I put this together and thought I would share. I realize that this isn't a perfect analysis, since player skill has a big impact, but there are still some correlations to be made when looking at the league wide sample. Note, I threw out any QB who had less than 200 career attempts. The closer to 1 the correlation value is, the more it influences. The obvious influence: Overall Accuracy Things that don't seem to matter: Arm Strength Awareness

There is also a strong correlation value between how many cigarettes your qb smokes and how little they give a shit

To follow up on Shaun's great work I put together a graph of the overall of the quarterback vs the number of interceptions thrown. The graph looks the opposite of what one would expect. I expected to see that as the overall of the player increases the number of interception drops, but I guess this tells us that what matters the most is most likely user error I was debating whether to remove Noles and his 11interceptions in one game since it is a once in a lifetime outlier achievement, but decided against it I will try to see if I have some time later to run a linear regression and see if we can see whether the relationship represented by the graph has statistical merits or is it not statistically significant

The only problem I also see here is the improvement of players over time. Take Kaelin Patterson for example. Last season he was an 83 overall when he accumulated the majority of his good passing stats, but now he is a 94 overall. So I think the correlation of overall to the numbers isn't as relevant because it is also a self-fulfilling prophecy, guys that do well will increase in overall. That all being said, the lack of correlation with arm strength I find fascinating.

This graph is the interception % vs overall of the player. I guess again there is a slight upward trend, which points to some user error as being an important factor of interception % as opposed to high overall. The trend line seems to be flatter implying that in terms of interception % there might not be any trend (dependency) relative to the overall of the players.

I feel like the linear line could actually be parabolic based on the data points somehow meaning that really bad overall QBs and really high rated QBs throw more

The first graph I posted is the total interceptions and the second graph is the interception percentage

There it is I averaged SAC, MAC and DAC. Again it seems that user error is the name of the game, even though the trend line is not very steep.

Nice analysis Shaun and Ehmo. Although this may take a lot more work than you may want to take on, perhaps you could take the analysis deeper and see if the QB success/failure correlates with the receiver caliber on the respective teams. Just a thought. Thanks for posting this.

I'd also be curious to see how confidence plays into it. Obviously it's something that cannot be tracked with attributes and stats, but confidence is an intangible that weighs heavily on player output for all positions on the field. I've started to value player confidence more than player progression with the game prep activities.

Patterson is all over the place this year. Even though all his ratings are high and his confidence is in the 80's. He is throwing wild passes a lot