By the Advanced Numbers: East Carolina

What does math say about Virginia Tech's chances to avoid a three-game skid against the Pirates.

Last week, By the Advanced Numbers predicted a 27-10 Hokies victory over Boston College. Of the many people who felt the prediction was too generous, all thought Virginia Tech was getting too many points. Needless to say, even that rosy forecast turned out to be far short.

As promised in that column, irrational optimism may now be fully engaged.

The Hokies' struggles against East Carolina over the past decade are well-known by fans, but none of those teams featured the most dynamic offense in the country and a defense that has allowed 0 ppg over the past five days. The national championship is inevitable, but sometimes it takes some time for the numbers to catch up to the obvious, so let's see what the ever-rational statistics say about the contest.

Note: Detailed offense, defense, and special teams ratings are not available until midway through the season. Until that time, "By the Advanced Numbers" is an abbreviated column.

Where the Hokies Season Stands

Virginia Tech is currently 2-1. The actual and predicted score difference and odds of winning each game are:

The odds of each possible regular season win total are now:

In 10,000 simulated seasons, Virginia Tech finished 11-1 twenty-five times. Just sayin'.

Rankings and Computer Predictions

The computer rankings and predictions for each team:

Despite one computer apparently loving a healthy serving of short passes and ranking the Pirates much better than all others, every prediction has Virginia Tech winning the game. What could go wrong? The odds of a 11-point favorite winning are 79.0%.

Next is a look at any overall offensive or defensive advantages:

Somehow the offense's rating dropped this week, and I've been unable to get an answer from the creator of S&P+ on how that could be. The Football Study Hall Advanced Statistical Profile of Virginia Tech indicates that the Boston College game was the 86th percentile of all offense performances for the season. The good news is that the defense has moved up into elite territory, and the Hokies have a large overall advantage.

Who To Watch Out For Against East Carolina

Oh hey, Shane Carden is gone, so who is this year's high-percentage East Carolina passer:

  1. Senior QB Philip Nelson is completing 79.8% of his passes and has the 24th-highest passer rating in the country. 13 spots behind Jerod Evans. Oh, and he's not a runner either.
  2. WR Zay Jones had 22 (!) catches last week — many on screens — and leads the country with 24 receptions of 10+ yards.

BONUS! Who NOT to watch out for:

  1. WR Quay Johnson is 87th in the country in kick return average, but what does it matter because Joey Slye is No. 1 in the country with a 65.7-yard kickoff average and 90% of kicks going for touchbacks.

Statistical Key to the Game

Zay Jones is not the only WR who will get the ball, and as usual with ECU the key will be limiting big plays. The high completion percentage is manageable if the many screens are getting stopped at or near the line. So plays of 10+ yards by the Pirates will be the statistical key to the game.

The Stats that Define Virginia Tech's Season

Revisiting my column from this summer on the stats that will define the season, here is how the Hokies performed thus far (each listed as VT - Opponent):

Opponent PPP Pace (sec/play) ToP
Liberty .40 - .21 22.7 - 25.5 33:42 - 26:18
Tennessee .33 - .71 26.1 - 26.1 31:44 - 28:16
Boston College .64 - 0 27.5 - 26.1 35:15 - 24:45

I will not value time of possession.

Statistical Prediction

Let's see, how is it this is calculated? Oh yeah...

Computer prediction + emotion - East Carolina WTF factor = Virginia Tech 34, East Carolina 20.

As always a thanks to Football Outsiders, cfbstats.com, and Minitab Statistical Software.

Comments

Big recruiting weekend for us too!!

5* WDE KJ Henry and 4* ATH Tre Turner are coming up for the game this weekend. VT will hopefully put on a show for these two

Edit: 3* RB Marcus Caldwell will be making the trip as well. Let's recruit some NC talent!!

Keep calm, Gobble on

Bring on the 4's and 5's!

Dat Massey ECU outlier tho... 10th?

I have no idea why my username is VT_Warthog.

Arkansas blew a 24-0 lead in the Belk Bowl.

...by a ranking system that has Stanford 21st and South Florida 5th.

Didn't Massey say his system doesn't really mean much until later in the season?

From his FAQ:

The BCS compliant ratings do not use preseason information, so everyone starts at zero. A team's rating may look funny or fluctuate wildly until there is enough evidence to get a more precise measurement of the team's strength. As games are played, the computer gradually 'learns' and the cream rises to the top.
For the main version, preseason ratings compensate for the lack of data early in a given season. They give the computer a realistic starting point from which to evaluate teams that have played zero or few games. This limits dramatic changes that could be caused by isolated results not buffered by the context of other games.

The effect of the preseason ratings gradually diminishes each week. When every team has played a sufficient number of games to be accurately evalulated based on this year alone, the preseason bias will be phased out.

Preseason ratings are based on an extrapolation recent years' results, tuned to fit historical trends and regression to the mean. A team's future performance is expected to be consistent with the strength of the program, but sometimes there may be temporary spikes.

Other potentially significant indicators (ex. returning starters, coaching changes, and recruiting) are ignored. Therefore, preseason ratings should not be taken too seriously,

Good research, but the low value isn't actually Massey's rating system. The data comes from his site (hence the link) which compares a huge number of computer ratings and combines them into a composite rating.

I think the one I mentioned had VT in the 50s, which isn't as unreasonable as its ECU and USF rankings.

I have no idea why my username is VT_Warthog.

Arkansas blew a 24-0 lead in the Belk Bowl.

I like that we scored infinite times as many points per play as BC. I'd like to do that more often.

Warning: this post occasionally contains strong language (which may be unsuitable for children), unusual humor (which may be unsuitable for adults), and advanced mathematics (which may be unsuitable for liberal-arts majors)..

So our odds to finish 11-1 have multiplied by 25 since last week. I approve.

I also enjoy seeing that we allowed 0 ppp for a whole game

Virginia Tech School of Architecture Class of 2014
Fan of Hokies, Ravens, NY Giants, Orioles

I'm just glad our odds are higher to have 11 wins than 1.

Warning: this post occasionally contains strong language (which may be unsuitable for children), unusual humor (which may be unsuitable for adults), and advanced mathematics (which may be unsuitable for liberal-arts majors)..

I wonder what UVA's chances to go 11-1 this year are?

Technically, they could end the season 11-3 but they would have to win out (ACC championship and bowl game).

So it's not yet 0% but it's probably .00001%

.....
95% of what I say is sarcastic. The other 5% is usually taken out of context.

Somehow the offense's rating dropped this week

How?

No, really, how?

That seems like either a math or an input error.

"I mean, you know, fuck them, but good for them." -Too Druck to Funk

That seems like either a math or an input error.


Even a business major knows there's a fly in the ointment on this calculation. I agree.

Yep, sand in the vaseline.

Reel men fish on Wednesdays

No idea how and Bill C. has no insights on it. This early in the season things can bounce quite a bit...keep putting up 49 a game and it will definitely climb.

Only thing I can imagine is that somehow the quality of our first two offensive performances took a hit due to our first two opponents' defenses playing horribly this past week. Ohio only scored 19 on Tennessee and SMU put 29 up on Liberty, both a little higher than expected but not terribly so, but that doesn't necessarily tell the story of the more advanced stats that the S&P+ looks at, so who knows.

I love these

If a tree falls in Scott Stadium does it make a sound?

I'm admittedly a nerd, but this is my favorite feature article.

"I mean, you know, fuck them, but good for them." -Too Druck to Funk

Computers seem to really like UNC still. They are not passing the eye test.

They don't pass tests as a rule. Sort of against their culture.

Well its supposed to be the tutors passing the tests

Did they even take the test?

You have to take a test to pass one.

Reel men fish on Wednesdays

I mocked your prediction of VT getting 27 against BC last weekend. Mocked it. Openly. In a british accent.

But you are low this week on VT at 34. 40+ or bust! 'Merica! Pirating is illegal!

you know if we don't score more than 34 points everyone is going to blame you, right?

I know I am

If a tree falls in Scott Stadium does it make a sound?

I thought we would score 16 last week. Clearly, I'm not so good at this.

well hopefully you're always low

If a tree falls in Scott Stadium does it make a sound?

oh for sure

via GIPHY

Come to Blacksburg and see what the Hokie Pokie is really all about

...and so is losing to Eastern Carolina

Warning: this post occasionally contains strong language (which may be unsuitable for children), unusual humor (which may be unsuitable for adults), and advanced mathematics (which may be unsuitable for liberal-arts majors)..

That was the loudest fucking ad (psa? who cares) of all time. Blow out the windows in the whole house loud.

"I'm a Miller Lite guy, always have been, since I was 8." -John Daly

Are we talking Cockney, Liverpudlian, Geordy, upper crust, Dick Van Dyke's murdered Cockney, Wiltshire, Brummie, South Wales (be still my heart)...what kind of accent were you mocking in? It's going to drive how I judge you.

At best, I can do a poor mans Chris Geere.

I wish it was full Brad Pitt from Snatch. Keep working on it!

but none of those teams featured the most dynamic offense in the country and a defense that has allowed 0 ppg over the past five days.

WOOO I'll ride this hype train wherever it takes me. Last week was too fun for my usual pessimism

By my count, were allowing 0 over the last eleven days.

"I mean, you know, fuck them, but good for them." -Too Druck to Funk

What does the x-axis represent in the S&P Ratings chart and what does the y-axis represent in the Computer Ranks and also the Computer Predictions and Spread charts? Am I slow or do these not make complete sense?

I've always wondered the same thing. Best guess and I could be very wrong: S&P x-axis doesn't mean anything. When he creates the plot, he has 125 data points. The x-values are 1, 2, 3, ..., 125 and the y-values are the ratings. Not sure how he has the teams sorted. The computer predictions is basically the same thing, but flip x- and y-, and then he must have an ID offset to separate the two teams. I guess he does it this way because it is easy, and maybe because it is a little easier to read with the data points being spread out.

IDK though. Would love an answer to this as well.

Main takeaway, our O is finally catching up to our D, but Bud doesn't want to be outdone so he has put out quite the D this year!!

--
"It's time to go play Virginia Tech Football longer and harder than anybody else in America!!" -- Justin Fuente
"I put a brick in Sacksburg today." -- Cam Phillips

This is correct...random jitter is added so that it is easier to differentiate the points. There is literally no meaning to it and it is completely random noise added in that direction.

Would it be possible to have the x axis represent the teams rank (s&p, ap, whatever). It'd be interesting to see, and I'm often curious who some of the outliers are

This is sort of along those lines, but one thing I'd like to do in the next column update would be to plot them on the same graph, like offense on one axis and defense on the other.

Ah yes, that would be interesting to see. Thanks!

I think there is only one Axis, the y-axis. Because it would be difficult to see one column with all 125 dots, the spread across the x-axis is so that the 125 teams are not all on top of one another. For visual purposes only.

🦃 🦃 🦃

Minor question, but is there any rhyme or reason to the color/shape of the plot points in the first graph? God I hope there isn't any obvious or easy answer, cause I can't for the life of me see any correlation amongst the like colored/shaped points...

"The TKP community is unrivaled."
-Justin Fuente, probably

I think maroon circle = home, green diamond = away, and black circle = neutral site.

Correct!

Thanks guys! Guess that has relevance because home field is given some weight in these predicted outcomes? Or it doesn't, and you show it that way so we can add our own mental weight to home field if we choose?

"The TKP community is unrivaled."
-Justin Fuente, probably

A good analytical mind! Good job, you earned you diploma.

The green diamonds are expected wins.
The maroon dots are previous or projected wins by large margins.
The black dot is a projected (or previously occurred) loss.

11-1. No brakes.

It was a catch

The snap pace going down is interesting. I thought the offense would need additional time and reps to gradually increase pace as the season went on. I assume the rate dropped against Tennessee because of the problems and the last thing you want to do with fumbling problems is ask people to press harder. Against Boston College, maybe the rate dropped simply because we didn't need to go hurry up. Fuente has stated time-and-time again that he doesn't want to go fast at the expense of the defense and I can only assume that was at play.

I think you're right about BC. Once you're up by 4+ touchdowns, it doesn't hurt to take your foot off the gas a bit and burn some clock.

It's not a perfect measure either, because it is clock time and not actual time. So when you run the ball you decrease the pace due to the clock continuing to run even though in actual time you might line back up and run the next play very quickly. Against BC, the Hokies rushed on nearly 70% of offensive plays which definitely drove the pace down. And of course there was more clock draining later in the game.

Oh, and he's not a runner either.

Oh, we'll see about that...

If you're reading the above post and thinking, "is this guy serious?!?," you can safely assume I'm not.

We only have a 97% chance at beating LOLUVA??!?!!?!

WTF? That's ridiculously low IMO

If a tree falls in Scott Stadium does it make a sound?

Give it time grasshopper, there is room for us to continue to get better and for LOLUVA to get worse.

ECU can kiss my axis!

Ok...I'm ok now

I wonder how much the computers will hate GT after last nights debacle.

I hate them enough already.
And stop calling me "computer".

not necessarily a debacle..they lost to the #5 team by 19 points.

Do I need to remind you that we lost to the #17 team by 21?

If a tree falls in Scott Stadium does it make a sound?

no, no you don't. But you decided to do it anyway.

Warning: this post occasionally contains strong language (which may be unsuitable for children), unusual humor (which may be unsuitable for adults), and advanced mathematics (which may be unsuitable for liberal-arts majors)..

Yeah, sorry 'bout that. Wasn't thinking

If a tree falls in Scott Stadium does it make a sound?

Only if you would call that anything less than a "debacle."

"Our job as coaches is to influence young people's lives for the better in terms of fundamental skills, work ethic, and doing the right thing. Every now and again, a player actually has that effect on the coaching staff." Justin Fuente on Sam Rogers

Just had it pointed out to me... The line for the game this week opened at VT being a 7.5 point favorite. The line is currently VT as a 12.5 point favorite.

Pretty interesting that Vegas has seen a swing like this. Someone is clearly seeing something about this game that makes us look like a heavy favorite. I'm starting to get cautiously optimistic about this one.

Don't let this comment take away from the fact that Arkansas blew a 24 point 2nd half lead in the Belk Bowl.
Don't let the Belk Bowl take away from the fact that Matt Ryan blew a 25 point 2nd half lead in the Super Bowl.