Percentage
Teams that had three or four penalties: 29-31-0, .483 Winning Percentage
Teams that had five or six penalties: 25-31-0, .446 Winning Percentage
Teams that had seven or more penalties: 17-33-2, .346 Winning Percentage
A similar phenomenon occurs when looking at penalty yardage.
Teams that had 19 or less penalty yards: 35-16-0, .686 Winning Percentage
Teams that had between 20 and 39 penalty yards: 38-37-0, .507 Winning Percentage
Teams that had between 40 and 59 penalty yards: 26-30-1, .465 Winning Percentage
Teams that had 60 or more penalty yards: 12-28-1, .305 Winning Percentage
But this didn't only occur on a per game basis. If we look at the team totals for penalties on the season, we see a link between penalties and wins.
Here is a look at the team by team totals in descending order based on total penalty yardage on the season.
- 48 Penalties for 410 yards - Team Record: 9-7
- 52 Penalties for 421 yards - Team Record: 14-2
- 52 Penalties for 468 yards - Team Record: 8-8
- 62 Penalties for 500 yards - Team Record: 12-4
- 67 Penalties for 505 yards - Team Record: 7-8-1
- 66 Penalties for 510 yards - Team Record: 13-2-1
- 70 Penalties for 563 yards - Team Record: 7-9
- 69 Penalties for 591 yards - Team Record: 8-8
- 82 Penalties for 616 yards - Team Record: 8-8
- 77 Penalties for 665 yards - Team Record: 7-9
- 82 Penalties for 693 yards - Team Record: 8-8
- 88 Penalties for 734 yards - Team Record: 6-10
- 95 Penalties for 735 yards - Team Record: 3-13
- 105 Penalties for 886 yards - Team Record: 1-15
Maybe the correlation is difficult to see in a list. Don't worry, I have the information graphed with a trendline and everything for a more visual explanation!
First up, comparing total penalties to win totals.
With this you can see a very clear negative correlation between the number of penalties and the win totals. To get a little bit more nerdy with this, the exact regression line equation is Wins = 19.8 - 0.163 * (Number of Penalties). The coefficient of determination (or R^2) was a fairly solid .595. With this information, we can actually put a (negative) value on each penalty as it would take away on average .163 wins. Of course not every team is on the regression line. And here is where we maybe see the gift that the sim gods have given to each team. The higher above the line a team is, the more they outperformed the league average when it comes to penalties (I would argue that they are sim-blessed!), while the further below a team is from the line, the more they underperformed compared to the league average in regards to penalties (should we say sim-cursed?).
But the regression was even better when looking at penalty yardage! (And why I ordered the list above by penalty yards instead of total penalties).
Comparing total penalty yardage to wins.
Here the regression equation is Wins = 20.3 - 0.0208 * (Penalty Yards). The coefficient of determination (R^2) was even higher at .647. This again gives us a (negative) value for each penalty yard. It even would differentiate between a 5 yard penalty and a 15 yard penalty. Looking at this chart, it comes as no surprise that 5 of the 6 playoff teams are either right on or above the trendline! Luck is never a bad thing!
So more penalties is bad, but can you really do anything about it?
The short answer to this is: Yes, most definitely!
Below is a list of all players that accumulated 9 or more penalties on the season. Why 9 or more? These are the players that averaged over half a penalty a game.
Shucks McShuckle - OL - 11 Penalties for 86 yards
Demarcus Cole - DT - 11 Penalties for 79 yards
Liath Squirrel - LB - 11 Penalties for 77 yards
Great Khali - OL - 11 Penalties for 70 yards
Daron Arnold - DE - 10 Penalties for 100 yards
Sack Sparrow - DT - 10 Penalties for 100 yards
Wälter Gunther - LB - 10 Penalties for 83 yards
Burnt Toast - OL - 10 Penalties for 80 yards
Bread Bowl - TE - 10 Penalties for 80 yards
Teenie Hodges - WR - 10 Penalties for 77 yards
Donatello McTurtle - OL - 10 Penalties for 69 yards
Ace Anderson - RB - 10 Penalties for 69 yards
Joshua Mabry - OL - 9 Penalties for 90 yards
Ignatious Cleetington - LB - 9 Penalties for 81 yards
Eli Prince - WR - 9 Penalties for 75 yards
Hawk Red-Tailed (BOT) - OL - 9 Penalties for 70 yards
Tank Commander (BOT) - OL - 9 Penalties for 68 yards
Bubba Hog - LB - 9 Penalties for 63 yards
So what did all of these players have in common to make it onto this list? All of these players had an Intelligence Stat at or below 60. So does the sim just hate you? Of course not! You just seem to have forgotten about that Intelligence stat. I know it doesn't seem that important but when the sim does stupid things, is it really the sim?
If you read this all the way to the end, thank you! I know this was a long one.
Do penalties even matter?
Penalties can be one of the most frustrating things to occur while watching a live stream. Always seeming to come at the most inopportune time, there are times it seems your team is playing the refs as well as the opponent! But every team deals with it right? Does the team that registers more penalties decrease their chance of losing? Or is it just a nuisance where we remember that one random time a penalty knocked us out of field goal range?
To answer these questions, I calculated "penalty differential" and "penalty yardage differential" and compared that to the outcome of the game. Just a quick summary, penalty differential is similar to turnover differential in that we take the losing team's number of penalties and subtract the winning team's number of penalties. The same was done with penalty yardage to create a penalty yardage differential. Below you will see every regular season game from season 43 plotted with penalty differential vs. point differential along with penalty yardage differential vs. point differential.
Just by looking at the data, it appears that maybe a team wins more if they have a positive penalty and penalty yardage differential but there doesn't appear to be a strong trend as teams win all the time with a negative differential. So a deeper analysis was needed.
First, let's take a look at Penalty Differential. During this season, the team with fewer penalties (a positive penalty differential) went 64-40 for a winning percentage of 0.615. To see if this is statistically significant, we can perform an inference test (specifically a one-sample z-test for a proportion for those you that want to know). This analysis takes the assumption that penalty differential doesn't matter (that the win percentage of a team with a positive penalty differential should .500) and see how likely our results would be if that was the case. Running this test, we get a p-value of .0093. This means that, if penalty differential doesn't actually impact which team wins, the winning percentage of 0.615 would only be surpassed 0.93% of the time. Or in other words, penalty differential does seem to influence which team wins!
One would expect a similar result with penalty yardage differential and that is indeed what happens. During season 43, teams with fewer penalty yards (a positive penalty yardage differential) went 65-42-1 for a winning percentage of .606. Using the same inference test, gave a p-value of .0131. Meaning that if penalty yardage differential didn't matter, we would surpass the win percentage of .606 only 1.31% of the time. This confirms the idea that penalty yardage differential also can be used as a metric to try and maximize to increase your teams chance of winning.
Ok so penalties do matter, but how much?
This question is a little harder to answer. However, let's take look at some commonalities that we saw this season. When posting a positive penalty differential the only teams without a winning record where New Orleans (0-5), Chicago (0-5), Sarasota (2-4), and Austin (5-5). Of those teams, New Orleans and Chicago would be hard pressed to have a winning record under any metric. Austin was at .500 so it barely counts. That leaves just one outlier team, Sarasota, that I am not sure how to explain. (Maybe the sim doesn't like them? Though results I show later may show the opposite!) Similar results were found as these four teams were the only ones without winning records with a positive penalty yardage differential as well. Below are each teams' record based on Penalty Differential and Penalty Yardage Differential.
However, maybe penalty differential isn't the best way to breakdown the importance of penalties to winning games. My documentation so far has not accounted for the difference between a -1 differential and a -10 differential. So instead let us take a closer look at some penalty volume stats.
Maybe penalty volume will give us a better indicator of success.
When recording the number of penalties for each team for each game, you can find that the a team had on average 4.53 penalties per game for a total of 37.04 yards. I went ahead and calculated the number of penalties and penalty yards for each winner and loser and found the record for each value. By splitting the results into four approximately equal groups, you can see a very obvious trend.
Teams that had two or fewer penalties: 40-16-0, .714 Winning Percentage
Teams that had three or four penalties: 29-31-0, .483 Winning Percentage
Teams that had five or six penalties: 25-31-0, .446 Winning Percentage
Teams that had seven or more penalties: 17-33-2, .346 Winning Percentage
A similar phenomenon occurs when looking at penalty yardage.
Teams that had 19 or less penalty yards: 35-16-0, .686 Winning Percentage
Teams that had between 20 and 39 penalty yards: 38-37-0, .507 Winning Percentage
Teams that had between 40 and 59 penalty yards: 26-30-1, .465 Winning Percentage
Teams that had 60 or more penalty yards: 12-28-1, .305 Winning Percentage
But this didn't only occur on a per game basis. If we look at the team totals for penalties on the season, we see a link between penalties and wins.
Here is a look at the team by team totals in descending order based on total penalty yardage on the season.
- 48 Penalties for 410 yards - Team Record: 9-7
- 52 Penalties for 421 yards - Team Record: 14-2
- 52 Penalties for 468 yards - Team Record: 8-8
- 62 Penalties for 500 yards - Team Record: 12-4
- 67 Penalties for 505 yards - Team Record: 7-8-1
- 66 Penalties for 510 yards - Team Record: 13-2-1
- 70 Penalties for 563 yards - Team Record: 7-9
- 69 Penalties for 591 yards - Team Record: 8-8
- 82 Penalties for 616 yards - Team Record: 8-8
- 77 Penalties for 665 yards - Team Record: 7-9
- 82 Penalties for 693 yards - Team Record: 8-8
- 88 Penalties for 734 yards - Team Record: 6-10
- 95 Penalties for 735 yards - Team Record: 3-13
- 105 Penalties for 886 yards - Team Record: 1-15
Maybe the correlation is difficult to see in a list. Don't worry, I have the information graphed with a trendline and everything for a more visual explanation!
First up, comparing total penalties to win totals.
With this you can see a very clear negative correlation between the number of penalties and the win totals. To get a little bit more nerdy with this, the exact regression line equation is Wins = 19.8 - 0.163 * (Number of Penalties). The coefficient of determination (or R^2) was a fairly solid .595. With this information, we can actually put a (negative) value on each penalty as it would take away on average .163 wins. Of course not every team is on the regression line. And here is where we maybe see the gift that the sim gods have given to each team. The higher above the line a team is, the more they outperformed the league average when it comes to penalties (I would argue that they are sim-blessed!), while the further below a team is from the line, the more they underperformed compared to the league average in regards to penalties (should we say sim-cursed?).
But the regression was even better when looking at penalty yardage! (And why I ordered the list above by penalty yards instead of total penalties).
Comparing total penalty yardage to wins.
Here the regression equation is Wins = 20.3 - 0.0208 * (Penalty Yards). The coefficient of determination (R^2) was even higher at .647. This again gives us a (negative) value for each penalty yard. It even would differentiate between a 5 yard penalty and a 15 yard penalty. Looking at this chart, it comes as no surprise that 5 of the 6 playoff teams are either right on or above the trendline! Luck is never a bad thing!
So more penalties is bad, but can you really do anything about it?
The short answer to this is: Yes, most definitely!
Below is a list of all players that accumulated 9 or more penalties on the season. Why 9 or more? These are the players that averaged over half a penalty a game.
Shucks McShuckle - OL - 11 Penalties for 86 yards
Demarcus Cole - DT - 11 Penalties for 79 yards
Liath Squirrel - LB - 11 Penalties for 77 yards
Great Khali - OL - 11 Penalties for 70 yards
Daron Arnold - DE - 10 Penalties for 100 yards
Sack Sparrow - DT - 10 Penalties for 100 yards
Wälter Gunther - LB - 10 Penalties for 83 yards
Burnt Toast - OL - 10 Penalties for 80 yards
Bread Bowl - TE - 10 Penalties for 80 yards
Teenie Hodges - WR - 10 Penalties for 77 yards
Donatello McTurtle - OL - 10 Penalties for 69 yards
Ace Anderson - RB - 10 Penalties for 69 yards
Joshua Mabry - OL - 9 Penalties for 90 yards
Ignatious Cleetington - LB - 9 Penalties for 81 yards
Eli Prince - WR - 9 Penalties for 75 yards
Hawk Red-Tailed (BOT) - OL - 9 Penalties for 70 yards
Tank Commander (BOT) - OL - 9 Penalties for 68 yards
Bubba Hog - LB - 9 Penalties for 63 yards
So what did all of these players have in common to make it onto this list? All of these players had an Intelligence Stat at or below 60. So does the sim just hate you? Of course not! You just seem to have forgotten about that Intelligence stat. I know it doesn't seem that important but when the sim does stupid things, is it really the sim?
If you read this all the way to the end, thank you! I know this was a long one.
DSFL S43 Draft - Round 4 Pick 30
ISFL S44 Draft - Round 1 Pick 12
S43 DSFL Pro Bowl
S44 DSFL Pro Bowl
S45 All Rookie Team
Gemini Media Awards
12 Gemini Nominations
Best Article (Limited) - S43 "Do Penalties Really Matter?"
Best Author - S43, S44