03-26-2024, 07:31 PM
(This post was last modified: 03-28-2024, 10:19 PM by wetwilleh. Edited 1 time in total.)
Hello and welcome to this statistical analysis of the ISFL and DSFL, using season 46 data. I will be analyzing the most important position in all of sports. Of course, I am talking about Left Guards Quarterbacks. I am slightly upset that it is a 2x media week, as I had planned on tackling defense next. Don’t get me wrong, I am using that 2x bonus as much as possible, and another analysis may be incoming (open to suggestions). I am open to suggestions about which position group to tackle next. I was going to tackle defense next, but it’s too hard it seems unrealistic as despite what these massive statistical analyses may suggest, I do almost have a life.
First, the data. I compiled the stats and rosters from every team in both the ISFL and DSFL after the S46 season ended.
ASSUMPTIONS:
1. Data
I assumed season 46 (my only full season) was a typical season with typical results. Outliers are, of course, expected, but this is the base assumption that the data is normal.
I am using the stats from both indices, assuming that they are canon and without errors. This does not include the pre- or post- season. In this analysis, I will be analyzing Passing Yards, Passing TD, Completion %, QB Rating, and Interceptions.
2. Ratings
I assumed that the ratings that players had at the end of the season were used for the entirety of the season. I am aware that this is only true for DSFL capped players. However, this article takes long enough without factoring in game-by-game analysis, and I don’t even know how to do that this assumption helps simplify the data enough to analyze.
3. Playbooks
I am a simple TE, who was barely included into my DSFL team’s draft war room (and I suspect it was by default). I don’t even know how many playbooks there are, let alone their effect on each player (hint this is a special clue for later).
4. Player Intangibles
Player Archetypes and Traits are not a factor in this analysis. Running QBs will likely have lower passing numbers because they scramble more, but this is not being factored into my analysis.
6. Defenses
Player game-by-game performances relative to the opposing defenses are not factored into this analysis.
7. Statistical Conclusions
Any chart with a R-squared value of .50 or more is enough to be statistically significant. Higher slopes on the trendlines suggests more effectiveness as an attribute toward a team’s passing game. I am using linear regression models to draw conclusions.
ANALYSIS:
Similar to my receiver analysis, I decided to run 3 sets of data: DSFL, ISFL, and Overall. While this is a useful framework (by giving valuable data on low attributes), it does struggle in that I am not factoring in either game-by-game performances nor the individual defenses they play. This means that the DSFL traits will seem stronger against a defense which is likely one of: bots, IA players, and low attribute totals.
First, I checked to confirm that Overall rating had an effect on all the stats.
There are several intriguing conclusions that this data suggests. It suggests a positive correlation between overall and Yards, TD, and Completion Percentage, which intuitively makes sense. However, it also suggests a negative correlation in Rating and a positive correlation in Interceptions (meaning that higher overall players get lower Ratings and more interceptions). To confirm this, I ran the numbers for the ISFL and DSFL.
ISFL:
DSFL:
As you can see, stats are confusing. Fear not, reader, for I’m going to expedite the word portion (it’s a pain to write, read, and probably grade, and I plan on eventually doing one of these articles for every position).
ARM:
A better arm leads to yards and TD. This is true in both leagues, and is thankfully one of the only conclusions I feel confident in drawing. The data suggests that at low levels, a worse arm is better for Rating and Completion Percentage. However, since this trend reverses in the ISFL, this is likely due to either the new create QBs in the league with minimal TPE which are getting good results, or that better defenses demand higher arm talent to be able to beat them (while you can sneak by on other traits in the DFSL)
ACCURACY:
The conclusion I am drawing here is that new players tend to prioritize Accuracy over Arm, which works early for all stats. It is positive at all levels for gaining yards, and helps the DSFL throw more accurately and more TD. Besides that, not much going on here.
INTELLIGENCE:
At least Intelligence as a positive correlation in everything except interceptions. Intelligence is definitely a worthwhile stat to take for all QBs. Intelligence definitely makes a QB more effective at their job.
RESULTS:
If you care about yardage and/or TD, Arm > Intelligence > Accuracy. For Completion percentage, accuracy is helpful in the DSFL, but Arm seems to get you over the top in the ISFL. However, this is only an approximation, as the sample size and R-squared values are not large enough to draw accurate conclusions.
This statistical analysis helps underline one of the core tenets of statistics: correlation does not imply causation. A base understanding of the data implies that getting higher attributes negatively impacts one’s Interception rate and passer rating. However, this does not make sense intuitively, so there are in fact, at least two possibilities.
A: Higher TPE players throw more picks and are worse
B: Higher TPE players play more, which gives them more minutes/plays, which inherently leads to more picks. Players with less TPE will play in relief in blowouts, and will have more conservative play calling to benefit their numbers.
I know which scenario I believe in. And it’s not A.
Moreover, there are two things that were ignored that certainly make a quarterback better, though the extent of each cannot be determined in this article. The purchased traits, and the playbooks. These are undoubtably factors to the equation which likely explain how the better overall QBs had worse ratings. One possibility is that the best QBs take traits instead of attributes, which lowers their overall and raises their effectiveness. This is just a theory, however, and cannot be proven with the data available to me now. Another possibility is that every QB in both the ISFL and DSFL are system quarterbacks, and the system determines how effective they are. In any case, being a quarterback is a thankless job which requires a lot of dedication from the users who choose it, and I am grateful to receive any targets I can as a lowly TE.
Note to graders:
All Charts: https://imgur.com/a/vPjOWFg
Time to write: available via DM
First, the data. I compiled the stats and rosters from every team in both the ISFL and DSFL after the S46 season ended.
ASSUMPTIONS:
1. Data
I assumed season 46 (my only full season) was a typical season with typical results. Outliers are, of course, expected, but this is the base assumption that the data is normal.
I am using the stats from both indices, assuming that they are canon and without errors. This does not include the pre- or post- season. In this analysis, I will be analyzing Passing Yards, Passing TD, Completion %, QB Rating, and Interceptions.
2. Ratings
I assumed that the ratings that players had at the end of the season were used for the entirety of the season. I am aware that this is only true for DSFL capped players. However, this article takes long enough without factoring in game-by-game analysis, and I don’t even know how to do that this assumption helps simplify the data enough to analyze.
3. Playbooks
I am a simple TE, who was barely included into my DSFL team’s draft war room (and I suspect it was by default). I don’t even know how many playbooks there are, let alone their effect on each player (hint this is a special clue for later).
4. Player Intangibles
Player Archetypes and Traits are not a factor in this analysis. Running QBs will likely have lower passing numbers because they scramble more, but this is not being factored into my analysis.
6. Defenses
Player game-by-game performances relative to the opposing defenses are not factored into this analysis.
7. Statistical Conclusions
Any chart with a R-squared value of .50 or more is enough to be statistically significant. Higher slopes on the trendlines suggests more effectiveness as an attribute toward a team’s passing game. I am using linear regression models to draw conclusions.
ANALYSIS:
Similar to my receiver analysis, I decided to run 3 sets of data: DSFL, ISFL, and Overall. While this is a useful framework (by giving valuable data on low attributes), it does struggle in that I am not factoring in either game-by-game performances nor the individual defenses they play. This means that the DSFL traits will seem stronger against a defense which is likely one of: bots, IA players, and low attribute totals.
First, I checked to confirm that Overall rating had an effect on all the stats.
There are several intriguing conclusions that this data suggests. It suggests a positive correlation between overall and Yards, TD, and Completion Percentage, which intuitively makes sense. However, it also suggests a negative correlation in Rating and a positive correlation in Interceptions (meaning that higher overall players get lower Ratings and more interceptions). To confirm this, I ran the numbers for the ISFL and DSFL.
ISFL:
DSFL:
As you can see, stats are confusing. Fear not, reader, for I’m going to expedite the word portion (it’s a pain to write, read, and probably grade, and I plan on eventually doing one of these articles for every position).
ARM:
A better arm leads to yards and TD. This is true in both leagues, and is thankfully one of the only conclusions I feel confident in drawing. The data suggests that at low levels, a worse arm is better for Rating and Completion Percentage. However, since this trend reverses in the ISFL, this is likely due to either the new create QBs in the league with minimal TPE which are getting good results, or that better defenses demand higher arm talent to be able to beat them (while you can sneak by on other traits in the DFSL)
ACCURACY:
The conclusion I am drawing here is that new players tend to prioritize Accuracy over Arm, which works early for all stats. It is positive at all levels for gaining yards, and helps the DSFL throw more accurately and more TD. Besides that, not much going on here.
INTELLIGENCE:
At least Intelligence as a positive correlation in everything except interceptions. Intelligence is definitely a worthwhile stat to take for all QBs. Intelligence definitely makes a QB more effective at their job.
RESULTS:
If you care about yardage and/or TD, Arm > Intelligence > Accuracy. For Completion percentage, accuracy is helpful in the DSFL, but Arm seems to get you over the top in the ISFL. However, this is only an approximation, as the sample size and R-squared values are not large enough to draw accurate conclusions.
This statistical analysis helps underline one of the core tenets of statistics: correlation does not imply causation. A base understanding of the data implies that getting higher attributes negatively impacts one’s Interception rate and passer rating. However, this does not make sense intuitively, so there are in fact, at least two possibilities.
A: Higher TPE players throw more picks and are worse
B: Higher TPE players play more, which gives them more minutes/plays, which inherently leads to more picks. Players with less TPE will play in relief in blowouts, and will have more conservative play calling to benefit their numbers.
I know which scenario I believe in. And it’s not A.
Moreover, there are two things that were ignored that certainly make a quarterback better, though the extent of each cannot be determined in this article. The purchased traits, and the playbooks. These are undoubtably factors to the equation which likely explain how the better overall QBs had worse ratings. One possibility is that the best QBs take traits instead of attributes, which lowers their overall and raises their effectiveness. This is just a theory, however, and cannot be proven with the data available to me now. Another possibility is that every QB in both the ISFL and DSFL are system quarterbacks, and the system determines how effective they are. In any case, being a quarterback is a thankless job which requires a lot of dedication from the users who choose it, and I am grateful to receive any targets I can as a lowly TE.
Note to graders:
All Charts: https://imgur.com/a/vPjOWFg
Time to write: available via DM