During the week leading up to the Georgia Tech game on Sept. 26, I saw a lot of Internet hate going Logan Thomas' way. Through four games in 2013, Thomas, once touted as a top-five NFL draft prospect, had completed just 65 of his 134 passes for a miserable 48.5 completion percentage, four touchdowns and six interceptions; already one more than his goal for the season.
As many know, Logan performed admirably in 2011, completing 59.8 percent of his passes for 3,013 yards, 19 touchdowns and 10 interceptions. However, during his first year as Frank Beamer's starting signal caller, Thomas was surrounded by a myriad playmakers, including the top two receivers, statiscally speaking, in school history in Danny Coale and Jarrett Boykin, as well as David Wilson, who would run for a school-record 1,709 yards that season.
In 2012, Logan's numbers dropped considerably, as he was plagued by fewer playmakers on offense and the complete lack of a running game. He finished with a 51.3 percent completion percentage with 2,976 yards, 18 touchdowns and 16 interceptions as the Hokies needed a strong finish to extend the school's 20-year bowl streak.
Through those first four games of 2013, Logan's stats resembled those he put up in 2012, rather than the 2011 numbers fans would prefer to have seen.
So it's not to say that the frustration among fans was unwarranted, if it was being based completely off Logan's stat line. Having watched every snap of every game he's ever started, I didn't believe that Logan's numbers adequately reflected his performance to that point in the season; thus began a mission.
In the past three weeks, I have watched every pass that Logan Thomas has thrown between the start of the 2011 season and Saturday's win over Pittsburgh. Sometime between now and then, "catch efficiency" was born.
Catch efficiency is a statistic I created to measure a pass catcher's ability to haul in balls thrown their way, based on difficulty. It has nothing to do with a quarterback; only those he throws to. The results I came up with did not come as a surprise, and proved what I originally believed to be true, but I will get to all that in a bit.
What is catch efficiency?
(er) + (sr)10/9 + (sd)1/4 + (tr)4/3 + (td)1/2
catchable targets
Before you make fun of my rudimentary equation (keep in mind I am a communication major), allow me to explain myself.
Every pass a quarterback throws can be put into four different categories:
Easily catchable: Expected to be caught by any pass catcher at any level. They require little to no adjustment on the receivers part and if you're on the receiving end of an easily catchable pass, it means there are most likely no defenders within a yard or two of where you are when you come into contact with the ball.
Should have been caught: Slightly more difficult to haul in than easily catchable passes, but are caught in most instances. May require some adjustment on a receivers part; having to slow up because a ball was thrown behind them, having to squat to catch a pass, etc. In some cases, pass could be easily catchable, if not for the defender on the receiver's back.
Tough catch: Are not caught most of the time. Pass usually requires major adjustment by the receiver; having to dive, maximizing their vertical, battling with a defender, taking a serious hit as soon as the ball hits their hands, etc. Defenders can often play a part in tough catches, but it is not necessary.
Out of reach: Receiver is unable to even get a hand on the ball, most often because of the impact of a defender or a quarterback's inaccuracy.
Before I go any further, understand that out of reach passes do not factor into catch efficiency. As I said, a quarterback has nothing to do with catch efficiency. It is strictly a measure of a receiver's ability to haul in passes that can to be caught.
That being said, let's talk a little about the formula, which puts on emphasis on catching tough catch passes and dropping easily catchable ones.
"Easily catchable" receptions get a base rate, with a receiver getting a little bit of credit for receptions that "should have been caught." A receiver gets even more credit for a "tough catch" reception. Likewise, drops that "should have been caught" still give a little credit to the receiver because they weren't easy receptions. "Tough catch" drops get even more credit, because they're not expected to be caught in most instances. Likewise, "easily catchable" drops receive zero credit. Keep in mind that all passes that are not out of reach are catchable targets.
(er) + (sr)10/9 + (sd)1/4 + (tr)4/3 + (td)1/2
catchable targets
er = easily catchable receptions
sr = should have been caught receptions
sd = should have been caught drops
tr = tough catch receptions
td = tough catch drops
So, in theory, a receiver can make up for dropping one easily catchable pass by hauling in three tough catches. Let's take a look at Dyrell Roberts in 2012, who I expected to have an awful catch efficiency.
Dyrell Roberts (2012)
64 targets, 33 catches
28 easily catchable (26)
6 should have been caught (3)
12 tough catch (4)
18 out of reach (0)
85.7% catch efficiency
Although Roberts dropped two easily catchable passes and failed to come up with half the should have been caught passes thrown his way, he did make one-third of the tough catches Logan threw him, giving him an average 85.7 percent catch efficiency. By comparison, in 2012, Corey Fuller finished at just 84.9 percent, while Marcus Davis came in at 84.7 percent.
A receiver can also exceed 100.0 percent if they're making all of the catches they should be making, as well as a few of the tough catches. I give you 2011 Danny Coale.
Danny Coale (2011)
83 targets, 59 catches
40 easily catchable (40)
12 should have been caught (12)
13 tough catch (7)
18 out of reach (0)
102.5% catch efficiency
Those numbers are absurd, especially considering the fact that the second-highest catch efficiency among Logan's pass-catchers in a single-season comes from Kalvin Cline, who currently stands at 94.7 percent, and that's with a small sample size of 20 targets (17 catchable). The next-highest single-season mark comes from 2011 D.J. Coles, who came in at 92.9 percent catch efficiency with 49 targets (41 catchable). So yes, 2011 Danny Coale really was THAT good.
Now that we've covered what catch efficiency is, it's important to understand what kind of numbers Logan's receivers have been putting up in 2013. I tweeted a screenshot of this list the other day; if you saw it, it should make a little bit more sense at this point.
Logan Thomas receivers' catch efficiency, ranked
1.Danny Coale (2011, 83 targets) 102.5%
2.Kalvin Cline (2013, 20 targets) 94.7%
3.D.J. Coles (2011, 49 targets) 92.9%
4.J.C. Coleman (2012, 32 targets) 92.3%
5.Demitri Knowles (2012, 41 targets) 90.8%
6.Willie Byrn (2013, 43 targets) 90.8%
7.Josh Stanford (2013, 32 targets) 88.2%
8.Ryan Malleck (2012, 28 targets) 87.6%
9.Jarrett Boykin (2011, 104 targets) 86.9%
10.Dyrell Roberts (2012, 64 targets) 85.7%
11.Chris Drager (2011, 28 targets) 85.3%
12.Corey Fuller (2012, 74 targets) 84.9%
13.Marcus Davis (2012, 100 targets) 84.7%
14.Marcus Davis (2011, 53 targets) 84.6%
15.David Wilson (2011, 30 targets) 83.6%
16.Demitri Knowles (2013, 51 targets) 80.9%
17.D.J. Coles (2013, 23 targets) 79.1%
18.Randall Dunn (2012, 25 targets) 77.7%
As you can see, Demitri Knowles, who leads the 2013 team in targets, comes in at just 80.9 percent. Knowles has seen his fair share of struggles so far this season, hauling in just 29 of 41 catchable passes.
Demitri Knowles
51 targets, 29 catches
24 easily catchable (20)
12 should have been caught (6)
5 tough catch (3)
10 out of reach (0)
80.9% catch efficiency
However, it is important to note that Logan's first four games and last three games are night and day. Knowles' are the same way.
*CB% stands for catchable ball percentage, or the number of catchable passes a quarterback throws
First four games of 2013
Logan Thomas: 65-134, 48.5% (70.1% CB), 4 TDs, 6 INTs
Demitri Knowles: 26 targets, 71.3% catch efficiency
Last three games of 2013
Logan Thomas: 57-87, 65.5% (78.2% CB), 5 TDs, 0 INTs
Demitri Knowles: 25 targets, 100.7% catch efficiency
So, as Knowles' catch efficiency has dramatically improved over the past three games, so have Logan's numbers. Coincidence? You tell me.
Receiver catch efficiency by game, 2013
Alabama
Josh Stanford: 100.0%
D.J. Coles: 58.3%
Demitri Knowles: 40.9%
Western Carolina
D.J. Coles: 100.0%
Willie Byrn: 98.8%
Josh Stanford: 81.3%
Kalvin Cline: 80.5%
Demitri Knowles: 77.0%
East Carolina
Willie Byrn: 101.6%
D.J. Coles: 100.0%
Kalvin Cline: 100.0%
Demitri Knowles: 88.6%
Josh Stanford: 79.4%
Marshall
Kalvin Cline: 122.0%
Josh Stanford: 94.4%
Willie Byrn: 71.6%
D.J. Coles: 68.8%
Demitri Knowles: 65.0%
Georgia Tech
Demitri Knowles: 111.0%
Josh Stanford: 111.0%
Willie Byrn: 102.8%
Kalvin Cline: 100.0%
D.J. Coles: 70.8%
North Carolina
D.J. Coles: 105.5%
Demitri Knowles: 102.8%
Kalvin Cline: 91.5%
Willie Byrn: 86.6%
Josh Stanford: 83.3%
Pittsburgh
Josh Stanford: 100.0%
Kalvin Cline: 92.2%
Demitri Knowles: 90.7%
Willie Byrn: 82.8%
So, even though Knowles had an abysmal 40.9 percent catch efficiency against Alabama, he put out +100.0 percent performances against Georgia Tech and North Carolina. And, although his total ranks on the lower end of Logan's receivers, if he continues to put out impressive performances, it will no doubt start to climb closer to his 2012 total.
D.J. Coles, who has just 23 targets on the season (17 catchable), currently has the second-worst single-season catch efficiency mark of any of Logan's top targets of the past three years, as he sits at just 79.1 percent. Keep in mind, that sample size is still very small, and we've seen in the past that Coles is capable of posting a high catch efficiency when he's healthy. We all know that Coles' playing time has been limited, for reasons that have never been made clear to the media, leaving everyone to guess that its because he's either A) not completely healthy or B) out of shape. My guess is the latter, and it has no doubt affected his catch efficiency numbers thus far in 2013.
D.J. Coles
23 targets, 11 catches
10 easily catchable (9)
3 should have been caught (1)
4 tough catch (1)
6 out of reach (0)
79.1% catch efficiency
The team's best catch efficiency performer in 2013 that's actually a wide receiver comes in the form of Willie Byrn. Despite not seeing a single target against Alabama, Byrn ranks second on the team in targets, behind Knowles, and has made the most of them.
Willie Byrn
43 targets, 25 catches
18 easily catchable (17)
7 should have been caught (5)
7 tough catch (3)
11 out of reach (0)
90.8% catch efficiency
Redshirt freshman Josh Stanford has performed admirably as well, coming in just below Byrn at 88.2 percent.
Josh Stanford
32 targets, 19 catches
11 easily catchable (10)
9 should have been caught (7)
6 tough catch (2)
6 out of reach (0)
88.2% catch efficiency
It's also worth noting that freshman walk-on fullback Sam Rogers has a 99.3 catch efficiency, as he's caught seven of the eight catchable passes thrown his way.
Now it's time to look at how Hokie pass catchers have fared for Logan in 2011 and 2012, so we have something to compare this year's numbers to. The list posted above covers individual numbers, but its important to look at how they performed as a team, in order to better determine how well Logan is actually playing.
2011: 75.2% CB, 90.3% catch efficiency, 59.8% completion percentage, 19 touchdowns, 10 interceptions
2012: 72.1% CB, 85.3% catch efficiency, 51.3% completion percentage, 18 touchdowns, 16 interceptions
So, yes, Logan did throw 3.1 percent more catchable balls in 2011 than he did in 2012, but a young offensive line and lack of confidence no doubt played a part in that. And while 5.0 percent seems like a very small number (the difference between 2011 and 2012 team catch efficiency), keep in mind that the difference in 2012 Corey Fuller and 2013 Demitri Knowles is just 4.0 percent. It may not seem like a lot, but every little bit means the world.
Unsurprisingly, Logan's 2013 numbers lie directly between his 2011 and 2012 campaigns. But again, the difference in his first four games and last three games is incredible. While Logan himself has improved during that time period, his receivers have been giving him a lot more help since that Thursday night down in Atlanta.
*Note that team catch efficiency is not just determined by Logan's top five targets. Every receiver who has at least one catchable pass thrown their way is factored into team catch efficiency.
Team catch efficiency by game, 2013
Alabama: 51.7%
Western Carolina: 86.1%
East Carolina: 93.3%
Marshall: 81.3%
Georgia Tech: 100.2%
North Carolina: 90.2%
Pittsburgh: 91.9%
The Georgia Tech game has easily been the receiving corps' best of the season, as they hauled in 19 out of 21 catchable targets, including three out of four tough catches. When Logan's pass catchers are making tough catches, especially early in games, it has a clear affect on his confidence and his ability to make impressive throws.
Logan completed his first nine passes against the Yellow Jackets, but only five of them were easily catchable. Sam Rogers made a tough catch on Logan's third pass of the game; the ball was thrown behind him as he was running out in the flat, and Rogers made an impressive, one-handed grab. Then, Knowles made a tough catch on Logan's fifth pass of the game, another one-handed grab on a throw down the field, while he was being face-guarded by the cornerback.
That game was Logan's second best in terms of completion percentage in the past two and a half seasons. His best outing in that time came against Miami in 2011, when he completed 23 of 25 passes for 310 yards and three touchdowns.
As was the case against Georgia Tech, Logan got a lot of help early from his receivers. His first two passes of the game were both tough catches, but he was bailed out by Danny Coale on the first and Jarrett Boykin on the second. The team's catch efficiency in that game was 101.8 percent.
First four games of 2013: 70.1% CB, 82.2% catch efficiency, 48.5% completion percentage, four touchdowns, six interceptions
Last three games of 2013: 78.2% CB, 93.9% catch efficiency, 65.5% completion percentage, five touchdowns, zero interceptions
Logan had an outstanding year in 2011, due largely in part to the confidence he had in the players around him. We saw the opposite in 2012, and through the first four games of 2013. But the Georgia Tech game seemed to be a turning point for both him and his offense, and if they continue to make plays for him, theres no reason his numbers at the end of this season won't resemble those of 2011.
If you're interested, here are the full game logs, player logs and research I did for 2011, 2012 and 2013.
Comments
Very cool analysis. I'm impressed by the amount of work you did to compile all this information. Well done!
Cool Stuff
This is a phenomenal post. I wish I could award you a million turkey legs for it.
Yeah this is pretty awesome. I know it would really complicate things, but is there any way to incorporate YAC and TD catches?
probably, but that wouldn't change catching efficiency per say. when you calculate QBR, it incorporates efficiency but not to the degree that this efficiency is calculated. If you include TD and YAC, it would become WRR (Wide Receiver Ratings)
love the idea of stat to measure catch efficiency, and you executed well
nice
Thanks for writing all of this up, it's always great to get some stats to go along with the eyeball test to see how a receiver is doing.
Love the effort, but c'mon man, its called TOTS, Sharkeys, Big Al's, Champs. By yourself a rail on me (by on me of course I mean buy a rail with your own money) for a job well done.
Epic film review! And some great info.
It never ceases to amaze me the level of knowledge and talent on TKP. Excellent article and research!
That's pretty cool. Interesting that 2012 Marcus Davis is better than 2011, probably just because he was targeted more. Out of curiosity does this or your film review take into account the receiver causing the throw to go from easily catchable to out of reach or hard to catch? Such as with DJ Coales giving up on that route in the Bama game and causing the pick 6?
Excellent question. I counted that particular pass as "out of reach," because when the ball crossed paths with Coles, he was unable to make a play on it, even though it was due to a lack of effort on his part.
So, to answer, a lack of effort on a receiver's part does not play a part in catch efficiency. Several times in 2012 when Logan was throwing to Dyrell Roberts, Roberts made little to no effort to catch a pass that he probably could have at least gotten a hand on however, he didn't, so all of those passes were filed as "out of reach."
This is terrific work Zach. Very insightful look into the receivers production, and it also demonstrates how each receiver is a credible threat, making it difficult for a defense to focus on shutting down just one guy. As Sid Gilliam often taught, the geometry of the game dictates that a defense can not defend the entire field effectively, so if every receiver is a threat and the patterns attack every bit of available space while forcing defenders to make choices, the defense can not defend the passing game.
... Unless they have a pass rush that totally F's up the quarterback. Comon, French - you of all people...
Those JUGS machine are paying off. Moorehead told me he used to do it and used to catch eggs (not hardboiled either), said the JUGS machine were confidence builder (obviously), and with the running backs and fullbacks helping out in the passing game, Thomas has a wealth of targets to choose from.
I'm just waiting for Sam Rogers to score on a catch in the end zone.
Awesome analysis! I love numbers. A few questions:
What happens to these numbers when you stop giving credit for drops? Perhaps the following equation might make the differences between good and bad recievers a liitle more explicit:
(er+sr+tr)/(catchable targets - (sr+sd)(1/4) - (tr+td)(1/2))
so, basically, every catch counts the same, but you only expect a reciever to catch 3/4 of the "should have been caught" and half of the "tough catches". I just pulled those out of my ass, you can make expected completion percentage for each catagory whatever you want.
Also, something that might be interesting while you've got the excel sheet with the data - what percentage of passes at a WR are uncatchable? They might be indicative of poor route running, not being on the same page with the qb, or just not having a great catchable radius. Likewise, a guy who always has a chance at coming down with a ball thrown at him is either an amazing physical specimen, or has a great connection with the qb.
ah yes, I was thinking just along those lines. Similar thinking would be expected from a fellow ae grad though. Not a fan of the weighting in the denominator though as it can produce unexpected results sometimes. Maybe have the weighting in the numerator? Or even based on a percentage, like this (which would mostly just penalize for uncaught balls that were catchable)
{A[er/(er+ed)] +B[sr/(sr+sd)] + C[tr/(tr+td)]} / (A+B+C) where A,B,C are your weighting coefficients
I'd also be interested to know what percentage of LTs throws have been in each catagory. IE, you have %Catchable Balls (CB), but it would be neat to see %Easily Catchable, %Should be Catchable and %Touch Catchable. Since it includes the influence of defenders it would be dependent on defenses, but over the course of the year might show some interesting trends.
Finally, I would not call it a "percent" (as it is not), but a rating. So Fuller was 4 pts (not 4%) better in 2012 than Knowles has been earlier in the year in 2013
2011:
391 attempts
184 easily catchable (47.1%)
58 should have been caught (14.8%)
52 tough catch (13.3%)
97 out of reach (24.8%)
2012:
429 attempts
168 easily catchable (39.2%)
67 should have been caught (15.6%)
74 tough catch (17.2%)
120 out of reach (28.0%)
2013:
221 attempts
90 easily catchable (40.7%)
42 should have been caught (19.0%)
30 tough catch (13.6%)
59 out of reach (26.7%)
where do throw aways work into this?
All throw aways count as out of reach passes. Logan had 13 in 2011, 14 in 2012. He has 12 already in 2013.
12 already because he is playing so much smarter in terms of possession
by the way, I'm just going to upvote the be-jeesus out of your comments on the page because I approve of the topic. Any way I could get my hands on the raw data? It's painful to admit, but I live for this crap
All of the data can be found in the links at the end of the post. If you have any trouble accessing them, let me know.
Hadn't clicked on them - wow. Bravo.
This is awesome, Zach. And holy hell, looks like this is about your first contribution to the community. Exceedingly well done, both the effort and the execution are impressive.
According to my own just-now-made-up-stat, your post efficiency is at about 519%, 5 for the Tyrod and 19 for the Open. Meaning that a recipient of your effort achieves knowledge more than five times the amount of effort required to read your offerings.
I'm surprised that Boykin graded so poorly. I seem to remember that he was able to catch just about any ball thrown his way.
As a Senior, yes. But as a youngster, not so much. I remember the XXXL glove commentary meaning nothing early in his career b/c of drops. Danny Coale as a youngster had great hands but didn't have the saavy to get open. He learned that as a Senior. Thus he is now always open.
Boykin had to learn how to position his body to make catches. He could get open, just drop the ball.
Upvote!
Well maybe we will see where he stands this weekend, as he likely will be up in a starting role for the Packers due to injury. I hope he has a great game.
Just think - had you been an engineering major you could have been a rocket scientist!
Nice work!
Impressive work, I think it really shows that although there was no complete excuse for his (Thomas) mediocre play at the start of the season it wasn't all his fault. As the receivers improve so will the QB...
A few years ago Football Outsiders tracked what they called Receiver Plus/Minus- looking at how well a receiver did at catching passes adjusted for depth of pattern.
http://www.footballoutsiders.com/stat-analysis/2009/receiving-plusminus-...
http://www.footballoutsiders.com/stat-analysis/2009/receiving-plusminus-...
I don't think they still track that stat, partially because it was Bill Barnwell's project (and he's left for Grantland), and partly because it's not terribly consistent year-to-year. You may want to look at that to see if you can improve your numbers. After all, it's easier to catch a bubble screen than it is to catch a 40 yard bomb.
Not from Logan its not.
I really like this
very good job. Tukery legs galore for you!!!
We're doing well with what we have. The more this new coaching staff inculcates the roster, the better our offensive performance will be down the line. Right now, mistake-free football is a crucial and field position battles must be won.
That was a hell of a thing. Great work.
Got another thing I'd be interested in seeing.
Counting the difference in the % involved when Yhomas receives a shotgun snap that requires him to drop down to get it.
Something I noted last year was the number of snaps coming in low.
Can't be good on timing routes or working through progressions when the ball comes in at your shins so you have to look down and bend at the waste to get the snap then look up and start to diagnose where the defense has moved.
How did that impact INT and in catchable balls, hurries and sacks.
Although it's been stated, I just want to reiterate how awesome of a post this is. Bravo
This is amazing work! Thank you. One of my few comments to Joe on the survey was that he could pick up another columnist so French and Mason don't get burnt out. I've got a nominee. This is fantastic work that you were kind enough to share with the community. Thank you!
Ditto for me with all the praise. Focusing on actually catching the ball and excluding everything else is quite interesting.
Though I personally don't like to see any stat that makes Marcus Davis look better than the colossal crap show that he was. And my former WR coach may call shenanigans as he always told us: "If you can touch it, you can catch it".
Anyway, this was extremely interesting and very well done. Thanks for that! I gave you a turkey leg for every reply you had up to this point as you should have the power to wield said turkey leg yourself.
Quick question, when analyzing the sugar bowl, how did you categorize the Danny Coale catch?
You answered in your question.
It was a catch.
As a tough catch that was incomplete. I had to go along with the record, even though you and I may know it's false.
What is this "may know" that you speak of?
Zach - just curious, but how much time did you put into this project? Very impressive!
Since I can't up-vote your OP, I gave you an up-vote for every response you made in this thread.
I would guess around 50-60 hours. I watched 34 games worth of Tech passing plays, and I would say those took about 40 minutes on average, depending on how many passes Logan threw. I mean, he only had 19 passing attempts against Appalachian State in 2011 and 49 against North Carolina in 2012, so there's obviously some disparity there. Plus all the time to log/compile the data, come up with the formula and write this post.
Thanks so much for your feedback! I really appreciate it.
This is awesome, I love this type of stuff. Just out of curiosity, how did you arrive at the coefficients used in your equation?
It sounds dumb, but honestly, I just weighted the different variables the way I thought they ought to be weighted. I'm far from a math major, but the results I found when testing these particular coefficients seemed legit, so I just went with it. It did take a good bit of tinkering to come up with the final formula.
Seems legit to me. I certainly wouldn't know another way to do it. Solid post, man. I really enjoyed it.
Zach, I'm a data guy by trade and also a guitarist, so allow me to add my 2 cents here.
Lots of people have played guitar and done it well since Hendrix, Eddie, Satch, Vai et all, but these few are considered ground-breakers (there are others but these are my guys). People have built upon what they started and have benefited from the experience of their vision. But they did their things, the way that they did them, first and we don't critique them- rather we are grateful that they came along in the first place.
Some of us probably have ideas to build upon this but I think we all can agree that we would just like to soak this one up for a bit first. Fan-freaking-tastic job bro.
I really appreciate that. As a fellow guitarist myself, I completely understand where you're coming from and wholeheartedly agree. There's definitely some adjustments that can be made to make this idea better and probably more accurate. But if I'm being honest, I'm probably going to take a little break from this research, if only for a little while. Don't wanna get too burnt out on it.
Turkey legs for all the guitarists!
Am I allowed to just jump in here asking for turkey legs for being a bassist, or would that just be turkey leg whoring? Know what, yes it is. But I need those turkey legs dammit!
the yin to our yang...turkey leg for you brother...
You know. What if people started TKP bands? I don't know what they would perform. Most likely Enter Sandman for all of them. Maybe Carry On My Wayward Son for a few of them.
Shane reference?
Hokey pokey, Na-Na-Na-Na (Hey, hey, hey), mock tomahawk chop, taking care of business, and a rocking edition of the Hokie Fight Song.
All of this makes me happy.
The group could probably cover some of Antone's raps as well.
I approve.
And the Harlem Shake.
Only if Frank Beamer is one of the dancers. No. No. I take that back. Only if Frank Beamer is the main dancer.
Bass man here too.
Commandmants of the bass: 'thou shalt not screw up the groove. Screw up the notes, but never the groove.'
'Thou shalt not lust after the guitarist's part. They are keepers of the fun, you are keeper of the groove.'
I've got the funk. I keep it in a box here in my dorm room. First time I saw it I just kind of lost my mind, and I started milking it. Made myself a funkshake. Started seeing around corners.
But yeah man, grooves on grooves on grooves is my motto.
Turkey leg for being Homozygous funk positive.
I'm not sure if I should be jealous, or in awe. I'll just be both.
I like the concept of homozygous funk positive. It makes me happy in my groovy bits.
Homozygous Funk Positive, exhibit A:
beginner bassist: key of A-major 1-4-5
lead guitarist: no man, G-major: 2-5-6
we get to the same place :)
and another
Upvote for being a bassist. Bassists unite!
We shall henceforth be known as the keepers of the funk.
Guitarist here also. Great post. Just for shits and laughs someone should post up the family guy "douchebag with a guitar song"
Crap. My daughter touched the screen on my phone. Please uptick this to counter my egregious error.
Zach, great write up. Extremely well done and an amazing effort.
According to your profile, it says you were Logan's back-up in high school. Is this true? If true, did all the "casual fans" of Brookeville HS call for you to replace Logan as starter?
Yes, that is true. You can read more about that here.
My older brother always likes to make the joke that if it wasn't for Logan, I would be Tech's starting QB (couldn't be farther from the truth).
Great analysis and great article! Whether you love him or not, we are going to miss Logan next year. I can't wait for the rest of this season as he keeps shutting his critics up with nothing more than a smile.
Thanks for the link, I hadn't read that. I'm impressed, keep up the good work. Reminds me a little of single-A Strasburg back in the '90s, where we "introduced" passing and kicking my senior year. Why weren't we going for 2??? Our offense, which got us to state final in '95, was pretty much "hand off to Frankie (Shoemaker)!" for 4 years. He got drafted as a baseball player for the Royals organization out of high school.
The more I read about LT3, the more I like the guy.
Zach, this post could be crucial to the NFL for next year's draft. You just settled a strenuous ongoing debate I've been having with someone. Very well done.
Well done, man. I really dig this post being a very math minded person (engineering). Being an engineer and seeing the lowest WRE (wide receiver efficiency) at 77.7%, I'd like to see a bit more disparity between the 102.5% danny and the 77.7% Randall Dunn. It's just my opinion of course, but I'd honestly give negative credit for drops on easy receptions because, as you defined it, these are "expected to be caught by any pass catcher at any level." In addition I would give the zero credit to the should-be-receptions that are dropped, yes they require receiver adjustment, but the receiver has to make that adjustment.
I kind of want to play around with this formula now...It's just me personally but I have never liked QBR where they can get a rating of 148. Blame it on being an engineer, but 100 is best.
Ok so I just spent some time trying some stuff in excel using your numbers:
I basically said that the reception values stay the same (1, 1.11, 1.33) but I assigned values to the corresponding drops (-0.5, -0.25, 0 : ed, sd, td). This was an easy adjustment, since making the drop coefficients positive would raise efficiency above 100%. The values are easily changeable since I have this in excel.
D. Coale had a rating of 101.03 % (I'm not sure why I didn't match your value, Zach, I matched all the others) before the adjustment, and a 96.4% after the adjustment, so that is still a very good number.
W. Byrn 90.8% adjusted to 79.68%
D. Knowles 80.89% adjusted to 66.26%
J. Stanford 88.25% adjusted to 74.79%
D.J. Coales 79.08% adjusted to 61.44%
Like I said, these are small adjustments that make ME feel better about the efficiency numbers, so in no way am I saying this is right...I just felt that our receivers were more in the C-range than the B+ to A range. But there is a definite correlation between this and LT's numbers
Really good write up, I enjoyed this a lot.
Zach, first of all great job, I'm a huge proponent of analyzing catch efficiency rather than just pass efficiency. This kind of thinking is what brought about some of the coolest statistics in baseball: WAR comes to mind...Anyhow, I, like others on this string, think that this type of analysis could have huge implications for the NFL draft. If you have access to a math/stats whiz it may be worth sitting down and discussing your methodology. He/she may be able to provide insight into how to use "big data" on passes/targets to really quantify the way you are weighting each category in the equation...food for thought. Well done and thank you for putting in the time an effort for this analysis.
I was actually thinking the same thing. But the data obviously has to be taken with a grain of salt...It's like the idea of moneyball vs. trouble with the curve. Numbers can get us to a point that our eyes have to then validate or invalidate. D. Coale has a crazy efficiency but he isn't on a team, whereas Jarrett Boykin (who I would guess wasn't as efficient) is going to be starting for Green Bay on Sunday.
Danny Coale has also been seriously injured twice in 2 years
Great analysis. If there was some way to grade the route as well as catching I bet there would be an even starker contrast with that 2011 group of receivers compared to now.
But the difference in receivers at simply catching the ball is very apparent. Before the GT game I re-watched the 2011 GT game, possibly Logan's best game, and it was apparent that much of his yardage came from throwing it up for grabs into single coverage down field. In that 2011 game, these resulted in catches. When he did the exact same thing early this year, the result was usually an INT.
When I looked at the raw data, almost all of Logan's interceptions were rated as 'out of reach' throws.
Now, whether they were out of reach because Logan mis-threw the ball versus the receiver not being in the right place or not adjusting appropriately is debatable.
This is an incredible post, awesome work brother! And it proves that Danny Coale is still open!
Awesome film breakdown. Incredible actually.
On a side note, based on the amount of time this had to have taken, you should really try and get out more. Go play hoops up at War, throw the pigskin on the drill field or go chase some tail downtown. That being said, outstanding post.
I'm a married college student with four jobs. I'll relax when I graduate (with a resume that stands out) and am finally able to just work one.
That being said, thank you for reading.
More power to you, brother.
You should attach your Key Play posts to your resume. You'll make Chief Emperor in no time.
Probably should stay off TKP since this will become like job #5. Not in a bad way, just so much stuff on here it can be time consuming.
Yeah TKP is kind of like Reddit for me. Before I knew about it I was great! Now, I like it too much and I spend FAR too much time on here...
Turner's FTW!
It sure is nice to see data instead of going with gut feelings from whatever the most recent game was. This is a metric worth tracking. I'm curious if you could do something similar with our running backs in the last couple of seasons regarding making the correct read and taking the available yardage.
Beautiful insight. I've always thought highly of Willie Byrn. All season I've been referring to him as the poor mans Danny Coale, and in no way is that an insult. He may very well end up as the Hokies best receiver over the course of this season and the next.
vtae13, I'm onboard with your adjustments to reflect the grade scale we're all accustomed to. Knowing a receiver can be given a C performance or a B+ performance is easier to follow.
So why don't we use JCC more? Efficiency isn't everything, but you think his speed and hands would count for a lot.
where is boykin?!
He's on the list, at No. 9 with an 86.9% catch efficiency.
#9, 86.9% in 2011
Starting this week for the Green Bay Packers
Great stuff!
Wow, this is just awesome stuff.
As somebody who has been reading up on Sabermetrics lately, this made me drool a bit.
Nice analysis! It leads me to believe we can develop a stat for for those passes "out of reach"? We would have to exclude "throw-aways" based on no one being open and simply throwing the ball away to avoid a sack. I would like to see a real number on "out of reach" where there WAS time, a bona fide throw to try to complete, only to be out of reach.
My guess is Logan still has way too many out of reach passes. He has improved, but he has also missed some really easy gimmes where the receiver was wide open and he just was not accurate.
Zach, if you have the patience and interest to carry out this type of project I recommend you read the Sartin Methodology, specifically Tom Brohamer's work on pace in horse races. You would probably love it if you could find the time on the side. There is money to be made. Still can't get over how awesome this post is.
Thank you, sir. If I ever find the time, I will look into that. Sounds like it could make for an interesting read.
Awesome job. Really impressed with this approach. Only thing I would love to have seen was in the section you put the 2013 by game ratings if you could have put the stats that went with it next to them. A guy might have been 100 but he may have only seen two targets. Just gives us a better idea of the game to game impact that each receiver had. Also would be great to see a section with the receivers in 2013 listed by total targets along with their ranking to see if the game plan is leading Logan to throw to his more efficient receivers or not.
Thank you for your time and effort putting this article together. The amount of research in this piece is really incredible and I absolutely love your concept of quantifying how receivers perform outside of their quarterback.
But even you admit that your equation is rudimentary so I'm going to throw out my ideas for improving it because 1) I love college football stats, 2) I really don't feel like doing my engineering homework, 3) Go Hokies and 4) fuck it.
So the biggest issue I see with the formula is that a receiver cannot negatively affect his score, he can only not positively affect his score. Take David Wilson is supremely talented but he can be prone to fumbles and tries to make too many things happen which more often than not dont work. And even when they do work, that means hes more tired for the next play so he either gets substituted out, he cant give maximum effort or the OC is forced to change his play/the opposing DC is more likely to play the pass. These negative stats would need to be kept on a per game basis and become exponentially more important with each successive negative play.
I firmly believe this principle was on full display against Alabama. If the receivers had been catching the easy passes, Loeffler would have called the game much differently and Kirby Smart may have had to taken another defender out of the box which opens up the run.
I looked at the stats you wrote down and unfortunately you dont have the other things Id like to include which are what Im going to call key play opportunities (pun fully intended) such as receptions or drops in the end zone/that would obviously lead to a touchdown, catches while losing, two-minute-drill offense and third down conversions. If youre going to give additional points to tougher catches and drops then you should apply the same principle for clutch players. In my opinion, a receiver should be given a fuck ton of negative points for dropping the game winning touchdown while hes wide open in the end zone as time expires. You could even throw in the relative importance of a game (championship game easily catchable drop as a massive underdog vs. easily catchable drop against a FCS team). Im definitely aware of how long finding all of these stats would take so Im by no means saying that you should go back and look at every pass play from Logan again but I do think that these should factor into the overall catch efficiency.
Originally I was going to say that you should factor in yards after catch/the receivers separation from the defensive player but that would be more of a wide receiver efficiency as opposed to catch efficiency. Id envision a wide receiver efficiency to be weighted similarly to your original formula with simple weights attached to catch efficiency, yac/separation efficiency, caliber of the defense played and, ideally, some way of compensating for the offensive system the receiver is in so you can account for the ratio of blown coverage to the amount of stress a defensive back has on a given play. In an extreme example, a guy who is targeted on every pass play but is Danny Coale open on every one of those plays should receive a lot more credit than a Gah Tech receiver getting open on blown assignments. The same principle can be used for trick plays like the halfback pass. This would also be where you could account for run blocking and route running.
I also cant decide whether this falls under catch efficiency or wide receiver efficiency but I also think that interceptions thrown while targeting a specific receiver should factor in.
Like I said in the beginning, I absolutely love college football stats. True story, I couldnt fall back to sleep this morning so I went stat hunting and found some really incredible stuff (check out Tulanes defense and Floridas offense this year if youre interested). Im an engineering student so Ive worked with statistics, certainly not at the level of a stat or math major, though Ive always loved and fantasized about the prospect of creating a computer code that would call offensive plays in a hurry-up setting. I read through the comments and I know youre an absurdly busy man so Im not sure if these types of statistical analyses are something you can realistically pursue going forward. However, if its something you or anyone else would like to continue doing, Id be more than happy to partner with you to split the work. Unfortunately I don't think I'll be able to churn out hardcore analysis by myself during the school year because of my 18 senior credits but I think this kind of quantification will be the next big push in football much like Sabermetrics was to baseball.
Math is power!...and all that other corny janx my elementary school teachers kept talking about. Thanks again for putting in all of that leg work, we all really appreciate it.
I can't help but be reminded of Professor Frink....
But seriously, nice job. I admire the LPD attitude it took to do the research, crunch the numbers, and write the post. This post puts a quantitative measurement to a trend/trends I have observed, as well. In other words, I came to the same conclusions, but didn't have the numbers to back it up.
As for LT this season, he is trending much like Tyrod did in 2009, with the incredible finish of the Nebraska game, being the turning point for him that season. He had only completed about 45% of his passes up to and really through that game. After those last two or three minutes, and throughout the remainder of that season, he was completing around 58%, I believe, thereby cementing the legend of the Tyrod we know and love and remember to this day. Logan still has time to follow in Number 5's footsteps.
Excellent Frink reference.
So, using Zach's raw data, and looking up a bit from ESPN, I made the following:
This first chart shows my tweaked version of his "Efficiency Scores" vs the Yds/Rec of each reciever. To tweak his formula, I changed it to E = (all receptions/expected receptions), where expected receptions = 100% of all easy catches, 80% of all moderately easy catches, and 50% of all tough catches.
Some interesting notes from this plot:
This next chart compares the percentage of passes "Out of Reach" for a given reciever, against how many yds/rec that reciever got:
As I said earlier, I think this percentage is a pretty good reflection of how effective the WR/QB relationship is, and might be a decent reflection for the offense as a whole. A healthy offense, with a comfortable qb, who is not behind the chains, and has a good rapport with his WR should generally be hitting his WR. Obviously, the longer the routes the WR runs, the more often hes going to be overthrown, but generally speaking, this number should be indicative of a healthy offense (I would think). The thing I think that's really apparent from this chart is how much Logan struggeled getting the balls to his WRs in 2012, and how now we are back pretty close to the 2011 levels. I'm not pointing any fingers, because we've had this discussion a million times here, just trying to show with math and stuff that we are defintely seeing a return to 2011 form in effectiveness for the entire passing game. Maybe not quite as deep, but perhaps even more effectively in the medium yardage range.
Awesome, but I can't pretend to get all that you presented.
Interesting to see gut observations backed up by data.
We knew Coale was special, but he was 'outlier' good.
I looked at the data charts and our worst percentage pass was the deep left out. We only complete it about 25% of the time.
Also, this year's GT numbers were crazy good. Logan was a machine that game.
One thing that's pretty easy to see in the 2nd chart - the slope of the line tells you how quickly our effectiveness falls off with deeper shots down the field. A steeper line means you rapidly lose effectiveness as your QB tries to reach deeper receivers. It means more deep incompletions. So 2012 really sucked in that regard - as we go downfield, we miss more passes. 2011 was great in that regard - we stay effective as we lengthen the field and are able to go deep more reliably. 2013 is closer to 2012. So, hooray!
really interesting analyses - the catch efficiency is just awesome. Trying to get trends and make decisions is just as interesting but the degree of difficulty goes up tremendously.
in terms of trends, there is a practical limit to getting trends out of data that isn't always reflected in the math. Sure, you can fit a line/trend through anything but the questions is how meaningful is it? On one of the last airplane programs I worked, an over-zealous engineer fit some curves through a smattering of flight test data that showed the air vehicle was 20 - 43% down on range. After the dust settled, everyone's conclusion was that the data was too scattered and a change of fuel flow meters tightened up the data and real trends were able to be calculated.
If I look at the charts above, I see an awful lot of scatter. For example, if you throw out the outlier of the second chart ("sir tripsalot" at 39% and 13yds/catch) 2012 could be trended to show BETTER % out of reach numbers as yds/catch increases.
Also, 2 data points for 2012 at 19 yds/catch with an average of 22% out of reach is a good indicator that the trend does not accurately reflect what is going on.
Anyway, just my $0.02 that it would be really difficult to conclude much in terms of trends from this data.
For the truly bored or energetic, take the charts above and make the marker size a function of the number of catches for each receiver - that "bubble chart" would give us an indication of the relative importance of each data point. Also, the trends through the data could take into account the number of catches/data points to graph the line ....
Yeah, I know. Was running out of time yesterday, and wanted to get it up before I had to leave. I came back and did some weighted averages, and I think they illustrate the point very well (note that these are weighted by total targets, not completions - I wanted to see who we were throwing at and how effective that was):
2011-
Average Completion Score - 91.65
Average %Out of Reach Targets - 21.3%
Yards/Reception - 13.87
2012-
Average Completion Score - 85.95
Average %Out of Reach Targets - 24.73%
Yards/Reception - 14.68
2013-
Average Completion Score - 87.76
Average %Out of Reach Targets - 22.1%
Yards/Reception - 12.50
So the general idea that we threw deeply and ineffectively in 2012, and are shorter and more effective this year holds up.
hmmm - maybe - did you look at 2011 v. 2013. If I understand your numbers, in 2011, we were throwing deeper on average than 2013 yet completing more passes with fewer out of reach. Anyway - good stuff.
btw, I looked at the raw numbers and because the yds are not recorded for the incompletions, I could not create the bubble chart I was talking about.
I know it's early, but I still think Willie Byrn is the next Danny Coale. Willie is still young, so I wouldn't doubt if he gets into that 102% range at some point, probably sooner rather than later.
I've got hopes for him and Josh Stanford. Also Kalvin Cline. It's weird watching our young receivers beast mode while our experienced receivers struggle.
Interesting point.
I wonder if its those guys having to unlearn less for the new offense.
I'm sure that has something to do with it. I think DJ is struggling, because he has always been hurt. In reality, he isn't a true senior when you think of it.
Then you also remember, that a lot of them are young. I mean, Knowles is a soph if I recall.
Stanford looks good. He has some great athleticism. Cline is a beast, and its great to think he is a freshman, and can only get better as time goes on. Gets even more exciting to think that we have more TE's in the system too.
I love Willie, he really reminds me of Danny. He is just there and makes some crazy catches. no fear of getting hit. His style reminds me a lot of coale.
Honestly, if we can get the running game up and going, and with some of these kids in the pipeline (Williams for instance), QB may not be that big of an issue next year. There are going to be a lot of good offensive weapons for a few years to come. Look at all the different WR getting touches.
great job, thank you for the effort.
I'm glad somebody is good at stats....great post!
Since I can't give a turkey leg for a post, here's one for a reply!