|
Post by 4verts on Mar 18, 2013 1:23:47 GMT -6
To what extent do you use analytics in your program? Offensively, Defensively, Special Teams, Clock/Game Management, Practice setup, Installations, Training, Recruitment/Retention, Staffing, etc...
Interested in any and all ways you use numbers/data to help you in running your program.
|
|
|
Post by Chris Clement on Mar 18, 2013 6:30:06 GMT -6
I run every regression I can think of on our data. Most of it comes out as garbage, but I get a few nuggets a year out of it. It also helps to do it in conjunction with film study, so I look at WHY some plays did or didn't work, or why, say, our IZ was our most explosive play but also had our most negatives and no-gains. The stats tipped me off and then I could do follow-up with film. There was a good thread on this called "Applying Statistical Methods."
|
|
|
Post by tigerpride on Mar 19, 2013 14:51:00 GMT -6
I had discussion of this on Saturday with the staff.
We spent countless hours on our quick game, and ran it about 3-4% of the time. This was expensive part of our offensive preparation.
I'm embarrassed to say that we spent countless hours on inside zone blocking and ran the play zero times this season. We trashed it in August because it simply was not working for us. We were better at down blocking, trapping, pulling, etc. This was a terribly expensive play for us, especially when you saw it on the field 0% of the time.
We were talking about the percentages of times we ran a specific play vs the time we devoted to practicing that play concept. We had 565 plays in 10 games and ran power 78 times. Did we devote 14% of our practice time to running power? Probably not. Did we devote 14% of our offensive run practice (indy and inside drill) to talking power? Probably so.
We ran an overbalanced set 38% of the time. By coincidence, it was our most successful formation. We did not devote 38% of our offensive team period towards this set/formation.
One thing I did notice (and i dont have a number) is that when we run power, two things can happen: kick out EMLOS and lead thru OR pin/log the EMLOS and wrap around. Well, we attempted to log the closing EMLOS more than kick him out and we did not prepare our guys enough to do this. We spent more time kicking out and leading thru. Once we had success running power, teams tried to squeeze/spill the play outside. We must devote more time to the pin/wrap component of power.
We punted 37 times this year, about 3% of the time. We punted as much as we ran our quick game - and did not spend near enough time on punt.
|
|
|
Post by larrymoe on Mar 19, 2013 17:39:40 GMT -6
Never.
|
|
|
Post by Chris Clement on Mar 19, 2013 17:53:37 GMT -6
Excellent analysis of practice time vs game time. Every play *should* be run more than it's practice time dictates because you inevitably have loss in your practices, time not directly attributable to one particular play or even a family of plays. In real terms, you will "overpractice" some plays because they're your home run plays, they emerge between once a game and once a year, but they work almost always because you hang on to them as a hole card.
It would also be interesting to compare those numbers to the % yardage gained vs % practice time, that might also be telling, as well as seeing first downs/practice time spent
Any minute now the old codgers will come in complaining that percentages don't help you down block. I think that really misses the point of the numbers, much as I would say that to anyone putting too much stock in these kinds of numbers.
|
|
|
Post by fantom on Mar 20, 2013 8:19:04 GMT -6
Any minute now the old codgers will come in complaining that percentages don't help you down block. I think that really misses the point of the numbers, much as I would say that to anyone putting too much stock in these kinds of numbers. Well, what IS the point of the numbers?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 20, 2013 8:32:12 GMT -6
This type of study has caught fire in baseball but there are many factors that play into football. In the examples mentioned, obviously if you practiced zone a lot and didn't run it in a game, those may have been wasted reps. We would occasionally practice a down play this year and I don't think we ever ran it in a game. However, our OL rules on that play were very similar to other plays we ran, so they got a rep at those skills. If we broke down the numbers, they would say that the times we ran Down in practice were wasted but I don't know if that's really true, at least in terms of the OL play. I'm not totally dismissing the study as it would have some value. The study of analytics would be an analytical study in itself--would the time you put into such a study justify the means you could get out of it?
I hope that makes sense. I feel like I rambled a bit!
|
|
|
Post by Luther Van Dam on Mar 20, 2013 11:06:30 GMT -6
Any minute now the old codgers will come in complaining that percentages don't help you down block. I think that really misses the point of the numbers, much as I would say that to anyone putting too much stock in these kinds of numbers. Well, what IS the point of the numbers? In this case, understanding how to be more efficient in your practice time management. I think that is pretty valuable.
|
|
|
Post by Scott Bronkema on Mar 20, 2013 11:27:19 GMT -6
the reasons why most football stats are garbage is because they are. as more math junkies try and replicate baseball metrics to football and coaches learn to correctly apply the information the better they will help.
|
|
|
Post by Chris Clement on Mar 20, 2013 11:43:36 GMT -6
The point of these numbers is to find the occasional glaring error. A 3% difference in the efficiency of two plays isn't something to go into, but if you run enough tests (easy and quick if you know how) you find one or two things that you were doing wrong, and then you can fix that. It's not like baseball where tiny differences matter because you have confidence in them, but if you notice that you never once completed a pass on 3-short but you were 60% on 3-long that's worth looking into.
|
|
|
Post by tigerpride on Mar 20, 2013 12:34:18 GMT -6
the reasons why most football stats are garbage is because they are. as more math junkies try and replicate baseball metrics to football and coaches learn to correctly apply the information the better they will help. I can see where stats could be garbage, especially in a year when you win 9 games by a margin of 35 points and you have so much junked up stats. We won 3 games this year with 10 sophomores on the field. Our largest margin of victory was 6. Every single yard, reception and first down conversion was a clutch play. More than any year, our staff simplified things, worked very hard on things and our players executed because of the work on the practice field and how it rolled into game time. I am not saying that since we were in a single formation X% of the time, that we need to line up in the formation X% this year. I am just saying what you get out of something (especially when you are not very talented) is what you get out of it. We were 100% (5-5) on 2 point plays. We ran three powers and two ISO's. Clements may be running some numbers that are a bit more sophisticated than my simple view of my numbers. I simply ran a HUDL hit chart and said wow, I called Hitch 6 times out of 565 plays and practiced it for 20 hours, lol.
|
|
|
Post by fantom on Mar 20, 2013 13:02:39 GMT -6
Well, what IS the point of the numbers? In this case, understanding how to be more efficient in your practice time management. I think that is pretty valuable. A few thoughts: 1. In some cases, if there's a really glaring problem, do you really need a study to tell you? I just read a story about a study that said that sugary drinks make you fat. No $hit? If you spent very little practice time on punt is it a surprise that your punt team was lousy? 2. As was asked above, is the time spent on this cost-effective? If you're a math guy and do it almost as a hobby, fine. If not is the info provided by a comprehensive study really valuable enough to do the study? 3. Does studying last year's practices really telling me much about how to practice next year? If you ended up never running inside zone does that mean that the time practicing it was wasted? Only if you knew ahead of time that you weren't going to run it. Otherwise, it certainly would be worth it to figure out why you didn't run it but do you really need a statistical analysis to tell you that?
|
|
|
Post by larrymoe on Mar 20, 2013 14:00:07 GMT -6
Ok- here's a question for the stat heads-
We practiced one play at least 2 times a team time during the season on offensive days and at least twice on Thursdays, with countless times running it in individual/7on7 type work. So, I would estimate that we ran it probably 9 times a week over a 7 week period. So let's just say we ran it 63 times in practice all year long. We ran it once.
Is that a waste of time?
|
|
|
Post by Chris Clement on Mar 20, 2013 14:04:32 GMT -6
Depends, did you score a touchdown with it?
Really it depends, how many total plays did you run in practice, why did or didn't the play work? The sample size is too small to be sure, but unless that play was a sure thing touchdown, it was probably a waste. Now you have to decide whether you scrap or if there's a justification for keeping it.
|
|
|
Post by Chris Clement on Mar 20, 2013 14:09:53 GMT -6
In this case, understanding how to be more efficient in your practice time management. I think that is pretty valuable. A few thoughts: 1. In some cases, if there's a really glaring problem, do you really need a study to tell you? I just read a story about a study that said that sugary drinks make you fat. No $hit? If you spent very little practice time on punt is it a surprise that your punt team was lousy? 2. As was asked above, is the time spent on this cost-effective? If you're a math guy and do it almost as a hobby, fine. If not is the info provided by a comprehensive study really valuable enough to do the study? 3. Does studying last year's practices really telling me much about how to practice next year? If you ended up never running inside zone does that mean that the time practicing it was wasted? Only if you knew ahead of time that you weren't going to run it. Otherwise, it certainly would be worth it to figure out why you didn't run it but do you really need a statistical analysis to tell you that? 1. Sometimes you use it to discover something you'd overlooked, and you use it to test a theory you have, sometimes you can misconstrue the facts. Usually it tells you what you already know, the utility is when it doesn't agree with the obvious, that's when you investigate. 2. It's easy for me, I enjoy it. If you don't and noone on your staff does, don't bother. 3. It means you need to think long and hard about whether it's worth keeping. If you didn't run it ever then you don't need math to tell you it wasn't useful. If you find it took double the practice time for half the effectiveness (by yards, yds/play, 1D, TD, whatever your metric is) compared to another base play then consider dumping or changing it.
|
|
|
Post by tigerpride on Mar 20, 2013 14:40:39 GMT -6
I don't think Clements said that analytics was the end all be all of football planning.
And for me, it took 5 minutes to run my report because the plays and formations were already in HUDL. So to Not sit down and view a report would be doing my kids an injustice. Isn't it my job to make sure that I evaluate the strengths and weaknesses of my players, my former season and how we prepared.
I did the same thing last off-season and came to some different conclusions: 1. I want to slow down the offense, huddle up and dedicate to the run. 2. Get back under center and away from the spread, and as a result, i will be a better coach. and that seemed to work.
|
|
|
Post by fantom on Mar 20, 2013 15:31:09 GMT -6
I don't think Clements said that analytics was the end all be all of football planning. And for me, it took 5 minutes to run my report because the plays and formations were already in HUDL. So to sit down and view a report would be doing my kids an injustice. Isn't it my job to make sure that I evaluate the strengths and weaknesses of my players, my former season and how we prepared. I did the same thing last off-season and came to some different conclusions: 1. I want to slow down the offense, huddle up and dedicate to the run. 2. Get back under center and away from the spread, and as a result, i will be a better coach. and that seemed to work. I'm not questioning the need to analyze the previous season. Of course you should. Do you need to run a statistical analysis? I doubt it. I still believe that all "wasted" practice isn't wasted. Why would you be embarrassed at the time you put into zone? You didn't know you were going to chuck it out. How else were you going to find out that you didn't like it? We spend practice time every week on stuff that we didn't use: maybe we didn't like how it looked in practice; maybe we found out during the game that we didn't need it. Was that a waste? Only in retrospect.
|
|
|
Post by tigerpride on Mar 20, 2013 15:43:44 GMT -6
I agree and I know that staffs throw a ton of x's and o's at their team. Over time they throw a lasso around the things they are good at and toss the stuff they are not effective with. Inside zone is an example that I used. I understand it fully well but have not been able to get my high school kids to do it wel enough. This has been a struggle for many years and I am not installing it at all this year.
My report is not sophisticated. It simply tells us % of times we ran specific plays and the number of times we were in specific formations. I also know that my opponents are studying my films this offseason and are going to have better adjustments this year! Im looking at what I want to keep the same from last year and what wrinkles I want to add.
I am just trying lessen the wasted time if I can help it. There is a need for it and we need to try new things, put players in new positions, etc.
|
|
|
Post by silkyice on Mar 20, 2013 20:14:45 GMT -6
Ok- here's a question for the stat heads- We practiced one play at least 2 times a team time during the season on offensive days and at least twice on Thursdays, with countless times running it in individual/7on7 type work. So, I would estimate that we ran it probably 9 times a week over a 7 week period. So let's just say we ran it 63 times in practice all year long. We ran it once. Is that a waste of time? We practiced one play probably 5 times a week for 16 weeks. We used it once. It went for 70 yards with 6 minutes to go in the state championship game with game tied 20-20. We won 27-16. It was worth it. That means we used it with 6 minutes left in the season. If we would have been up at that time, the play NEVER would have been called. It was worth it. But it also was a play that I knew was for that special situation. I was going to call it the series before but we were on the wrong hash and it was 3rd and long. I knew that it just wasn't time for it. We needed to be on the left hash and still be in a running situation since it was play-action. But you can't have too many of these types of plays.
|
|
|
Post by silkyice on Mar 20, 2013 20:20:52 GMT -6
Analytics are great for pro baseball. Every batter is quality and every pitcher is quality. It is a one on one matchup. Not 11 on 11. 95% of the situations are the same. The batter is trying to get a hit and the pitcher is trying to get him out. Most batters get over 500 plate appearance in a season.
In football, your opponent is different. Their scheme is different. Their personnel is different. 3rd and 10 is vastly different than 2nd and 1. Being on the 40 is vastly different than being on 2 on either end of the field. If your starting center in or the backup? Is their stud MLB in or is he hurt? Is it raining? What is the score? The situations change so much from play to play and game to game, and their really aren't that many chances. If you run hitch 5 times a game, that actually is a decent amount. That is 50 for the season, not 500.
|
|
|
Post by Scott Bronkema on Mar 20, 2013 20:35:00 GMT -6
silkyice said it best. This is not baseball its the truest team game out there. there are some advanced metrics that are worth your time that were discussed in the hidden game of football. A couple that I've found in some of my research is a 1st down conversion rate which most teams only get 35% rate. that is first downs/total plays. Anytime your getting first downs consistently your going to score. the other is yards above average. This is a little tougher for HS but you find the average for yards per carry by all running backs and find the average for that year. my 10 year study of the NFL is roughly 4. something something. about 25% of the league had more than that. Obviously you have some outliars the guy that carried it twice and had 20 yards. A lot of them have a good amount of carries and maybe you have a young tailback that needs to get some more carries than your giving them. maybe you have a bruiser that is just getting you 3 yards a pop. Its not ground breaking but numbers tend not to lie. It just depends where you look. Are they perfect? absolutely not, are they helpful possibly?
|
|
|
Post by Chris Clement on Mar 20, 2013 21:17:02 GMT -6
First down conversion rate isn't first downs/total plays, it's how many 1-10 you turn into another 1-10. It is a very useful metric that I look at.
|
|
|
Post by larrymoe on Mar 21, 2013 7:22:16 GMT -6
We practiced one play probably 5 times a week for 16 weeks. We used it once. It went for 70 yards with 6 minutes to go in the state championship game with game tied 20-20. We won 27-16. It was worth it. That means we used it with 6 minutes left in the season. If we would have been up at that time, the play NEVER would have been called. It was worth it. But it also was a play that I knew was for that special situation. I was going to call it the series before but we were on the wrong hash and it was 3rd and long. I knew that it just wasn't time for it. We needed to be on the left hash and still be in a running situation since it was play-action. But you can't have too many of these types of plays. This is sort of why I ask. The play I was asking about was a Tailback throwback pass to our QB. We practiced it all the time, ran it once. I would say it was worth it as it went for a 95 yard touchdown that eventually won us a conference championship and led to an undefeated season. We also similarly practiced a WR reverse the same amount of time last year and only used it once. It lead to a 50 yard gain that would lead to the game winning score in the same game as the TB pass. Look, if you have the time and resources and desire to look the numbers of this stuff up, God bless you and good for you. I have neither, nor a desire. I know what we do well, and I know what we don't. I don't need to run a bunch of numbers to figure it out. Last year we tried to run midline and sucked at it. About week 4 one of my assts asks me "Why don't we ever run midline in practice?" I said "Because we suck at it and I'm never going to call it in a game." If you need numbers to reaffirm what you already know, knock yourself out. I would think the only thing running numbers on that would be helpful for me is to find out if I have any playcalling tendencies. I imagine I do. Just like anyone else.
|
|
|
Post by tigerpride on Mar 21, 2013 8:20:07 GMT -6
Maybe the analytic guys need to figure how much time they are spending on analytics, lol. This is my 5th reply in two days.
I think it's good to reflect on our past season. As mentioned earlier, we were 5-5 on 2 pt plays. We ran all 5 times. One of them was a "walk off win" in overtime when I went for the win. On all 5 occasions, we motioned a WR and ran power or ISO away from the motion. Man coverage took a defensive player out of the picture. We ran to that area and scored. Two point plays are difficult calls, but we called simple plays that we executed well and practiced in those situations. Dont know if this is analytics or simple football common sense, but its worth discussing. Most teams are in man coverage on the goal line, so doesn't that fact come from analytics? And there are so many concepts that coaches try to take advantage of near the goal line vs man coverage.
|
|
|
Post by Scott Bronkema on Mar 21, 2013 11:06:44 GMT -6
First down conversion rate isn't first downs/total plays, it's how many 1-10 you turn into another 1-10. It is a very useful metric that I look at. maybe that's the way you look at first down conversions and maybe I need a better name for it but it is important to not how many of your plays are getting you first downs.
|
|
|
Post by 33coach on Mar 21, 2013 11:24:21 GMT -6
well, my background is databases and computer science, and 40% of my day to day job is reporting and data analysis.
and i dont do much analysis when it comes to football, the basics...self scouting...and even then that is pretty light.
I believe i should. the power of past data is proven...i mean do you enjoy facing an opponent without having watched film on them?
|
|
jmg999
Junior Member
Posts: 263
|
Post by jmg999 on Mar 28, 2014 18:49:09 GMT -6
There are different ways to look at data. Most of what's been discussed in this thread relates to descriptive statistics. This includes means, which are a notoriously unreliable method for evaluating significance in data sets. And, that's the key: Significance. Statistical analyses, in and of itself, isn't what's useful. It's significant data derived from these analyses that are useful. Sometimes, data that isn't significant is useful, as well. Running a regression analysis can give you all sorts of information about the relationship amongst variables, but, in the case of football, w/out performing a post-hoc analysis on the results, there's no way to determine if the variance amongst the data sets is significant, or if it's simply the result of error plus treatment conditions. For the non-statistically inclined, it's really not as complicated as it may sound.
For instance, you may have a difference of 0.07 between data sets. And, that may be statistically significant. You may also have a difference of 32 between two other sets of data, and that may not be statistically significant. It depends on context and the rate at which error is attributable. However, there are tests to ensure the validity and reliability of data. This is really why statistical analyses are important: What seems obviously significant or insignificant is not always the case. An actual mathematical analysis is required to measure this significance.
As for football vs. baseball. While it may seem that baseball requires a much less complicated model to perform a regression analysis, this really isn't the case. If I had to venture a guess, I would think that the control variables for each would be relatively on par w/ one another. In any case, an analysis of this type could be performed on football. It would just be a complex model that accounts for all the variables, but it could be done.
|
|
|
Post by coachd5085 on Mar 28, 2014 19:05:46 GMT -6
We practiced one play probably 5 times a week for 16 weeks. We used it once. It went for 70 yards with 6 minutes to go in the state championship game with game tied 20-20. We won 27-16. It was worth it. A play that scores 6 points AND takes away 4 of your opponents points? Must be the greatest play in all of football
|
|
|
Post by 4verts on Mar 28, 2014 19:09:16 GMT -6
|
|
jmg999
Junior Member
Posts: 263
|
Post by jmg999 on Mar 28, 2014 19:17:21 GMT -6
What did you do there, coach?
|
|