Unless you're new to The Key Play, you've probably seen the advanced statistical previews done before each football game. To prep for 2014, we need to know what kind of information is interesting!
To see an example of last year's go to http://www.thekeyplay.com/content/2013/november/14/advanced-numbers-mary....
Let me know what you like, what you don't like, what you don't understand, and any other rambling thoughts you have about statistics. One thing I'd like to add is more individual statistics rather than just team stuff but what do you want to see?
Let's set the gold standard for analysis in 2014!
Forums:
DISCLAIMER: Forum topics may not have been written or edited by The Key Play staff.
Comments
How about more rushing yds/att, more TOs from the D and better coverage from STs.
I think he's asking what stats you want to see analyzed, not what you want the stats to be
and any other rambling thoughts you have about statistics.
That's another thing I'm great at.
But you can't disagree with his demands...
Had to think about what an ST was for a second. Though maybe we had developed some hybrid player called a Super Tackle. A defensive tackle that is really fast and has to cover TE's. Then I realized it was special teams...
We have that phenomenon. It is called Corey Marshall.
I saw The Corey Marshall Phenomenon open for Gwar back in '09 at the PA Electric Factory.
any gwar reference on this board gets a leg
Here's a stat I'd like to see/hear more of: MORE GAMES WITH "STICK IT IN" CHANT BROUGHT BACK.
Sadly that's a graph that is unlikely to need updating during the season.
I want to see the stats of how many times timecop tries to call 4 timeouts in a half.
Woah, woah, we don't want to cause a system overload.
Early in the season it may be tougher to do, but I'd like to see how the offense and defense do against opposing teams vs the mean. It's more a post game analysis, but for instance, how many fewer yards the defense gave up to an offense than the average team. And then average those and I guess doing the same for opposing teams, just to see the impact that a team has on its opponents. Does that make sense?
Yes - maybe not those exact numbers, but the offensive and defensive ratings currently used are all adjusted in the way you mean....in other words, give up 14 to Florida State and your defense has a significantly better rating that giving up 14 to Virginia. Well, 14 to FSU is probably seen as better than 0 to Virginia.
That would probably be a pretty difficult stat to work with. I mean unless you knew some actual probability distribution for the points of each particular football game (and if you do you should be a millionaire somewhere), it would be pretty difficult to analyze if deviations from the mean of that team are due to random chance or not. Especially at the beginning of the year that stat would be particularly misleading.
It is true that asymptotically the average would follow a normal distribution, but without knowing some actual distribution its pretty difficult to talk about the rate at which that average approaches the asymptotic distribution. That makes it difficult to talk about how you do versus a team from any inferential standpoint.
I am sure there are techniques to discern whether you are doing better than random chance would indicate, but I haven't learned any up to this point.
Skipper Shots vs. Number of Turkey Legs Consumed. I have a theory that we have been conditioned like Pavlov's experiements.
-1,000 yard receiver
-500 yards from a TE
I think both are reasonable and means the PASSING part of our offense is progressing.
Those number will be tough with the inexperience at QB, but I really think Stanford can be big for us this year.
Some predictive stats prior to the game so we could evaluate the QB (whoever it is) and passing game would be interesting. For instance:
Also, some other similar metrics to evaluate the offense as a whole might be useful:
These metrics compared to our final performance in the game would give us a pretty good idea about where we actually stand offensively, and where the weaknesses are. You could also go one step further, and by adjusting for our metrics, provide our expected stats. Then, if we're exceeding those, we'll know if we're improving/having a good day.
A final thing that you could do, would be to compare our offense last year to this year using the predicted stats. IE - our offense this year was in the Xth quartile for completion percentage. This D gives up on average a 52% completion rate. Our last year's offense would probably output X. This year's offense output Y. Hopefully Y>X, although maybe not in every area. I think it might make for an interesting comparison. A similar study for our D would be interesting as well.
EDIT - I think these should be adjusted for opponents as well. Not sure I made that clear, and I'm not sure how, exactly. Maybe some sort of iterative ranking system?
Wait a minute...lost in the multiple comments listing actual numbers people want to see is this useful comment! Thanks Hokieshibe!
Everyone else: I clearly want the actual numbers to be good, but I'm looking for responses like "It would be good to see each teams top three receivers versus the opponent's top three cornerbacks" and not "It would be great to have a 2,000 yard rusher and a 90% completion rate and a 1,500 yard receiver".
It's been funny seeing everyone ask you for production stats for next year though. I personally would like 5 Forced Fumbles by our secondary next year please.
I'll raise you 5 Forced Fumbles by our punter.
What about the place kicker?
Good point. #AJHughes4Placekicker
Man, reviewing those Maryland stats...still can't believe we lost that game.
I'm having a tough time thinking of team stats that would add to your analysis without reinventing the wheel.
For individual stats, maybe track "catching efficiency?" It'd be a lot of work to compile that data, but it's an interesting stat that adds some local TKP flavor.
10,000 yards passing, 5,000 yards rushing. Just to be modest.
If I had to pick one reasonable stat - a 1200 yard running back. Second choice - >50% third down conversion. Third - opponents TO : our turnover ratio - better than 2:1. Fourth - score more than 95% of times we are in opponents RZ. 5 - our beamerball plays to opponents beamerball plays ratio of better than 4:1.
Okay, I will stop wishing. I want my old Tech game back. As in scoring 55 points on GT. Or running the ball in with ease for a TD within their 5 yd line.
Edit: after all this, I realized that all I want for next year is a team that fights tooth and nail to win. When the other team is on Offense, they need to feel like every yard they gained drained their life. When their defense is on, I want them to be gasping for air between plays. When they come out to punt or kick a FG, I want their punters and kickers to lose composure thinking of our STs flying in. I want visiting teams to go weak in their knees when they hear Lane stadium roar. And I want to beat the heck out of UVA and show recruits that Mike London can recruit all he wants, but the real deal is in Blacksburg. That's what I want more than any statistic next year.
14-0 ..... yep, that'll do it
This one right here. All other stats are secondary to the W column.
Don't you mean 15-0? Don't settle for 14-0.
This may be too much work, but for each opponent it would be interesting to see how their previous opponents performed relative to their season average. Obviously this would only become meaningful after a number of games have been played.
For example, when we play Miami we would examine all of Miami's previous opponents. Let X be each of Miami's previous opponents. For each x in X, was x's offensive/defensive performance (using various metrics) better or worse than their average against Miami. I think that would give a better indication of how our upcoming opponent (in this case Miami) is performing. If Miami previously played UNC, FSU, and TT then Miami giving up more yards than the national average doesn't mean a whole lot. Miami holding those teams to 20% below their average does.
Thanks - this is like the one above. The offensive and defensive rankings I use are actually doing that...everything is opponent-adjusted so the numbers are effectively based on a big "web" of all games played to that point during the season. But since two people have suggested this now, that points to the previews needing to do a better job of communicating what the hell people are looking at!
I think a short primer before the season starts would be good, especially for people who aren't used to looking at more complex breakdowns. Then you can just link it to any write-up you do, so new people who come along can jump right in.
Points per possession
% of possessions that resulted in points
Odds of opponent getting trucked by Sam Rogers: 100%
Odds of Juice Williams flattening some poor, unsuspecting opponent's safety: 100%
Odds of one game of #ALLMAROONEVERYTHING: 100% (In Whit I Trust)
Odds of opponent being shamed by a Fuller: 200% (because Kendall doesn't play by the same rules of probability)
Odds of opponent dropping a punt out of sheer fear of AJ Hughes: 100%
Odds of Bucky Hodges growing to over 10' by the end of the season: I think you sense a trend here.
Done and done.
Greatest. Post. Ever.
don't know how hard it would be...but for individual stats, comp % when thrown at for the defensive backs, and when thrown to for WR/TE/RB
I don't expect 14-0 but I would like us to get 12 Ws this season (for once) to get that monkey off our back. At least 11 Ws to show we're back on track.
Yards per carry, broken down by play type (dive, off tackle, draw, etc.).
Are we analyzing FAINTs this year?
I'm amused by the number of posters who read the title and replied without reading the post itself.
At first I thought "dammit, that's not what we are going for here" but screw it, the comments section is getting fun.
Throw in some real suggestions every now and then too though.
Ok, here goes then.
I want the O-line to rollover defenses until they look like a deflated whoopee cushion.
Not sure how to stat that though.
Someone tracks that stat. They're called pancakes. And now I'm hungry for some reason...
Number of snow angels the UVA linemen make per game.
When your coach gets so many timeouts each half you have to do something to pass the time...
I'd really like to see the stats on TKP users that read before they post and those that post before they read. Interesting dynamic there.
Hmmm...I wonder where we could collect that data from? We would have to have some kind of post where it was really obvious which commenters had read the post and which hadn't...
It's just not possible. Too many outliers. You're not even considering those people who read half the post or the people who just look at the pictures.
Stats?
I want wins. To hell with the rest.
I REALLY don't know how anyone could down-vote this stat
As has been discussed in other threads... there seems to be a mass influx of downvotes recently. Some think its acceptable to just downvote something they don't like, or an opinion they don't agree with. Replying with their objection is too hard, they'd rather sit back in the shadows anonymously.
These people also need a refresher of the Community Guidelines. Downvotes are for trolls, not disagreements. If you don't like what someone has to say, counter their argument with one of your own. If you can't, that is not justification for a downvote.
Right or not, some people are getting tired of posts that clearly ignore the point of the thread, and are trying to discourage them through downvotes.
The poster who keeps downvoting gifs is another story.
Comment redacted by Alum07
Okay, you still didn't read the original post, did you?
He's asking what sort of original statistics you'd like to see in game previews, so saying "as long as we win the game" is stupid because he's talking about before the game has even been played. He asked a legitimate question about what TKP players would like to know in advance of each football game, and appears to be willing to do this analysis for the sole benefit of the readers, and yet you ignore the post again, after it's been pointed out multiple times that people are ignoring the post, and then say that "the original question was a stupid one"?
The OP is trying to make TKP even better. Read the post.
Seeing as how you've been a member of this site for just over a month, you may be unfamiliar with what joelestra does during the season. He included one of his posts in the OP, so I suggest looking at it. His contributions are part of a slew of game week posts that look at the opponent and determine what we can expect. He compares the offensive and defensive productivity of each team to show where each team's strengths lie. It's not the end-all-be-all, but it is a look at numbers that help some people see what to expect for the next week. So rather than call a question from someone who puts a good deal of effort into his work for the benefit of everyone on this FREE site stupid because you could care less, refrain from reading them and move on.
You're still missing the point.
IT'S NOT WHAT STATS YOU WANT US TO HAVE. JOELESTRA RUNS AN ANALYTICS-BASED PREVIEW BEFORE EVERY GAME USING ADVANCED DEFENSE AND OFFENSE METRICS AND RANKINGS. HE WANTS TO KNOW WHAT KIND OF ANALYSIS HR SHOULD ADD TO MAKE HIS COLUMN BETTER.
The stats in question are used to evaluate VT and their opponents before the game. If you don't believe stats can help tell the story of how good a team is, that's fine. I think it's ignorant to ignore the detail they can provide, but whatever, man.
tl;dr read before posting
I find it funny that you're blasting me for (and I will admit this was the case) not reading the original comment, but then at the very end of this comment you put 'tl;dr'
Seriously? Too long, didn't read? Isn't that the EXACT SAME THING you're going off about right now? Hypocrite much?
That's one stat I'd like to see. Less downvotes (not speaking exclusively for myself, but all around).
3rd down conversions, red zone efficiency, rushing yards/attempt, scoring both on offense and defense.
LESS TURNOVERS!!!!!!!
also, better yards per rush and 3rd Down conversation
Try reading the post.
I suggest having a conversation with Mike London about 3rd Downs. You know they were pretty good at 3rd down defense last year.
Some of the comments on this from people not reading the original post have me like ....
I'm not sure if there are statistics other than passing efficiency to go QB vs Secondary, really, but that would be nice to see.
Like what's the likelihood of seeing the QB either on the ground, or completing passes.
Or, is there a coaching competency comparison stat? Like OC playcaller vs DC playcaller for each team?