*TPE Allocation in terms of On Field Success - TE

[nunccoepi]

As I've set out attempting to develop the best TE ever, I wanted to know in which attributes to invest the TPE I've earned. After all, we only have a limited amount--even if you are hitting all PT's and churning out media. In order to do that, I determined the best way to see a direct link between attributes and on-field success would be to calculate a measure of correlation (through Pearson's product-moment correlation coefficient). Essentially, this correlation is represented by a number which, as it approaches 1 indicates a greater correlation between two variables--the higher the number the greater the link between the two factors.

Here is the data that I used for the analysis: all TE's sorted by overall rating.

NSFL in blue, DSFL in red. I excluded the bots from PB and NOR since I knew they were bots, but the kept all the rest since I wasn't sure which are bots and which aren't. Doesn't matter though--their TPE allocations should still affect their on-field success in the same way. I believe I might also be missing a TE or two that I didn't have complete data for. Obviously, this is data is only relevant up to this point in the season but should be more accurate for a larger data set.


And the main event.

I eliminated the attributes that I felt had no realistic advantage towards receptions, yards, and touchdowns: QB attributes, Kicker attributes, and Blocking attributes (although this last category does apparently affect the overall rating of a player). Remember, as the number approaches 1, it indicates a greater correlation between the two variables--the higher the number the greater the link between the two factors.
At first blush we have some interesting conclusions:
-No areas have a negative correlation meaning as they increase, the on-field performance decreases. That's expected.
-In the document describing which attributes are most important for overall rating, it ranks speed, hands, and strength as the highest followed by endurance and intelligence. Agility has apparently no bearing on overall rating for TEs. This mostly holds true in the analysis I did. However, hands seem to demonstrate very little correlation, and agility enough for the second most important category (other than overall which can be excluded as it cannot be raised without raising the other categories).
-Strength also has a strong correlation. This might indicate that strength is a key component in avoiding or breaking tackles. I'd be interested to see how it correlates with YAC.
-It seems that higher correlations can be seen for data presented throughout the game as a whole rather than a per/reception basis. This might mean that more information is needed as a whole to be accurate in determining correlation. Or it might just mean that a game gives a better snapshot of a player's skill as determined by TPE allocations than isolated plays--which would make sense.
-Finally, there appear to be weaker correlations between TPE allocations and touchdowns across the board. This might not have as much to do with TPE allocations and more on how the sim dictates which positions get more goalline looks regardless of skill level or what each team's game plan includes. The only exception might be endurance which seems to correlate more strongly with touchdowns than any other stat line.

Here's another graphic to see these findings represented more clearly.

I think this image really highlights the categories that do and do not matter: speed, agility, strength; and intelligence and hands respectively.

Hopefully this was illuminating for you as it was for me. Remember, this is only for TE's at this point in the season. If you're interested in seeing more and would like for me to do other positions, let me know. I'll probably try to use data from previous seasons if I do this again, so that I can have a more complete data set. Thanks for reading.

Edit: at the risk of taking some of the teeth out of the analysis, it's worth noting (in case you haven't heard) that correlation doesn't imply causation. So, just because doesn't a player has high skills in one area; that doesn't mean the stats necessarily cause an increase in one area of stats or another (at least not according to this data). Other causative factors not taken into account here likely include game strategy, play calling, matchups, gameflow, and just the sim being unpredictable.

[bovovovo]

Awesome stuff!

Its hard to be able to really pin point contributors like this without using custom-made rosters in the sim because of all the different variables (Are there other good WRs on the TEs offense hogging targets? How good is the TE's QB? How often do they even pass it? What playbook are they using? etc) but you mentioned that already.

Forreal, good stuff dude :cheers:

[nunccoepi]

Awesome stuff!

Its hard to be able to really pin point contributors like this without using custom-made rosters in the sim because of all the different variables (Are there other good WRs on the TEs offense hogging targets? How good is the TE's QB? How often do they even pass it? What playbook are they using? etc) but you mentioned that already.

Forreal, good stuff dude  :cheers:

Thanks, it's a start anyway.

[iamslm22]

Also interesting to note that the Hawks are the only NSFL that doesn't use a TE at all. Have used WRs at the TE position since S3.

Great work here man.

[nunccoepi]

Oh interesting. I just thought it was a personnel thing, I didn't know it was more of a policy

And thank you

[jaskins811]

Awesome awesome stuff! Love seeing cool statistics on stuff from the league!

[timeconsumer]

Also interesting to note that the Hawks are the only NSFL that doesn't use a TE at all. Have used WRs at the TE position since S3.

Great work here man.

Which has a much greater effect on how targets are distributed than one would expect. Another thing that can distort statistics is that Dimirio while being listed as a TE actually spends a lot of time starting at the WR position which is going to inflate his efficiency numbers because of how this sim covers receivers.

@nunccoepi if you want some better data sets to work with I can pull an extract of the advanced statistics out of the sim which would give you the hidden stats like number of snaps played, number of drops recorded, number of tackles missed, etc.

[nunccoepi]

Which has a much greater effect on how targets are distributed than one would expect. Another thing that can distort statistics is that Dimirio while being listed as a TE actually spends a lot of time starting at the WR position which is going to inflate his efficiency numbers because of how this sim covers receivers.

@nunccoepi if you want some better data sets to work with I can pull an extract of the advanced statistics out of the sim which would give you the hidden stats like number of snaps played, number of drops recorded, number of tackles missed, etc.

I had a feeling that was the case with Dimirio. So yeah, there's only so much this data can show. But yeah I'll take all the data you can give me! Although I'd be interested in a complete seasons set so s5 would be best.

[nunccoepi]

@nunccoepi if you want some better data sets to work with I can pull an extract of the advanced statistics out of the sim which would give you the hidden stats like number of snaps played, number of drops recorded, number of tackles missed, etc.

@timeconsumer would you happen to have that extract?

[timeconsumer]

@timeconsumer would you happen to have that extract?

When I get home. You'll need to splice some of the data together to get anything worthwhile out of it, index-match worked for me.