League Forums

I've been starting to dig through the stats and have noticed that there seem to be quite a few dropped passes.

Looking at the 2010, 2012 and 2013 NFL seasons we see the following data:

2010
1. Worst 5 receivers dropped between 10 and 13 passes per player for the season.
2. Number of receivers with 10+ drops was 5.

2012
1. Worst 5 receivers dropped between 10 and 14 passes per player for the season.
2. Number of receivers with 10+ drops was 6.

2013
1. Worst 5 receivers dropped between 9 and 12 passes per player for the season.
2. Number of receivers with 10+ drops was 3

MFN-1
2013 - League creation season
1. Worst 5 receivers dropped between 9 and 12 passes per player for the season.
2. Number of receivers with 10+ drops was 4.

2014
1. Worst 5 receivers dropped between 12 and 14 passes per player for the season.
2. Number of receivers with 10+ drops was 23.

2015
1. Worst 5 receivers dropped between 14 and 19 passes per player for the season.
2. Number of receivers with 10+ drops was 25.

MFN-2
2013 - League creation season
1. Worst 5 receivers dropped between 11 and 16 passes per player for the season.
2. Number of receivers with 10+ drops was 5.

2014
1. Worst 5 receivers dropped between 12 and 17 passes per player for the season.
2. Number of receivers with 10+ drops was 18.

The NFL data seems pretty consistent while the MFN data seems to be trending upward, especially where the number of receivers with 10 or more drops are concerned.

I wonder if the pass catching numbers are a bit low for the top receivers in the draft. Also, perhaps the receivers pass catching development is also a bit slow.

One reason dropped passes are high is an antidote to bring pass completion percentage down around 60% - if you look at the first season (and maybe the 2nd) you'll see QB completion percentages in the 80's (and lower-rated qb's struggling to get in the 50's). I spent a lot of my football-watching time this season studying why passes are incomplete and I think I there are a lot more intentional incompletes (or incompletes from throwing the ball in the only place that the defender can't make a play on it) than drops or bad passes. I have some plans to update some of the pass targeting logic to play more in this vein.

How does MFN determine if a pass was dropped or just incomplete?

Looking a bit further into this area we have some additional NFL data for the average number of drops for the top 20 receivers. I am assuming this is based on the number of receptions.

2005 - 8.45
2006 - 8.90
2007 - 8.60
2008 - 8.60
2009 - 8.85
2010 - 8.65
2012 - 9.30
2013 - 6.25

MFN-1
2013 - 4.35
2014 - 9.00
2015 - 9.65

MFN-2
2013 - 3.70
2014 - 8.35

These numbers look much better, but more seasons will be needed to see which way the data will trend.

When the ball is thrown from the quarterback, the optimal launch vector is determined, and then error is added to it based on the QB's accuracy. The ball is then launched, and the receiver adjusts to get to the optimal receiving location based on the flight path of the ball. Once the receiver reaches his closest point to the ball, if the ball is outside of his arm's length or too high then he is unable to catch it, so he does not get a drop stat. If the ball is within his arm's length, he gets a probability factor of whether or not the catch is made. This takes into consideration his pass catching ability, the distance away from him that the ball is, the altitude of the ball, whether the ball is behind him vs. in front of him (behind him reduces his probability of making the catch), and also the proximity of the defenders and their velocity vectors (if he's about to get a head-on collision, he is more likely to drop the ball, driven by his courage bar and the defender's Punish Receiver bar). If the defender has a strong enough influence in the drop, then instead of a drop stat the defender gets a knock-down stat.

That is very sophisticated!

I think it's very important to switch the burden of the incomplete pass to the defense rather than the dropped pass. What's happening now is that we're keeping the offense down by artificial means, but the result is that defense matters less as it has a lesser impact on the game than what the offense does.

edit: Thought I'd add an example. I recently benched my 71/71 TE for a lesser player because of drops. Since defense can't stop receivers anyway, the rest of the bars have much less weight than pass catching!