*What does it take to catch balls?

[papalinator]

Hello and welcome to this statistical analysis of the ISFL and DSFL, using season 46 data. I will be analyzing all 3 positions that catch the ball (WR, TE, RB, hereon generally referred to as receivers), and analyzing what traits are useful for all 3 to thrive in their own team's respective offense.

I was inspired to write this analysis by Pvtpenne's analysis on TE traits, and decided to expand on that idea while drawing my own conclusions.

*Disclaimer: I apologize ahead of time to the grader(s) who have to review this data.  But also, I’m not sorry, because that’s your job and it appears that there have just been so few statistical analyses recently that I’m making up for the past users.

First, the data. I compiled the stats and rosters from every team in both the ISFL and DSFL after the season ended.

Assumptions:

1. Ratings
I assumed that the ratings that players had at the end of the season were used for the entirety of the season. This is, of course, not true. However, this statistical analysis is working on a general basis using overall season stats, and I am using the end of season stats. The growth of players throughout a season is not accounted for and can be a confounding variable in the final result which must be accounted for.

2. Stats
I am using the stats from both indices, assuming that they are canon and without errors. These do not appear to include post-season play.  I am only including stats from season 46.  In this analysis, I will be analyzing both catches and yards.

3. Playbooks
I am a simple DSFL TE, and do not know anything about any team's (including my own) playbook. I attempted to correct for this variable by comparing player stats to their team's total, instead of in a vacuum, which should better accommodate different team styles.

4. TDs
In both real life and the simulation, touchdowns are random.  I will not analyze the TD rate of players in this analysis, but this may change in future with more data.

5. Player Intangibles
Player Archetypes and Traits are not a factor in this 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 R2 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.

8. Rosters
I am not factoring in the rest of the roster in this analysis. Receiving is a zero-sum game, where if a TE catches the ball, a WR inherently cannot.  High performers on a team will likely skew the data downward of their teammates, as their efficiency will cut into their teammates potential production.

Analysis:

Part I:
I started by checking each player's Overall score (per roster page in index) against their contributions toward both yardage and completions.  This led to the following results:

Overall:


This chart led me to my first realization.  While incorporating DSFL data is useful to gain data points on lower attribute values, the high value players in DSFL will skew the results toward lower attributes being more valuable.  Since the DSFL has lower attributes around the league, I decided that every attribute check should be run 3 times.  Once combined, once with exclusively ISFL data, and once with exclusively DSFL data.  This will hopefully reveal data trends that hold true for all attribute values, while giving information about so-called “low-tier” and “high-tier” play, which may reveal some dichotomy on split strategies between the two leagues.

The charts for the ISFL and DSFL are as follows:





I then decided to check other potential stats that can help a team, based on intuition. I chose Agility,Hands, Intelligence, Strength, and Speed.

A basic charting yields the following charts.

Agility:







The R2 value is not high enough to draw any conclusions in overall, ISFL, or DSFL.

Hands:







The R2 value suggests that Hands has a high correlation with contribution to success in both leagues.  The Hands stat seems especially important in the DSFL, but it still has a benefit toward completions in the ISFL (but not yards).  The DSFL has higher trendline slopes, so Hands matter more, but it is still important in the ISFL to maintain a high hands attribute.

Intelligence:







The data actually suggests that in the DSFL, Intelligence is counterproductive.  However, I would posit that since this trend is reversed in both the ISFL and the overall data sets, this is due to DSFL players having less overall TPE to invest, and people who invest in other skills are more productive than those who invest in Intelligence first.

Strength:







This is the Intelligence stat with less correlation.  Strength seems completely arbitrary when looking at receiving potential.

Speed:







The R2 value is not high enough to draw any conclusions.

However, the focus on speed around both leagues, especially when compared to other stats, suggests a meta build which proves effective.  I decided to test combinations of Speed with other stats.  I decided to use Speed as a multiplicative factor with the aforementioned attributes, which matches the scaling factor of the TPE system.  The closer to 100 one’s speed is, the more the other attributes can shine.

Speed x Agility:







The R2 value is not high enough to draw any conclusions in overall, ISFL, or DSFL.  There seems to almost be a correlation in the DSFL, but the R2 value is not high enough to suggest it.

Speed x Hands:







The R2 value suggests that Hands still has a high correlation with contribution to success in both leagues.  If anything, Hands and Speed seem to have a multiplicative effect on each other, and having a high combination of the two seems very commensurate with team contribution in both leagues.

Speed x Intelligence and Speed x Strength:













I believe the Speed stat multiplier just reinforces my hypotheses about Intelligence and Strength.  The Strength and Intelligence attributes seem to not be particularly useful for receiving yards in either the DSFL or the ISFL.

Part I Conclusion:
For pass catchers, the optimal route appears to be
Speed
Hands
Agility
Intelligence
Strength
for both the ISFL and the DSFL.  Hands is an extremely valuable attribute in the DSFL in particular, and has potential to be more valuable.  However, there is little data about low speed builds, and so I cannot in good faith give the recommendation to take it over Speed, even with a stronger correlation to team success.

Part II:
Part I yielded valuable conclusions.  However, as many may know, TE and RB are different in the sim from a WR.  Therefore, they may have different optimal attributes.  I decided to analyze each position’s breakdown of the basic and advanced stats to better gleam if there are any differences in each position’s preferred attribute distribution.  Even though Intelligence and Strength seemed unimportant, I kept them in the positional analysis to confirm that each position wasn’t being drowned out by the others.

WR:
The proverbial stars of the show. This is a majority of the pass catchers, so this will likely follow the same trends as the overall data.

Overall:







Good WRs are good, especially in the DSFL. No surprises here.

Agility:







Similar to Combined stats, not high enough R2 value to draw any conclusions.

Hands:







The R2 values are actually shocking, given the combined stats.  The only statistically significant conclusion that can be made with this raw data is that Hands contribute to completions in the DSFL.  However, the advanced version of these stats may tell a different story.

Intelligence/Strength:













Similar to combined, negative correlations are seen in the DSFL only, and these stats seem unhelpful.

Speed:







The R2 value suggests that in the DSFL, Speed is enough.  Speed is king.  In the ISFL, the data shows that speed definitely leads to more yards, but not necessarily more receptions.

Advanced Stats:
On the one hand, the high correlation that the DSFL has with speed leads to skewed data in that league.  However, this can actually reveal how effective other attributes are, given that speed is so valuable.

Speed x Agility:







Speed + Agility is a very good combination which plays off itself in the DSFL.  However, it seems that combination is not as effective later in the ISFL, though still strong.

Speed x Hands:







The R2 value suggests that Hands is a solid attribute for WR in both the DSFL and the ISFL when paired with good speed.

Speed x Intelligence and Speed x Strength:













I am combining these two combinations again because they are just not on the same level as the other attributes.  All Positive correlations for WR, but not worth per TPE as the others.


Final Ranking:
Speed
Agility
Hands
Strength
Intelligence

RB:
I predict most of the WR scrolled until seeing the word WR, will only read that section, and their enjoyment of the analysis will depend on how much it confirms their own experience. For those of you who are still reading, thank you.

This is the same format as the WR section.  There was no consideration for archetypes, so receiving backs are not separated out.


Overall:







There isn’t any correlation between overall and receiving contribution.  This can be attributed to different playbooks and archetypes, as some running backs will not be used in the receiving game.  This almost made me just write off RB as a position group, but for those RB who are interested, I continued on.  There are likely not many statistically significant conclusions that can be drawn from this data, however, but the show must go on.

Agility:







As predicted, not high enough R2 value to draw any conclusions.

Hands:






Same conclusion as the WR breakdown in this category.

Intelligence/Strength:







Same conclusion as the WR breakdown in this category.

Speed:







Unlike the WRs, Speed will not single handedly make a RB contribute to the passing game in either league.

Advanced Stats:
This actually allows for the inverse analysis from the WRs.  A low speed base correlation allows us to analyze the effectiveness of multiple attributes on receiving.

Speed x Agility:







Speed + Agility does not tell us enough information to draw any conclusions.

Speed x Hands:







The R2 value can actually help us draw conclusions.  The Hands attribute, when combined with speed, helps RB in both leagues contribute to the receiving game.  It is highly plausible that receiving backs are the ones primarily focusing on hands, and this is the stat which demonstrates how effective they can be in the passing game.

Speed x Intelligence and Speed x Strength:













I am combining these two combinations again because they are just not on the same level as the other attributes.  I am only including these results out of intellectual honesty.  These stats are not very helpful for a team’s success.


Final Ranking:
Hands*
Speed
Agility
Strength
Intelligence

*Caveat: This is by no means an excuse to ignore Speed.  This is only to say that Hands has a bigger impact on the RB’s impact on the passing game than Speed does.


TE:

I predict the only people reading this are one of the following:
a TE (now and prospective)
an ex TE (fell to the chart)
an article completionists who can’t leave without finishing an article
the grader
people who scrolled too far when looking for WR

The problem with TE is that there are not enough of them to honestly draw statistically significant conclusions.  In order to get more data, there needs to be more TEs to get more data to determine the best course of action for TE.  However, for the 5 groups of people listed (and any I missed), here is my analysis.


Overall:







Good TEs contribute to teams.  Enough said, we can keep moving.

Agility:







With such a small sample size, it is pretty damning that the R2 value cannot come to a conclusion about the importance of agility.

Hands:







Shockingly, this is not strong enough of a correlation to draw any conclusions.  Going into this, I thought that hands would be the go to attribute to adjust.  The advanced stats may show another story.

Intelligence/Strength:













Similar to everyone else, Strength seems wasted on this position.  It appears that Intelligence may have a role to play in the ISFL (with the high R2 value), but not for the DSFL.

Speed:







Like the WRs, speed will carry a TE in the DSFL (for yardage, at least), but not enough evidence to show for the ISFL.

Advanced Stats:
This is moderately useful for ranking the skills, but is not going to be enough to draw any statistically significant conclusions.  I will draw all conclusions at the end of the TE section

Speed x Agility:









Speed x Hands:







Speed x Intelligence:







Speed x Strength:









Final Ranking:
Speed
Hands
Intelligence
Strength
Agility

Intelligence has a shocking amount of correlation and impact when combined with Speed.  However, since the TE data set is so small, there can be too many confounding variables to truly say which is actually the best.  TE is the wild west, and because there is such limited data available, there is actually no way to determine if one’s build is statistically significantly better than another (with the same number of TPE, don’t @ me)

Part II Conclusion:
The different positions have unique preferred attributes to succeed in their individual passing games.  However, Speed is definitely king in this sim.  The rest of the abilities almost unilaterally are secondary to speed (with some possible aforementioned exceptions in the RB department).


Note to Graders:

All the charts are located at the below.  There are 120 charts in total, so in case I messed up the link to any, the charts are here.  I also can provide the Google Sheet document I used for this entire article upon request.
Basic Combined
Advanced Combined
WR
RB
TE

[Baron1898]

this article has the sauce

[soevil]

[.HoshinoIchika]

every time you put R2 instead of R-squared my sanity decreased a little

[roquefort]

but i thought tpe didn't mean anything and it was all luck at the dsfl level ..

[papalinator]

every time you put R2 instead of R-squared my sanity decreased a little

That's a formatting issue that didn't transfer from google docs

[.HoshinoIchika]

but i thought tpe didn't mean anything and it was all luck at the dsfl level ..

You don't believe in the Sim God? Get him! He's a heretic!

[JJ5]

wow! amazing analysis. one has to wonder how much of this is correlation vs. causation... YPC, and yards per target would be cool to analyze. I guess the only way to really test is to plug 2 different builds with different attributes into a game and run it thousands of times...

[Tmoney6996]

This is really awesome dude, and a very put together correlative analysis of the stats, at both levels. Love seeing this sort of in-depth look at the underlying attributes of the sim!